The article mostly focuses on ChatGPT uses, but hard to say if ChatGPT is going to be the main revenue driver. It could be! Also unclear if the underlying report is underconsidering the other products.
It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me. There will be challenges in capturing it and challenges with user trust, but it seems super promising because it will likely be harder to block and has a lot of intent context that should make it like search advertising++. And for context, search advertising is 40% of digital ad revenue.
Seems like the error bars have to be pretty big on these estimates.
IMO the key problem that OpenAI have is that they are all-in on AGI. Unlike a Google, they don't have anything else of any value. If AGI is not possible, or is at least not in reach within the next decade or so, OpenAI will have a product in the form of AI models that have basically zero moat. They will be Netscape in a world where Microsoft is giving away Internet Explorer for free.
Meanwhile, Google would be perfectly fine. They can just integrate whatever improvements the actually existing AI models offer into their other products.
OpenAI is still de facto the market leader in terms of selling tokens.
"zero moat" - it's a big enough moat that only maybe four companies in the world have that level of capability, they have the strongest global brand awareness and direct user base, they have some tooling and integrations which are relatively unique etc..
'Cloud' is a bigger business than AI at least today, and what is 'AWS moat'? When AWS started out, they had 0 reach into Enterprise while Google and AWS had infinity capital and integration with business and they still lost.
There's a lot of talk of this tech as though it's a commodity, it really isn't.
The evidence is in the context of the article aka this is an extraordinary expensive market to compete in. Their lack of deep pockets may be the problem, less so than everything else.
This should be an existential concern for AI market as a whole, much like Oil companies before highway project buildout as the only entities able to afford to build toll roads. Did we want Exxon owning all of the Highways 'because free market'?
Even more than Chips, the costs are energy and other issues, for which Chinese government has a national strategy which is absolutely already impacting the AI market. If they're able to build out 10x data centres at offer 1/10th the price at least for all the non-Frontier LLM, and some right at the Frontier, well, that would be bad in the geopolitical sense.
I think you're measuring the moat of developing the first LLMs but the moat to care about is what it'll take to clone the final profit generating product. Sometimes the OG tech leader is also the long term winner, many times they are not. Until you know what the actual giant profit generator is (e.g. for Google it was ads) then it's not really possible to say how much of a moat will be kept around it. Right now, the giant profit generator is not seeming to be the number of tokens generated itself - that is really coming at a massive loss.
Selling tokens at a massive loss, burning billions a quarter isn't the win you think it is. They don't have a moat bc they literally just lost the lead, you only can have a moat when you are the dominant market leader which they never were in the first place.
The AWS moat is a web of bespoke product lock-in and exorbitant egress fees. Switching cloud providers can be a huge hassle if you didn't architect your whole system to be as vendor-agnostic as possible.
If OpenAI eliminated their free tier today, how many customers would actually stick around instead is going to Google's free AI? It's way easier to swap out a model. I use multiple models every day until the free frontier tokens run out, then I switch.
That said, idk why Claude seems to be the only one that does decent agents, but that's not exactly a moat; it's just product superiority. Google and OAI offer the same exact product (albeit at a slightly lower level of quality) and switching is effortless.
I mean, on your Cloud point I think AWS' moat might arguably be a set of deep integrations between services, and friendly API's that allow developers to quickly integrate and iterate.
If AWS' was still just EC2, and S3 then I would argue they had very little moat indeed.
Now, when it comes to Generative AI models, we will need to see where the dust settles. But open-weight alternatives have shown that you can get a decent level of performance on consumer grade hardware.
Training AI is absolutely a task that needs deep pockets, and heavy scale. If we settle into a world where improvements are iterative, the tooling is largely interoperable... Then OpenAI are going to have to start finding ways of making money that are not providing API access to a model. They will have to build a moat. And that moat may well be a deep set of integrations, and an ecosystem that makes moving away hard, as it arguably is with the cloud.
EC2 and S3 moat comes from extreme economies of scale. Only Google and Microsoft can compete. You would never be able to achieve S3 profitability because you are not going to get same hardware deals, same peering agreements, same data center optimization advantages. On top of that there is extremely optimized software stack (S3 runs at ~98% utilization, capacity deployed just couple weeks in advance, i.e. if they don’t install new storage, they will run out of capacity in a month).
I wouldn't call it a moat. A moat is more about switching costs rather than quality differentiation. You have a moat when your customers don't want to switch to a competitor despite that competitor having a superior product at a better price.
I've also thought of this and what's more, Google's platform provides them with training from YouTube, optimal backend access to the Google Search index for grounding from an engine they've honed for decades, training from their smartphones, smart home devices and TV's, Google Cloud... And as you say, also the reverse; empowering their services from said AI, too.
They can also run AI as a loss leader like with Antigravity.
Meanwhile, OpenAI looks like they're fumbling with that immediately controversial statement about allowing NSFW after adult verification, and that strange AI social network which mostly led to Sora memes outside of it.
I think they're going to need to do better. As for coding tools, Anthropic is an ever stronger contender there, if they weren't pressured from Google already.
The moat for any frontier LLM developer will be access to proprietary training data. OpenAI is spending some of their cash to license exclusive rights to third party data, and also hiring human experts in certain fields just to create more internal training data. Of course their competitors are also doing the same. We may end up in a situation where each LLM ends up superior in some domains and inferior in others depending on access to high quality training data.
> OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome.
Note that it doesn't say: "Our mission is to maximize shareholder value, and we develop AI systems to do that".
In fairness, no company’s mission statement says “maximize shareholder value” because it doesn’t need to be said - it’s implicit. But I agree that AGI is at the forefront of OpenAI’s mission in a way it isn’t for Google - the nonprofit roots are not gone.
I don't know what the moneyed insiders think OpenAI is about, but Sam Altman's public facing thoughts (which I consider to be marketing) are definitely oriented toward making it look like they are all-in on AGI:
See:
(1) https://blog.samaltman.com/the-gentle-singularity (June, 2025)
- "We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence, and at least so far it’s much less weird than it seems like it should be."
- " It’s hard to even imagine today what we will have discovered by 2035; maybe we will go from solving high-energy physics one year to beginning space colonization the next year; or from a major materials science breakthrough one year to true high-bandwidth brain-computer interfaces the next year."
- "In a decade, perhaps everyone on earth will be capable of accomplishing more than the most impactful person can today."
(3) https://blog.samaltman.com/reflections (Jan, 2025)
- "We started OpenAI almost nine years ago because we believed that AGI was possible, and that it could be the most impactful technology in human history"
- "We are now confident we know how to build AGI as we have traditionally understood it."
(4) https://ia.samaltman.com/ (Sep, 2024)
- "This may turn out to be the most consequential fact about all of history so far. It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there."
(5) https://blog.samaltman.com/the-merge (Dec, 2017)
- "A popular topic in Silicon Valley is talking about what year humans and machines will merge (or, if not, what year humans will get surpassed by rapidly improving AI or a genetically enhanced species). Most guesses seem to be between 2025 and 2075."
(I omitted about as many essays. The hype is strong in this one.)
Not only this, but there is a compounded bet that it’ll be OpenAI that cracks AGI and not another lab, particularly Google from which LLMs come in the first place. What makes OpenAI researchers so special at this point?
This is how I look at Meta as well. Despite how much it is hated on here fb/ig/whatsapp aren’t dying.
AI not getting much better from here is probably in their best interest even.
It’s just good enough to create the slop their users love to post and engage with. The tools for advertisers are pretty good and just need better products around current models.
And without new training costs “everyone” says inference is profitable now, so they can keep all the slopgen tools around for users after the bubble.
Right now the media is riding the wave of TPUs they for some reason didn’t know existed last week. But Google and meta have the most to gain from AI not having any more massive leaps towards agi.
> IMO the key problem that OpenAI have is that they are all-in on AGI
I think this needs to be said again.
Also, not only do we not know if AGI is possible, but generally speaking, it doesn't bring much value if it is.
At that point we're talking about up-ending 10,000 years of human society and economics, assuming that the AGI doesn't decide humans are too dangerous to keep around and have the ability to wipe us out.
If I'm a worker or business owner, I don't need AGI. I need something that gets x task done with a y increase in efficiency. Most models today can do that provided the right training for the person using the model.
The SV obsession with AGI is more of a self-important Frankenstein-meets-Pascal's Wager proposition than it is a value proposition. It needs to end.
Also, they'll have garbage because the curve is sinusoidal and not anything else. Regardless of the moat, the models won't be powerful enough to do a significant amount of work.
They're both all in on being a starting point to the Internet. Painting with a broad brush that was Facebook or Google Search. Now it's Facebook, Google Search, and ChatGPT.
There is absolutely a moat. OpenAI is going to have a staggering amount of data on its users. People tell ChatGPT everything and it probably won't be limited to what people directly tell ChatGPT.
I think the future is something like how everyone built their website with Google Analytics. Everyone will use OpenAI because they will have a ton of context on their users that will make your chatbot better. It's a self perpetuating cycle because OpenAI will have the users to refine their product against.
yeah but your argument is true for every llm provider. so i don't see how it's a moat since everyone who can raise money to offer an llm can do the same thing. and google and microsoft doesn't need to find llm revenue it can always offer it at a loss if it chooses unless it's other revenue streams suddenly evaporate. and tbh i kind of doubt personalization is as deep of a moat as you think it is.
Google can offer their services for free for a lot longer than OpenAI can, and already does to students. DeepSeek offers their competitor product to ChatGPT for free to everyone already.
> It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me.
I'm not super bullish on "AI" in general (despite, or maybe because of working in this space the last few years), but strongly agree that the advertising revenue that LLM providers will capture can be potentially huge.
Even if LLMs never deliver on their big technical promises, I know so many casual users of LLMs that basically have replaced their own thought process with "AI". But this is an insane opportunity for marketing/advertising that stands to be a much of a sea change in the space as Google was (if not more so).
People trust LLMs with tons of personal information, and then also trust it to advise them. Give this behavior a few more years to continue to normalize and product recommendations from AI will be as trusted as those from a close friends. This is the holy grail of marketing.
I was having dinner with some friends and one asked "Why doesn't Claude link to Amazon when recommending a book? Couldn't they make a ton in affiliate links?" My response was that I suspect Anthropic would rather pass on that easy revenue to build trust so that one day they can recommend and sell the book to you.
And, because everything about LLMs is closed and private, I suspect we won't even know when this is happening. There's a world where you ask an LLM for a recipe, it provides all the ingredients for your meal from paid sponsors, then schedules to have them delivered to your door bypassing Amazon all together.
All of this can be achieved with just adding layers on to what AI already is today.
> It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me. There will be challenges in capturing it and challenges with user trust, but it seems super promising because it will likely be harder to block and has a lot of intent context that should make it like search advertising++. And for context, search advertising is 40% of digital ad revenue.
Yeah, I don't like that estimate. It's either way too low, or much too high. Like, I've seen no sign of OpenAI building an ads team or product, which they'd need to do soon if it's going to contribute meaningful revenue by 2030.
> Like, I've seen no sign of OpenAI building an ads team or product
You just haven't been paying attention. They hired Fidji Simo to lead applications in may, she led monetization/ads at facebook for a decade and have been staffing up aggressively with pros.
Reading between the lines in interview with wired last week[0], they're about to go all in with ads across the board, not just the free version. Start with free, expand everywhere. The monetization opportunities in chatgpt are going to make what google offers with adwords look quaint, and every CMO/performance marketer is going to go in head first. 2% is tiny IMO.
At least the description is not at all about building an adtech platform inside OpenAI, it's about optimizing their marketing spend (which being a big brand, makes sense).
There are a bunch of people from FB at OpenAI, so they could staff an adtech team internally I think, but I also think they might not be looking at ads yet, with having "higher" ambitions (at least not the typical ads machine ala FB/Google). Also if they really needed to monetize, I bet they could wire up Meta ads platform to buy on ChatGPT, saving themselves a decade of building a solid buying platform for marketers.
> There are a bunch of people from FB at OpenAI, so they could staff an adtech team internally I think
Well they have Fidji, so she could definitely recruit enough people to make it work.
> with having "higher" ambitions (at least not the typical ads machine ala FB/Google)
Everyone has higher ambitions till the bills come due. Instagram was once going to only have thoughtfully artisan brand content and now it's just DR (like every other place on the Internet).
> At least the description is not at all about building an adtech platform inside OpenAI, it's about optimizing their marketing spend (which being a big brand, makes sense).
The job description has both, suggesting that they're hedging their bets. They want someone to build attribution systems which is both wildly, wildly ambitious and not necessary unless they want to sell ads.
> I bet they could wire up Meta ads platform to buy on ChatGPT, saving themselves a decade of building a solid buying platform for marketers.
Wouldn't work. The Meta ads system is so tuned for feed based ranking that I suspect they wouldn't gain much from this approach.
no. this role is for running ads campaigns at scale (on google, meta, etc) to grow openai users. its at a large enough scale it's called "platform" but it would be internal use only.
> Your role will include projects such as developing campaign management tools, integrating with major ad platforms, building real-time attribution and reporting pipelines, and enabling experimentation frameworks to optimize our objectives.
Actually yes (I did mean to check again but I hadn't seen evidence of this before).
I do think that this seems odd, looks like they're hiring an IC to build some of this stuff, which seems odd as I would have expected them to be hiring multiple teams.
That being said, the earliest they could start making decent money from this is 2028, and if we don't see them hire a real sales team by next March then it's more likely to be 2030 or so.
Thanks for calling this out. Here is a better comparison. Before Google was founded, the market for online search advertising was negligible. But the global market for all advertising media spend was on the order of 400B (NYT 1998). Today, Google's advertising revenue is around 260B / year or about 60% of the entire global advertising spend circa 1998.
If you think of openAI like a new google, as in a new category-defining primary channel for consumers to search and discover products. Well, 2% does seem pretty low.
>Today, Google's advertising revenue is around 260B / year or about 60% of the entire global advertising spend circa 1998.
Or about 30% of the global advertising spend circa 2024.
I wonder if there is an upper bound on what portion of the economy can be advertising. At some point it must become saturated. People can only consume so much marketing.
But that occurred with a new form of media that people now use in more of their time than back before Google. It implies AI is growth in time spent. I think the trend is more likely that AI will replace other media.
There's also a possible scenario where the online ads market around search engines gets completely disrupted and the only remaining avenues for ad spending are around content delivery systems (social media, youtube, streaming, webpages, etc.). All other discovery happens within chatbots and they just get a revenue share whenever a chatbot refers a user to a particular product. I think ChatGPT is soon going to roll out this feature where you can do walmart shopping without leaving the chat.
Google, Meta and Microsoft have AI search as well, so OAI with no ad product or real time bidding platform isn't going to just walk in and take their market.
Google, Meta and Microsoft would have to compete on demand, i.e. users of the chat product. Not saying they won't manage, but I don't think the competition is about ad tech infrastructure as much as it is about eyeballs.
It might take Microsoft's Bing share, but Google and Meta pioneered the application of slot machine variable-reward mechanics to Facebook, Instagram and Youtube, so it would take a lot more than competing on demand to challenge them.
Several ways, although I'm not sure whether the below will happen:
1. Paid ads - ChatGPT could offer paid listings at the top of its answers, just like Google does when it provides a results page. Not all people will necessarily leave Google/Gemini for future search queries, but some of the money that used to go to Google/Bing could now go to OpenAI.
2. Behavioral targeting based on past ChatGPT queries. If you have been asking about headache remedies, you might see ads for painkillers - both within ChatGPT and as display ads across the web.
3. Affiliate / commission revenue - if you've asked for product recommendations, at least some might be affiliate links.
The revenue from the above likely wouldn't cover all costs based on their current expenditure. But it would help a bit - particularly for monetizing free users.
Plus, I'm sure there will be new advertising models that emerge in time. If an advertiser could say "I can offer $30 per new customer" and let AI figure out how to get them and send a bill, that's very different to someone setting up an ad campaign - which involves everything from audience selection and creative, to bid management and conversion rate optimization.
So I don't necessarily disagree with your suggestions, but that is just not a $1T company you're describing. That's basically a X/Twitter size company, and most agree that $44B was overpaying.
It's not that OpenAI hasn't created something impressive, it just came at to high a price. We're talking space program money, but without all the neat technologies that came along as a result. OpenAI more or less develop ONE technology, no related product or technologies are spun out of the program. To top it all off, the thing they built, apparently not that hard to replicate.
You have to take into consideration the source. FT is part of the Anthropic circle of media outlets and financial ties. It benefits them to create a draft of support to OpenAI competition, primarily Anthropic, but they(FT) also have deep ties to Google and the adtech regime.
They benefit from slowing and attacking OpenAI because there's no clear purpose for these centralized media platforms except as feeds for AI, and even then, social media and independents are higher quality sources and filters. Independents are often making more money doing their own journalism directly than the 9 to 5 office drones the big outlets are running. Print media has been on the decline for almost 3 decades now, and AI is just the latest asteroid impact, so they're desperate to stay relevant and profitable.
They're not dead yet, and they're using lawsuits and backroom deals to insert themselves into the ecosystem wherever they can.
This stuff boils down to heavily biased industry propaganda, subtly propping up their allies, overtly bashing and degrading their opponents. Maybe this will be the decade the old media institutions finally wither up and die. New media already captures more than 90% of the available attention in the market. There will be one last feeding frenzy as they bilk the boomers as hard as possible, but boomers are on their last hurrah, and they'll be the last generation for whom TV ads are meaningfully relevant.
Newspapers, broadcast TV, and radio are dead, long live the media. I, for one, welcome our new AI overlords.
All of which is great theory without any kind of evidence? Whereas the evidence pretty clearly shows OpenAI is losing tons of money and the revenue is not on track to recover it?
Tapping into AdTech is extremely hard, as it's hard driven by network effects. What you mean is "displaying ads inside OpenAI products" then, yes, achievable, but that's a miniscule part of targeted Ad markets - 2% is actually very optimistic. Otherwise, they can sell literally 0 products to existing players as they all have already established "AI" toolsets to help them for ad generation and targeting.
Query: LibraGPT, create a plan for my trip to Italia
Response: Book a car at <totally not an ad> and it will be waiting for you at arrival terminal, drive to Napoli and stay at <totally not an ad> with an amazing view. There's an amazing <totally not an ad> place that serves grandma's favorite carbonara! Do you want me to make the bookings with a totally not fake 20% discount?
I'm traveling like this all the time already, I don't understand why it's hard for people to understand that ad placement is actually easier for chat than search
But who wants that? And you're going to say that's exactly what a travel agent does, selling me stuff so he can get a kickback. But when stuff goes wrong, I'll yell at the travel agent so he has some incentive to curate his ads.
I'm not aware of any FTC rule that would preempt this sort of product as long as it met the endorsement disclosure rules (16 CFR Part 255), same as paid influencers do today.
friendzis's example showed a plausible way to generate revenue by inserting paid placements into the chat bot response without disclosures by pretending they are just honest, organic suggestions.
Right. That's not a novel idea, and this is a well-trod area of concern. That's why these FTC rules have been around for many years.
edit: to be clear, I am saying that in the absence of clear disclosures, that would run afoul of current FTC rules. And historically they have been quick to react to novel ways of misleading consumers.
I expect all hosted model providers will serve ads (regular, marked advertisements, no need for them to pretend not to, people don't care) once the first provider takes the lid off on the whole thing and it proves to be making money. There's no point in differentiating as the one hosted model with no ads because it only attracts a few people, the same way existing Google search and YouTube alternatives that respect the user are niche. Only offline, self hosted models will be ad free (in my estimation).
Assuming you know it's an ad. Ads in answers will generate a ton of revenue and you'll never know if that Hilton really is the best hotel or if they just paid the most.
This isn't a realistic concern unless FTC rules changed substantially from where they are today (see my other comment on this post for links). Sponsored link disclosures would be in place.
Everything else aside, it's simply not worth it for them to try to skirt these rules because the majority of their users (or Google's) simply don't care if something is paid placement or not, provided it meets their needs.
That's only true if you can demonstrate a substantial percentage of people would be unaware of it. The reason influencers have to disclose is some but not all take endorsement money. It would be pretty easy for OpenAI to say it was common knowledge or to bury disclosures in the fine print of the services and not every time it happened
The US federal government is now a mob-style organization. The laws, rules, and regulations that are written down are only applicable as far as Trump and those around him want them to be. Loyalty to the boss is the only inviolable rule.
In other words, if they want to put ads into chat, they just need to be perceived as well aligned to Trump to avoid any actual punishment.
Seems like ad targeting might be a tough sell here though, it’d basically have to be “trust me bro”. Like - I want to advertise coca-cola when people ask about terraforming deserts? I think I wouldn’t be either surprised by amazing success or terrifying failure.
Perplexity actually did search with references linked to websites they could relate in a graph and even that only made them like $27k.
I think the problem is on Facebook and Google you can build an actual graph because content is a thing (a url, video link etc). It will be much harder to I think convert my philosophical musings into active insights.
So few people understand how advertising on the internet works and that is I guess why Google and Meta basically print money.
Even here the idea that it’s as simple as “just sell ads” is utterly laughable and yet it’s literally the mechanism by which most of the internet operates.
It is to the point of yellow journalism. They know that the "OpenAI is going to go belly up in a week!" take is going to be popular with AI skeptics, which includes a large number of HN viewers. This thread shot up to the top of the front page almost immediately. All of that adds to the chances of roping in more subscribers.
The team also assumes LLM companies will capture 2 per cent of the digital advertising market in revenue, from slightly more than zero currently.
This seems quite low. Meta has 3.5 billion users and projected ~$200b revenue in 2025. ChatGPT is at ~1 billion so far. By 2030, let's just stay ChatGPT reaches 2 billion years or 57% of Meta's current users.
I'd like to think that OpenAI's digital ad revenue should reach 10% by 2030 an then accelerate from there. In my opinion, the data that ChatGPT has on a user is better than the inferred user data from Instagram/FB usage. I think ChatGPT can build a better advertisement profile of each user than Meta can which can lead to better ad targeting. Further more, I think ChatGPT can really create a novel advertisement platform such as learning about sponsored products directly via chat. I'm already asking ChatGPT about potential products and services everyday like medicine, travel, gadgets, etc.
I think people are severely underestimating ChatGPT as a way to make money other than subscriptions. I also think people are underestimating the branding power ChatGPT has already. All my friends have ChatGPT on their phone. None of them except me has Gemini or Claude app.
This doesn't account for OpenAI's other ambitions such as Sora app.
Hey Sam Altman or OpenAI employee, if you are reading this, I think you should buy the North American version of TikTok if the opportunity presents itself. The future of short videos will be heavily AI generated/assisted. Combine Tiktok's audience with your Sora tools and ChatGPT data and you got yourself a true Instagram competitor immediately. If the $14b sales price of US Tiktok is real, that's an absolute bargain in the grand scheme of things.
> Meta has 3.5 billion users and projected ~$200b revenue in 2025
Meta makes about $200B on ads, Google makes about $235B on ads. Advertising is roughly 1.5% of the total GDP of the US and hasn't changed in 20+ years. So what you have is a big ass pie with a few players fighting for it that barely grows every year.
OpenAI has to somehow:
1. Compete directly with Google Gemini and Meta's Llama for a piece of users pie with a product that has very little differentiator (functionality and technically speaking).
2. Have to prove to advertisers that their single dollar ad purchase on OpenAI is categorically worth more than any other channel.
3. Have enough forward capital to continue purchasing capital-intense hardware purchases.
4. Having enough capital to weather any potential economic headwinds.
OpenAI has branding power, a clear product focused mindset, focused attention, and moves faster than Google and Meta. Nearly 1 billion users in 3 years is no joke.
We're on Hacker News. Y Combinator literally teaches their companies that they can beat incumbents using focus and speed.
My bet is on OpenAI. When they IPO, I can easily see them with $1 trillion in valuation and raise the a record amount of money in an IPO.
If Meta and Google don't see OpenAI and LLMs as an existential threat, they wouldn't invest so much. I think AI has that potential to completely disrupt Google and Meta because it fundamentally changes the way people behave. It's a paradigm shift. It isn't just playing the same game.
It seems to me you're comparing apples and oranges here.
I don't think so. 1 billion users and a clear intention to deliver ads with an immense amount of data on users. That's a clear threat to both Meta and Google.
PS. That's why Meta and Google are all in on AI. OpenAI is an existential threat to both in my humble opinion.
> You didn't state reasons why not being a social platform matters here.
There's nothing to pointlessly waste your time on. You open it to do a thing, you either do the thing or get frustrated or leave. Social networks are designed to waste your time even when they outlive their usefulness, therefore they can serve you more ads.
You could argue Google is the same as ChatGPT in that regard, but that's why Google has Adsense in almost any search result you click on.
As for your group chats feature argument, anyone can make a social network, that's the easy part. Getting friend groups to switch is the more difficult part.
> PS. That's why Meta and Google are all in on AI. OpenAI is an existential threat to both in my humble opinion.
They're all in on AI because that's what their investors want them to do to "not be left behind". Meta was all in Metaverse. And on a cryptocurrency before that (Diem). And on Free Basics before that. The fact that none of those succeeded didn't hurt them at all precisely because they had an infinite money glitch known as ads.
They can afford to waste amounts of money equivalent to a yearly budget of a small country, ChatGPT can't.
> You didn't state reasons why not being a social platform matters here.
The network effects matter so much more for a social platform than a chat bot. The switching costs for a user are much lower, so users can move to a different one much easier.
How sticky will chat bots prove to be in the long term? Will OpenAI be able to maintain a lead in the space in the long term, the way Google was over Bing? It's possible, but it's also pretty easy to imagine other providers being competitive and a landscape where users move between different LLMs more fluidly
Another reason why social media matters is that people actually spend their free time in all those feeds. On the other hand, using LLMs is much more oriented towards specific utilities. Having 1B users who visit you once a day to ask for an email proofread will not make you profitable.
But, like search, it captures intent to buy much better, while looking at feeds for internatinment does not. Adwords worked because of that, they could capture ad revenue on queries that lead to sales. The amount of extra context in AI chat is even better, as is the ability to steer the conversation to different options.
I think Google is the much bigger threat. I've more or elss stopped using ChatGPT now, it's easier to just type the question directly into Google and get the response from their AI rather than navigating to chatgpt first. Anecdotal but I don't see anything long term keeping people on that site.
Problem is the day Google puts a similarly priced and more powerful chatbot similar to ChatGPT and powered by Gemini most users will switch without a second thought, and Google will get all that data about their personal life too. OpenAI having a free tier with powerful models is unsustainable, brand loyalty is a joke at this stage, people will not bother going to ChatGPT if the search bar in their browser starts a discussion with Google. I’m sure big players like Google know OpenAI is a temporary scheme to get free money, they are surfing along now but they are playing the long game.
Maybe, but OpenAI only get the data you decide to type into ChatGPT in a vibe therapy session or whatever, and they have to burn a load of GPU time to keep you feeding it more slop.
Google just passively collects email and browsing history, much better data for targeting ads and way less cost to run.
I wonder if meta is a poor comparison for advertising because they're users tend to spend more time on their products doom scrolling, as opposed to something like google, where you get your answer right away and move on.
ChatGPT is a hybrid between Google and Meta. People use it for product and service search and research. People also use it as a companion - especially young people.
>I think you should buy the North American version of TikTok if the opportunity presents itself. The future of short videos will be heavily AI generated/assisted.
I will have whatever you're smoking. If a social media platform literally proves the dead internet theory, it's not making any money.
I just realized that other industries are way larger than AI. Assuming they capture the entire advertising market, only $390 Billion was spent in the US last year. Compare that to health care, where 4.3 Trillion was spent in the US last year, or commercial banking's revenue of 1.5 Trillion, commercial real estate's 1.5 Trillion, gasoline stations' 1.10 Trillion, etc. What's amazing is, despite the fact that AI isn't making much money, is taking on considerable debt, and isn't even assured to be all that useful, one third of the stock market is now just AI crap. The economy is going to collapse because of a small, brand-new industry. This... shouldn't be possible.
Build a system which can read MRI images, now AI took over the whole MRI Analysis sector.
Let AI diagnose the avg cold, now AI took over the household / local doctor for 90% of cases.
Use AI for support, now AI took over the call-center business.
AI can already code. It might not be perfect, it might not be always good but NO ONE assumed that 2025 some matrix multiplication can mimick another human being so well that you can write with it and it will produce working code at all.
thats the hype, thats the market of ai.
And in parallel we get robotic too. Only possible because of this AI thing. Because robotic with ML is so much better than whatever we had before. Now you can talk to a robot, the robot can move, the robot can plan actions. All of these robots use some type of ML in the background. Segment Anything is possible because of AI.
I think they are missing a few very significant revenue sources that likely going to grow beyond just simple enterprise user accounts for Chat GPT.
I'll call out two of them.
Image, video, and other content generation is going to become more important and companies will be spending on that. We've seen some impressive improvements there. IMHO near term a lot of that stuff might start showing up in advertising and news content, the whole media industry is going to be a massive consumer of this stuff. And there's going to be a lot of competition for the really high quality models that can be run at scale. Five years until 2030 is a lot of time for some pretty serious improvements to land.
Another area that the article skips over is agentic tools. Those are showing a lot of promise right now. Agentic coding tools are just the tip of the iceberg here. A lot of these tools are going to be using APIs. So API revenue is a source of revenue. There are applications across the entire IT industry. SAAS, legacy software, productivity tools, etc.
Yes 207B is a lot of money. And there's no guarantee that OpenAI comes out on top "winning" all these markets of course and we can argue about how big these markets will be. But OpenAI does have a good starting position and some street credibility here. It's a big bet on revenue and potential here. But so is betting against all that and dismissing things. And there's a lot of middle ground here.
> Image, video, and other content generation is going to become more important and companies will be spending on that. We've seen some impressive improvements there.
They are assuming that people _want_ automated content enough to pay for it.
Like at the moment its great because its essentially free, but paying >200billion?
Are we assuming that we can automate the equivalent of instagram, netflix/disney and whatsapp and make advertising revenue?
I think the unmentioned pivot is into robotics, which seems to have actual tangible value (ie automation of things that are hard to do now)
Everything professionally produced you currently see uses a lot of CGI, and you can’t even tell.
Essentially what they are doing is cheaper, more accessible CGI. And in the same way at it is now, you are not going to be able to tell it was used in expensive productions and will be able to see it in the cheap productions.
Actually, I sense a mounting "AI fatigue". Angry comments under AI contents. Social media accounts that proclaim to be "AI Free". I was thinking today that the killer app for AI could be a filter that automatically exclude all AI content.
I agree. Back in the 90's there was plenty of skepticism, along with some mockery, surrounding the internet. But you never saw headlines like "Lawyer caught using the internet" or "Artist busted using the World Wide Web".
Google are experimenting with different filters that intentionally blur the lines between real and fake video in their shorts. A creator I follow who makes human-in-shot videos suddenly looked fake, like some animation filter was applied. I commented on how bad it looked, and someone said YT are doing it to people's videos as some sort of "trial". The effects were to make the human look more AI generated. The only use for this I can surmise is to make AI content even harder to detect.
It is very obviously there. People are not simply accepting the AI-bro justification that one day they won't notice the difference so it's all fine.
Ordinary people are understandably either exhausted by or angry at AI content, because it is relentless and misleading, and because they know instinctively that the people who want this to succeed also want to destroy employment and concentrate profits in the hands of a few Silicon Valley sociopaths.
> They are assuming that people _want_ automated content enough to pay for it.
I’m assuming gen ai will supplement so instead of ten artists producing content you’ll have three orchestrating and correcting but the output will be indistinguishable from what it was before. Just now the margin is much higher.
I never think of direct to consumer sales in this context.
Anecdotally I've seen lots of anti AI posts on instagram from gen-z accounts, and on YouTube ive seen "ai not used to make this music" in the description.
Also anecdotal, but I do think there's gen z anti-ai sentiment, these kids are joining a world where they will never own a house and have bleak prospects and now even art is something they can't do if AI takes off, so the market might not be there
Seems like older people love AI art, people over 50, which is a sizeable market don't get me wrong, but in 10-15 years idk
I would assume that generative AI will be able to make endless hours of personalized reality TV content that will hit a surprisingly large demographic. And then it can use recommendation algorithms to distribute that reality TV AI slop to all the people that might or might not want it in order to turn it profitable. Evidence: all the god-awful foreign reality TV on Netflix currently.
It's a bit harder to replace Disney at this point, but give it 20 years. I'm bullish on AI slop on par with The Apple Dumpling Gang or Herbie Goes to Monte Carlo within a decade or so. Evidence: The rapid evolution of AI generated video in the past few years extrapolated into the future.
I keep thinking about how, 20 years, 3D printers became accessible for tech/price, and there was a lot of talk about how we'd all just print what we need (pretty much like in Stephenson's book Diamond Age), and you'd get rich selling digital patterns for material goods.
Instead, we got the elevation of the handmade, the verifiably human created, typified by the rise of Etsy. The last 20 years have been a boom time for artists and craftspeople.
I keep seeing AI slop and thinking that all this will do is make verifiably human created content more valuable by comparison, while generative AI content will seem lowbrow and not worth the cost to make it.
Creative people are on the whole a lot better at keeping AI slop out of art and craft fairs than they have been with the lazier output of 3D printers, CNC, laser engravers and off-the-shelf resin mould art.
It is as if the relatively unspoken feelings about the downsides of technologies as a gateway to art have been rapidly refined to deal with AI (and of course, even the CNC and laser engraver people have common cause).
But I think it is fair to say that if they feel success, there will be a growing pushback against the use of 3D printers, eufyMake resin printing and CNC in a niche where hand tools used to be the norm.
And speaking even as someone who has niche product ideas that will be entirely 3D printed/CNC cut/engraved, I don't really disagree with it. I am mostly not that kind of creative person (putting aside experimental photography techniques) and I see no reason why they shouldn't push back.
The reality is that "craft" fairs are an odd mix of people who spent a lot of time refining their art and selling things that are the work of hours of expressive creativity and effort, and table upon table of glittery resin mould art and vinyl cut stencil output stuck on off-the-shelf products. I think AI might help people refine their feelings about this stuff they once felt bad/incorrect/unkind about excluding.
It's a bit like the way the art photography market is rediscovering things like carbon printing, photosensitive etching and experimental cyanotype, and getting a lot more choosy about inkjet-printed DSLR output.
Yeah, I have a hard time believing that there’s a massive demand for AI-generated videos and images. Like, why would the news industry want to generate images and videos with AI? It’s not news. The advertising industry maybe, but even then it’s probably not your top brands that go full in on it. If you see that all of Apple’s adverts are generated with AI, it’ll probably lower your brand perception.
We already put more importance on handmade goods vs. factory-made, even if the latter is cheaper and better quality. I have my doubts about humanity collectively embracing content generated from prompts by black boxes.
> Yeah, I have a hard time believing that there’s a massive demand for AI-generated videos and images. Like, why would the news industry want to generate images and videos with AI? It’s not news. The advertising industry maybe, but even then it’s probably not your top brands that go full in on it. If you see that all of Apple’s adverts are generated with AI, it’ll probably lower your brand perception.
(disclaimer: I don't work in ad and don't know more about it than the next person)
Whether the end product (the ad) is AI-generated or not is almost irrelevant. The whole production chain will likely be AIfied: to produce one ad you need to go through many concepts, gather reference images/videos, make prototypes, iterate on all that, and probably a ton of other things that I don't know about... The final ad is 1 image/video, but there's been dozens/hundreds of other images/videos produced in this process. Whether the final ad is AI-generated or not, AI will almost certainly (for better or worse...) have a major place in the production chain.
The cost of operating television studios and paying related staff, including on-air talent, is probably significant. I can easily see major news networks turning to AI-generated newsreaders. TikTok is already full of AI voiceovers; seems like a short leap, to me.
I don't quite follow this point. Master Chief is recognizable. So is Lara Croft. So is Darth Vader's voice. Networks could easily develop virtual personalities with distinctive, bankable, appealing characteristics.
They wouldn't have off-air scandals, require insurance, pensions, teams of wardrobe and makeup artists, security details; They wouldn't need to travel. And that is just the on-air talent. You can replace thousands of tv studios all over the world with a handful of workstations and compute power.
And why haven't they? Master Chief has been around since 2001, Lara Croft since 1996 and Darth Vader since 1977. The technology has been around for ages, and as far as I know, no networks have opted for virtual anchors.
Just from where you are pulling the data that on-air personalities are too expensive?
I don't have a good answer for why they haven't already. I have wondered about the possibility of doing this for 10 years or more.
"The data that on-air personalities are too expensive?" It doesn't seem to me , for the purposes of this conversation, that identification of a cost center required a quantitative analysis. The cost of human talent is non-zero, presumably large enough to merit scrutiny, and unpredictable; that is sufficient, to my way of thinking. So is the cost of the equipment and infrastructure to capture and transmit video image of that human talent, and the humans who maintain and operate that infrastructure.
We've seen several decades of human cost-reduction initiatives, across multiple industries and fields of endeavor, so I'm taking that as evidence that if there is a cost that can be reduced or removed, someone somewhere is looking at doing so. Everything from assembly-line automation, to switching to email over inter-office memos and mailrooms, to the abandonment of fixed-benefit pensions, to self-service kiosks in fast-food restaurants, demonstrates that costs will be cut where they can be cut.
> We already put more importance on handmade goods vs. factory-made, even if the latter is cheaper and better quality
I guess if 'importance' means sentimentality, however, the "factory-made things" industry certainly makes the vast majority of the money to be made in manufacturing. Handmade is a niche. I think the argument is that whether people prefer it or not, AI-generated content will have a major place in the marketplace for content because of its natural advantages. In other words, a movie heavily generated with AI won't win any Oscars but there is a world where such things would make a lot of money and some of that money would flow to OpenAI.
I have a hard time understanding why I would want to waste my time watching something generated by AI when they clearly didn't value the time spent making it.
I have a limited lifespan, I'm not going to use that to consume slop.
> IMHO near term a lot of that stuff might start showing up in advertising and news content, the whole media industry is going to be a massive consumer of this stuff.
One of the candidate for mayor of New York ran a whole AI video ad campaign.
Half of the real estate listings have AI remodeled pictures.
Probably a quarter of the printed ads I see each day are AI generated.
Real estate agents only use AI because it's cheaper than paying $70 CAD per image for human-made images. There's a cap on how valuable the AI generated images are.
The last time I sold a house, a professional photographer took some beautiful photos, gave the realtor like 60-80 and he picked 40 for the listing. He absorbed the cost of that with his commission. No way in hell did he pay $2800-5600 (I'm also in Canada) for those photos.
I think actually selling media is going to be tough. 1: you have to beat the current price point so people will use it, 2: it can’t be too cheap or you can’t make money. That mad market price point makes it hard to get right. The more content the fewer eyeballs, meaning the less value, meaning people willing to pay less. If it makes making movies a lot cheaper, then there’s a lot less money to be made.
The thing about tech is it has to scale to meet tech valuations, and there’s a reason instagram basically gives content creation tools away for free.
The main thread through the case laid out has a fundamental flaw "assumes a spherical cow in a vaccuum" .
There are cheaper/faster more open competitors out there doing as much as OAI and getting integrated in real product pipelines across the planet.
Its not a Blueray vs HD-DVD situation (winner takes all and competition dies on the vine) or early Google vs Yahoo Search i would posit its more like bitkeeper vs git - and will probably eveolve into a git/hg/bazaar type situation. Costs will only keep dropping and the easiest/cost effective options will get faster uptake in the backend dev world where it matters.If im integrating a LLM and the whole field can kinda do the equivalent capex/opex will push the decision.
What I don't understand is how Elon will ever compete with Chinese humanoid robot pricing and output. He's known for not caring about privacy, so there's no real difference between buying a Tesla one or a Chinese one. Both will look and listen into your home and transfer as much data as possible to their true owners.
He congratulates Jeff Bezos for New Glenn, and rightly so, but he's absolutely silent about the Chinese robots. One has to admit that they're already ahead of the US. This comes from a German who well recognizes how far behind Germany is in comparison to the US, in regards to any digital technology.
Oh that's easy: as soon as China gets good enough to compete at any market seen as vital, "the west" either makes up some nebulous national security threat to cut them off or imposes high enough import taxes to make the home-made versions cheaper in comparison. The same way "the west" has done it with cars, phones, and social media in recent years.
There is no free market competition between "the west" and China, they're two completely separate markets for a reason. You're only allowed to cross that isle in a limited capacity, as soon as you get big enough to be seen as a legit threat, you're getting cut off. Works both ways, Tesla can sell in China, but they can't beat BYD. Apple can sell iPhones, but only because their market share is not big enough to matter.
I don’t know if this is ridiculous, but I’m curious if access to LLMs will one day be priced like the Bloomberg Terminal or something. Where access for one user is like 20,000 dollars. Maybe less than that, but like 5k per person.
Seems crazy by most software standards, but when Bloomberg became a software only program (they stopped selling physical terminals) and people were shocked when they paid almost nothing for excel but then so much for the second tool they needed as traders.
The difference is that Bloomberg Terminals were always expensive, and so people expected to pay. LLMs are basically free (subsidized) at this point, and people are very sensitive to large price increases.
Sure and I’m sure there would be a huge shock, but simple economics would dictate that if that’s the true equilibrium of price for LLMs to be economical, then it would have to get to the price eventually
I think there could be a few directions. Consumer level LLM usage will become free. Corporate-grade LLM use will cost a lot of money. Maybe the highest grade LLM use will be regulated and only done through government labs.
What’s a high grade LLM though in such a competitive environment? And if China releases more high grade open source models, that pricing model is f8cked.
One interesting thing I heard someone say about LLM’s is this could be a people’s innovation, basically something so low margin it actually provides more value to “the people” than billionaires.
The top two comments mention that capturing, at minimum, 2% of the digital advertising market seems low. I disagree. I see that as an enormous hill to climb.
The LLMs will create new content, but they aren't creating new business channels in the advertising industry. As an example even once Google achieved search domination they still didn't have this. They had purchase a lot of things to make that happen like Urchin, Adscape, DoubleClick, YouTube, and a lot more.
It's really hard for start ups to compete for VC money: "Hello Mr. VC I'm going to burn your money for buzzword" just no longer works, first openAi has an industrial buzz word generator, second their money burning plan scales much better than my money burning plan.
Yeah but they aren't a diversified portfolio of money burning plans and that's the real secret to responsible investing. I think Warren Buffet said that.
This should be a easy barrier for OpenAI, cgpt has only been public for 3 or so years and they are already close to mediating most of the global economy and human decision making
AI credits will likely replace current money anyway
I tried it yesterday by comparing the specs of a few low-budget desktops for my father. It worked well and saved me a lot of time and the hassle of comparing multiple tabs on a PC.
I am a bit worried about the feature where it calls the shops for inventory checking. That's the whole point of having a website. Now we are going to have expensive AI that calls other AI answering machines and no value will be added. Meanwhile, it will become even more difficult to talk to a human when necessary.
Honestly, this is huge for people like me who tend to over-research and over-think the hell out of product choices. "Find me a top-fill warm-mist humidifier that looks nice, is competitively priced against similar products, and is available from a retailer in $city_name. Now watch for it to go on sale and lmk."
If they can figure out how to get the right kickbacks/referrals without compromising user trust and really nail the search and aggregation of data this could be a real money-maker.
Why would I want to spend 1-2h researching humidifiers if I can spend that time in any other way, and still end up with a humidifier that fits my needs first try?
This kind of task is perfect for AI in a way that doesn't take away too much from the human experience. I'll keep my art, but shopping can die off.
Because you end up with a $1 value piece of crap that someone spent considerable time optimizing and faking reviews for LLMs instead of on the product. Basically in the medium term this strategy will get you temu stuff
How often do you buy the first result on an Amazon search? Because that's delegating your labour, isn't it? Surely the best products are getting to the top, right? Well no, they're being paid to get to the top. An LLM that has in-app shopping is gonna be the same thing
for the same problems with amazon, youre relying on a computer to tell you what to buy, which is very shortly going to be infested with promoted products and adverts instead of genuine advice. The AI implementers will poison the responses in the name of advertising, of this i have zero doubt in my mind.
10+ years ago you could in fact just pick the best-reviewed product on Amazon at a certain price point and have a great experience! God help you if you tried that today.
Well, you can always do the same thing that an LLM would: open SEO spam ranking sites "best humidifiers 2025", filled with referral links to Amazon or other sellers, which basically copy product descriptions and assign rankings that aren't based on any tests or real data.
For this to be useful they need up to date information, so it just Googles shit and reads Reddit comments. I just don't see how that is likely to be any better than Googling shit and reading Reddit comments yourself.
If they had some direct feed of quality product information it could be interesting. But who would trust that to be impartial?
Do you buy the first item that pops up on Amazon for a search that you've made? Because that's letting the robot do it for you.
If the answer is "no because that's an ad", well, how do you know that the output from ChatGPT isn't all just products that have bought their rank in the results?
Project Farm solves the trust problem with methodology + video documentation and the monetization problem with affiliate links for every product tested.
> If they can figure out how to get the right kickbacks/referrals without compromising user trust
i'm trying to envision a situation in which the former doesn't cancel out the latter but i'm having a pretty hard time doing that. it seems inevitable that these LLM services will just become another way to deliver advertised content to users.
> If they can figure out how to get the right kickbacks/referrals without compromising user trust and really nail the search and aggregation of data this could be a real money-maker.
As another commenter points out, "not compromising user trust" seems at odds with "money-maker" in the long-term. Surely Google and other large tech companies have demonstrated that to you at this point? I don't understand why so many people think OpenAI or any of them will be any different?
LLMs have high utility for coding. The PMF is so strong that even with just coding i feel they will see an ARPU expansion beyond conservative assumptions.
There are adjacancies in white collar work like financial analysis that they will go after. All these will capture high ARPU usage.
Consumer is not their only path to revenue but it is probably the easiest to model. The enterprise play to automate and accelerate some white collar workers is a clear target not reflected here.
You misunderstand the Amazon story. Amazon did two pre-IPO funding rounds: <1m angel round and an 8m Series A led by Kleiner Perkins. That's it. In contrast, OAI has raised almost 60 billion in 10 funding rounds, and if you RtFA you can see the estimates are they have to raise hundreds of billions more.
For the IPO itself Amazon sold 3 million shares at $18, for a raise of 54m (from the IPO alone they had enough cash to pay off every investor up to that point). In July 2001, in the heart of the dotcom crash, they raised 100m by selling equity to undisclosed investors, and of course they have been using shares as part of their compensation packages for a very long time, but that's about it for Amazon's entire equity raises.
They did raise 15b in a bond issuance a few weeks ago, their first bonds issued since 2022, with the money going to several things but mostly AI. However, since bond payoffs are very different from selling equity this is a very different play from what OAI is doing. Amazon will never pay more than a fixed amount for that money, capped upside to the bond-holders.
The reason this is different is that Amazon has largely run either a small profit or a small loss, quarter after quarter, because they take their profits and instead of recording it, putting it in a bank, or dividend-ing it to shareholders they put it into building datacenters and warehouses and software and the like. But because of that enormous cash generation they have only rarely tapped outside investors, either in bonds or equity markets. OAI is not generating near enough cash to fund their operations, so they have been selling equity in absolutely enormous quantities- they have already raised more cash pre-IPO than any company in history and outside estimates like this one from HSBC call for them to blow past the amount they've already raised. This is fundamentally very very different.
OpenAI doesn’t have a product? Have we existed in the same reality for the last 3 years? Something something fastest grown user base in the history of tech
Selling books online was certainly profitable, but I'm not sure about unique. Amazon's big success is that they had no particular ties to any existing publisher so they didn't have the corporate headwinds of "this will kill our brick and mortar stores and their distribution systems!".
>In total, OpenAI aims to invest approximately $1.4 trillion in computing infrastructure – encompassing Google Cloud, Nvidia chips, and data center expansions.
Huh yeah fair. That's more than the yearly defense budget. Absurd.
Though I'm sure it's not _yearly_
The sovereign wealth funds and billionaires need something to do with all their available cash. It's no fun just watching the number increase anymore, that got boring quite a while ago.
To add to the valid counterpoints in the existing replies, Amazon was cash flow positive for most of those 20 years, so they were investing their own money back into things that could generate even more revenue (and profit!) for the company.
OpenAI is still losing money much faster than it can make it and is planning to accelerate those losses indefinitely.
I think the most likely outcome is that it turns into something like Uber, where they continue to lose money waiting for a major technological leap (truly unassisted and reliable AI in this case, fully self-driving cars in the case of Uber) and then pivot a bit to a largely unnecessary and poorly executed business model that people reluctantly use for the most part (with some eager advocates) but makes some money.
And unlike taxi drivers / companies, these competitors have much deeper pockets.
It is not hard to imagine Google being able to outlast OpenAI for a decade and it is hard to imagine OpenAI being able to survive for more than another couple of years given that.
Is there any link directly to the report? I am unable to find any in this article or a number of others which seem to just copy the same information here.
I think beyond the number of crazy assumptions (no Google taking market share in the consumer market?? only 2% of digital advertising expected to be captured by OpenAI?) it is hard to nail down which levers could move which might make this funding hole disappear.
I don’t think LLM leader is going to come out of this AI race on top. OpenAI is like Yahoo! in my view. They were there in the beginning and led by default. Someone somewhere will be the Google of this era. It won’t be LLM but the next step. Leaps and bounds better than LLM. I also think we won’t use even 1/100th of the energy needs projected right now.
The reason for this is that leverage is extremely expensive and punished in the current macro regime. OpenAI will lose money faster if they scale up. That's because the Fed can capitulate all it wants, but real interest rates and real yields won't move. The market is saying "pay me".
That are 4 words. ;-) Other then that yes, goverment contracts, meaning offloading the costs to the regular people. I think the US goes all in with AI, trying to get world dominant and controlling the market. But I think this is a mistake and the US will get into trouble with this.
It is not losing money, the whole industry haven't figured out how to generate revenue yet.
it is the starting phase of LLM like we experienced in .com bubble.
Is it a bubble at this point? Yes. Can it generate revenue in the future? Yes.
Will openAI survive in this bubble? Maybe.
It is quite literally losing money at an unsustainable rate. They need a path to profitability otherwise this is a massive boondoggle for all of the investors
Yeah, financial bubble != useless technology. Maybe coding agents do cost $50 per month in the long term, but I might just pay that for entertainment and personal stuff. Like, I don't even try to vibe code my job, but in the evenings having a cool slop generator is good times.
Right now I use a Chinese vibe code plan, really good value.
The most interesting point about OpenAI I have heard lately is they are literally trying to make themselves too big to fail. If they go down so does everyone else, which explains all those strange deals with everybody and the comment from their (cfo?) about being backstopped by the gov.
If OpenAI fails, they could potentially bring in their downfall the major cloud providers who invested in hardware for them, expecting that it would pay-off over time.
Only Oracle went into debt to fund this expansion, and may well die. The rest of the cloud oriented mag7 used cash, can afford to write it off, and will continue being monopolies unimpeded.
I don't see a world where there is such a catastrophic failure, unless someone comes up with a significantly more efficient architecture.
We're barely scratching the surface of the utility of LLMs with today's models. They aren't more pervasive because of their costs today, but what happens if they drop another order of magnitude with the current capabilities?
'Potentially' but we are nowhere close to this. Hyperscalers print a _lot_ of money they can afford to lose. Even Nvidia wouldn't be in too much trouble yet. (The pure LLM companies are already toast, IMHO).
Hot take: a flagship silicon valley startup built on hype and overzealous ambition crashing and burning in 2026 is exactly what the industry needs right now.
Let's say OpenAI signed a commitment contract that they agreed to spend XXX USD in your company over N years. You invest in infrastructure, your contractors sometimes take loans, the construction companies take loans. Countries / Funds lend money to such companies (example: Saudi Arabia fund), these funds themselves raised debt, it can quickly spiral down.
If OpenAI fails, wouldn't their customers just move to other AI providers? So the total hardware demand by AI companies wouldn't dramatically change over a reasonable time frame?
If OpenAI fails, the assumed case is overcapitalisation relative to demand. In such a world, if the reason OpenAI has no liquidity is because there's insufficient demand, there's no customers to move across.
What happens if the OpenAI failure is because of lack of demand from their customers and the market in general? None of the AI providers or hardware providers will survive, no?
Similar to certain banks in 2007/2008, the idea would be “so much investment is tied to one company that if that company went bankrupt, it could have consequences for the broader economy”
The thing is, it is not 2007/2008 any more. The US government is holding record amounts of debt and countries around the world are now trying to become independent of it. This includes its bond markets on which the dollar relies upon to give it its reserve currency status, which in turn is what gives it its power to print money and bail industries out. If something happens that requires Big Tech to be bailed out and international bond holders decide the US is no longer reliable, it could very well end up triggering the collapse of the US dollar as the world's reserve currency and the downfall of the US as we know it.
The stock market will lose faith in AI companies, which will crash the stock price of Google, Microsoft, Oracle, Nvidia and CoreWeave. Investors will lose billions, many of those investors are pension funds. Any AI projects that aren't already profitable will shutdown.
And, because AI is currently what prevents the US economy from being in a recession (at least that what some people speculate), the US economy will stumble, which means that everyone else will to.
Is it unclear? Compared to other times a "too big to fail" industry failed?
If OpenAI crashes, for example funding stops, they go broke, fall behind, nobody buys anything, then all the money they invested for data centers or demand they created for NVIDIA chips and compute collapses. That creates surplus of hardware, causes lots of construction/buildout / stockup orders to get cancelled, and the whole thing ripples as suppliers and construction and data center providers etc etc suddenly lose a ton of anticipated profits.
Share prices drop as people dump to protect their portfolios, anticipating dips in the prices because share prices will drop as people dump to protect their portfolios (I'm not kidding).
Given that the big 7 AI companies are basically _all_ of the market growth lately, it doesn't even take a serious panic / paranoia episode to see the market itself stagnate or significantly regress, as people pull from anything AI related, and then pull from the market itself anticipating the market will fall.
That's different from saying that boeing is too big to fail for example. The US can't accept to lose its only major commercial aircraft manufacturer. Or Intel for similar reasons.
But what you're describing is about keeping the AI bubble from popping. Can a bubble really be too big too fail?
People crying about the revenue gap constantly forget that OpenAI still hasn't turned on the ads, porn, and gambling. Trust, they will turn it on eventually.
Why would OpenAI have any sort of advantage in these industries compared to the established players?
it's pretty sobering to think that the so-called harbingers of SkyNet AGI have to fall back to mafia-era revenue streams like vice to convince shareholders that their money wasn't wasted.
I am suspect that ads will be able to cover the cost of inference, much less model building, salaries, product development, etc. The ad industry is built upon _and priced for_ websites that return results for virtually zero cost. That isn't going to be true for rendering AI.
Define "attracted". I believe porn and gambling both do incredible amounts of damage to society. Ads probably do some damage to society, though I can recognize some benefits in being able to advertise and market products and services. I don't believe there is any benefit to the other two.
HSBC, the firm who did the analysis in the article, took into account projected revenues matching user numbers of 44% of the world's non-Chinese population.
44% for OpenAI wouldn't mean 0% for Google, but considering that even Apple only has about 30% of the global smartphone marketshare, OpenAI getting there seems unlikely.
> People crying about the revenue gap constantly forget that OpenAI still hasn't turned on the ads, porn, and gambling
But quite the opposite, HSBC assumed that they will have a virtual global monopoly on AI, and even under those projections they will still need to take on hundreds of billions of debt. I'm sure if they get there getting access to that debt will be easier than I'm assuming currently.
I think that a company with a global monopoly on a thinking machine that replaces wide swaths of current internet business categories, whose user base makes up most of the world outside of the GFW, would be very valuable on the public market and would probably have access to as much debt as they want, even if it had negative consequences for the world economy. Even the 2nd or 3rd place company in that race will be quite profitable. It's not impossible, just implausible.
They may not be shipping good enough products, but on the flip side it still feels like they have almost no competition outside of coding and image gen. The EOL termination of 4-o should be some evidence of this.
OTOH, why don't they ship good enough products? To me all of OpenAIs recent investments strongly suggest they hit a dead end with their current LLM approach. After all, if they knew the path ahead for GPT looks great, why don't they invest into training the next big thing instead of doing datacenters with the intention of renting them out?
Ok, probably true, but it’s still a pretty far fall from “we’re just about to deliver AGI which will put us in the driver seat of the entire global economy.” Which was the core value prop of their $trillion investment pitch. Like, there are way cheaper ways to deliver ads, porn, and gambling than training and operating huge LLMs.
I think you're right, however that only defeats themselves in the long run. AI is something that can be run locally, while search, shopping, movies, social, etc. cannot. And once the ads are baked into the product, you will need a force like the EU to remove them else shareholders will riot. Ads will be the perfect weapon to shoot themselves in the foot.
The cost per generation is still too expensive, they would not profit from ads and very few people pay for porn. I don't see how they'd profit from gambling either.
That is exactly the parent's point, that while those things could bring in revenue, the competition in those industries is already incredibly fierce. And that's before potential new entrants into the space on the AI front. Even in the best of cases I don't think anyone is gonna be able to pull anything other than middling margins.
I'm not saying they should or will give up on these markets, just that their margins aren't likely to be great.
There's already so much AI porn right now. There's no reason to believe that OpenAI will be any better or able to command higher margins than their competitors. They also
have the problem of being under a ton of public and regulatory scrutiny so the highest paying clientele, the people with niche unsavory fetishizes, won't be able
to be serviced by them.
And that's ignoring the fact that getting into this space at all is pure HN speculation and unlikely to ever happen anyway.
This is absolutely what techbro MBAs raised on case studies of Apple and Toyota are likely to think. Gamblers and porn users do not have any sort of brand loyalty and are likely to be hostile to the idea of "vertical integration" and "360 degree views of the customer".
(Imo) They will turn to all of these(especially to porn and gambling) when the core model of "enhance your life" will slowly fade away. The "academia space", the teens/boomers demographics, all of those will stop using OpenAI at scale if they're bombarded with vices (porn, gambling, etc).
Ads & referrals are already in the works, and people are generally tolerant of those. But, as with any company, appearances matter. ChatGPT will definitely lose users at the slightest possibility of having non-sanitized content served to more morally sensible groups.
But this is not a bank, or an airline, or a real estate giant.
If OpenAI goes bankrupt, what happens? People won’t be able to write their precious slop oh no and serious professionals will just switch to any other LLM provider
TL;DR - financial analysts look at current charts and project them foward by 5 years and go "wow the numbers look bad".
Sure OpenAI may well be bleeding money into the 2030s, or may even go bust completely depending on how pessimistic you are, but the analysis completely skips:
- They are building their own data centers, and will be less reliant on renting compute from Microsoft and Amazon over time.
- Once the AI bubble has subsided costs for GPU purchases and rentals will decrease significantly. Plus there will be more advancements and competition in the space (e.g. Google TPUs) and Nvidia will no longer be able to name their own price.
- We will write more efficient software for training and inference.
- Once user growth is tapped out OpenAI will no longer need to have the overly generous free tier that they do today. And if they decide to turn up the advertising faucet these users could bring in a ton more revenue than in the projection. Thinking that every AI company combined will capture only 2% of the total digital advertising market is ridiculous. AI apps are already challenging social media for scrolling time.
Basically, the entire space is evolving so rapidly that it's pointless to make a projection with the assumption that the landscape isn't going to change from here on.
Capitalism makes more sense if it's thought about purely in terms of individual self interest as supposed to something that leads to an efficient market on aggregate.
The CEOs making big calls across the economy have already negotiated golden parachutes in the event of their failure.
The financiers and lawyers getting a chunk of each bond deal they close have every incentive to raise more than what's actually needed. Investment funds flush with ZIRP dollars have every incentive to plow it back into investments to show that "the money is at work".
Boom and bust hype cycles have always been a feature of capitalism, especially for advanced technology. Perhaps the only way to know the limit is to overshoot.
Not just the economy. We usually deal with these busts and crisis collectively through the public sector as well.
There's also a hidden opportunity cost in regards to hype cycles. So much energy, attention and money flows into the hype, while other businesses or entire sectors get overlooked and underappreciated.
Except the 7 firms make up >30% of all market growth, and move more capital than entire countries. Yes, they are hemorrhaging cash with every new customer, and burdening taxpayers by eating community electrical and water infrastructure.
Not when the government steps in because OpenAI, NVIDIA, Microsoft, etc are all going down together. There is so much pseudo-money-laundering going on between those companies. I would be shocked if the government (e.g. 2008 style socialism for the rich) steps in to make them “too big to fail”.
It's so sad how much money leaders will effortlessly pump into something like this, when we still have existential threats of climate change, incurable diseases, poverty, housing, and so on.
Meanwhile ungodly amounts of money are being used so some boomer can generate a AI video of a baby riding a puppy.
https://archive.ph/PyLnT
It's sort of hard to judge this.
The article mostly focuses on ChatGPT uses, but hard to say if ChatGPT is going to be the main revenue driver. It could be! Also unclear if the underlying report is underconsidering the other products.
It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me. There will be challenges in capturing it and challenges with user trust, but it seems super promising because it will likely be harder to block and has a lot of intent context that should make it like search advertising++. And for context, search advertising is 40% of digital ad revenue.
Seems like the error bars have to be pretty big on these estimates.
IMO the key problem that OpenAI have is that they are all-in on AGI. Unlike a Google, they don't have anything else of any value. If AGI is not possible, or is at least not in reach within the next decade or so, OpenAI will have a product in the form of AI models that have basically zero moat. They will be Netscape in a world where Microsoft is giving away Internet Explorer for free.
Meanwhile, Google would be perfectly fine. They can just integrate whatever improvements the actually existing AI models offer into their other products.
"don't have anything else of any value. " ?
OpenAI is still de facto the market leader in terms of selling tokens.
"zero moat" - it's a big enough moat that only maybe four companies in the world have that level of capability, they have the strongest global brand awareness and direct user base, they have some tooling and integrations which are relatively unique etc..
'Cloud' is a bigger business than AI at least today, and what is 'AWS moat'? When AWS started out, they had 0 reach into Enterprise while Google and AWS had infinity capital and integration with business and they still lost.
There's a lot of talk of this tech as though it's a commodity, it really isn't.
The evidence is in the context of the article aka this is an extraordinary expensive market to compete in. Their lack of deep pockets may be the problem, less so than everything else.
This should be an existential concern for AI market as a whole, much like Oil companies before highway project buildout as the only entities able to afford to build toll roads. Did we want Exxon owning all of the Highways 'because free market'?
Even more than Chips, the costs are energy and other issues, for which Chinese government has a national strategy which is absolutely already impacting the AI market. If they're able to build out 10x data centres at offer 1/10th the price at least for all the non-Frontier LLM, and some right at the Frontier, well, that would be bad in the geopolitical sense.
I think you're measuring the moat of developing the first LLMs but the moat to care about is what it'll take to clone the final profit generating product. Sometimes the OG tech leader is also the long term winner, many times they are not. Until you know what the actual giant profit generator is (e.g. for Google it was ads) then it's not really possible to say how much of a moat will be kept around it. Right now, the giant profit generator is not seeming to be the number of tokens generated itself - that is really coming at a massive loss.
Selling tokens at a massive loss, burning billions a quarter isn't the win you think it is. They don't have a moat bc they literally just lost the lead, you only can have a moat when you are the dominant market leader which they never were in the first place.
All indications are that selling tokens is a profitable activity for all of the AI companies - at least in terms of compute.
OpenAI loses money on free users and paying the absurdly high salaries that they've chosen to offer.
The AWS moat is a web of bespoke product lock-in and exorbitant egress fees. Switching cloud providers can be a huge hassle if you didn't architect your whole system to be as vendor-agnostic as possible.
If OpenAI eliminated their free tier today, how many customers would actually stick around instead is going to Google's free AI? It's way easier to swap out a model. I use multiple models every day until the free frontier tokens run out, then I switch.
That said, idk why Claude seems to be the only one that does decent agents, but that's not exactly a moat; it's just product superiority. Google and OAI offer the same exact product (albeit at a slightly lower level of quality) and switching is effortless.
I mean, on your Cloud point I think AWS' moat might arguably be a set of deep integrations between services, and friendly API's that allow developers to quickly integrate and iterate.
If AWS' was still just EC2, and S3 then I would argue they had very little moat indeed.
Now, when it comes to Generative AI models, we will need to see where the dust settles. But open-weight alternatives have shown that you can get a decent level of performance on consumer grade hardware.
Training AI is absolutely a task that needs deep pockets, and heavy scale. If we settle into a world where improvements are iterative, the tooling is largely interoperable... Then OpenAI are going to have to start finding ways of making money that are not providing API access to a model. They will have to build a moat. And that moat may well be a deep set of integrations, and an ecosystem that makes moving away hard, as it arguably is with the cloud.
EC2 and S3 moat comes from extreme economies of scale. Only Google and Microsoft can compete. You would never be able to achieve S3 profitability because you are not going to get same hardware deals, same peering agreements, same data center optimization advantages. On top of that there is extremely optimized software stack (S3 runs at ~98% utilization, capacity deployed just couple weeks in advance, i.e. if they don’t install new storage, they will run out of capacity in a month).
I wouldn't call it a moat. A moat is more about switching costs rather than quality differentiation. You have a moat when your customers don't want to switch to a competitor despite that competitor having a superior product at a better price.
I've also thought of this and what's more, Google's platform provides them with training from YouTube, optimal backend access to the Google Search index for grounding from an engine they've honed for decades, training from their smartphones, smart home devices and TV's, Google Cloud... And as you say, also the reverse; empowering their services from said AI, too.
They can also run AI as a loss leader like with Antigravity.
Meanwhile, OpenAI looks like they're fumbling with that immediately controversial statement about allowing NSFW after adult verification, and that strange AI social network which mostly led to Sora memes outside of it.
I think they're going to need to do better. As for coding tools, Anthropic is an ever stronger contender there, if they weren't pressured from Google already.
The moat for any frontier LLM developer will be access to proprietary training data. OpenAI is spending some of their cash to license exclusive rights to third party data, and also hiring human experts in certain fields just to create more internal training data. Of course their competitors are also doing the same. We may end up in a situation where each LLM ends up superior in some domains and inferior in others depending on access to high quality training data.
"Needs cash" is not a moat.
> they are all-in on AGI
What are you basing this on? None of their investor-oriented marketing says this.
https://openai.com/charter/
> OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome.
Note that it doesn't say: "Our mission is to maximize shareholder value, and we develop AI systems to do that".
In fairness, no company’s mission statement says “maximize shareholder value” because it doesn’t need to be said - it’s implicit. But I agree that AGI is at the forefront of OpenAI’s mission in a way it isn’t for Google - the nonprofit roots are not gone.
The opening lines of their mission statement is direct about this:
"OpenAI is an AI research and deployment company. Our mission is to ensure that artificial general intelligence benefits all of humanity."
and
"We are building safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome."
https://openai.com/about/
I don't know what the moneyed insiders think OpenAI is about, but Sam Altman's public facing thoughts (which I consider to be marketing) are definitely oriented toward making it look like they are all-in on AGI:
See:
(1) https://blog.samaltman.com/the-gentle-singularity (June, 2025) - "We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence, and at least so far it’s much less weird than it seems like it should be."
- " It’s hard to even imagine today what we will have discovered by 2035; maybe we will go from solving high-energy physics one year to beginning space colonization the next year; or from a major materials science breakthrough one year to true high-bandwidth brain-computer interfaces the next year."
(2) https://blog.samaltman.com/three-observations (Feb, 2025) - "Our mission is to ensure that AGI (Artificial General Intelligence) benefits all of humanity."
- "In a decade, perhaps everyone on earth will be capable of accomplishing more than the most impactful person can today."
(3) https://blog.samaltman.com/reflections (Jan, 2025) - "We started OpenAI almost nine years ago because we believed that AGI was possible, and that it could be the most impactful technology in human history"
- "We are now confident we know how to build AGI as we have traditionally understood it."
(4) https://ia.samaltman.com/ (Sep, 2024) - "This may turn out to be the most consequential fact about all of history so far. It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there."
(5) https://blog.samaltman.com/the-merge (Dec, 2017) - "A popular topic in Silicon Valley is talking about what year humans and machines will merge (or, if not, what year humans will get surpassed by rapidly improving AI or a genetically enhanced species). Most guesses seem to be between 2025 and 2075."
(I omitted about as many essays. The hype is strong in this one.)
Not only this, but there is a compounded bet that it’ll be OpenAI that cracks AGI and not another lab, particularly Google from which LLMs come in the first place. What makes OpenAI researchers so special at this point?
This is how I look at Meta as well. Despite how much it is hated on here fb/ig/whatsapp aren’t dying.
AI not getting much better from here is probably in their best interest even.
It’s just good enough to create the slop their users love to post and engage with. The tools for advertisers are pretty good and just need better products around current models.
And without new training costs “everyone” says inference is profitable now, so they can keep all the slopgen tools around for users after the bubble.
Right now the media is riding the wave of TPUs they for some reason didn’t know existed last week. But Google and meta have the most to gain from AI not having any more massive leaps towards agi.
> They can just integrate whatever improvements the actually existing AI models offer into their other products.
If this is what users actually want.
Yes, as is implied by the word "improvements"
> IMO the key problem that OpenAI have is that they are all-in on AGI
I think this needs to be said again.
Also, not only do we not know if AGI is possible, but generally speaking, it doesn't bring much value if it is.
At that point we're talking about up-ending 10,000 years of human society and economics, assuming that the AGI doesn't decide humans are too dangerous to keep around and have the ability to wipe us out.
If I'm a worker or business owner, I don't need AGI. I need something that gets x task done with a y increase in efficiency. Most models today can do that provided the right training for the person using the model.
The SV obsession with AGI is more of a self-important Frankenstein-meets-Pascal's Wager proposition than it is a value proposition. It needs to end.
Also, they'll have garbage because the curve is sinusoidal and not anything else. Regardless of the moat, the models won't be powerful enough to do a significant amount of work.
[dead]
They're both all in on being a starting point to the Internet. Painting with a broad brush that was Facebook or Google Search. Now it's Facebook, Google Search, and ChatGPT.
There is absolutely a moat. OpenAI is going to have a staggering amount of data on its users. People tell ChatGPT everything and it probably won't be limited to what people directly tell ChatGPT.
I think the future is something like how everyone built their website with Google Analytics. Everyone will use OpenAI because they will have a ton of context on their users that will make your chatbot better. It's a self perpetuating cycle because OpenAI will have the users to refine their product against.
yeah but your argument is true for every llm provider. so i don't see how it's a moat since everyone who can raise money to offer an llm can do the same thing. and google and microsoft doesn't need to find llm revenue it can always offer it at a loss if it chooses unless it's other revenue streams suddenly evaporate. and tbh i kind of doubt personalization is as deep of a moat as you think it is.
Everyone could raise and build a search engine or social network. Many did and none of them dethroned Google or Facebook.
Social networks have network effects, the value comes from other people on the network, not the platform itself.
Google can offer their services for free for a lot longer than OpenAI can, and already does to students. DeepSeek offers their competitor product to ChatGPT for free to everyone already.
> It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me.
I'm not super bullish on "AI" in general (despite, or maybe because of working in this space the last few years), but strongly agree that the advertising revenue that LLM providers will capture can be potentially huge.
Even if LLMs never deliver on their big technical promises, I know so many casual users of LLMs that basically have replaced their own thought process with "AI". But this is an insane opportunity for marketing/advertising that stands to be a much of a sea change in the space as Google was (if not more so).
People trust LLMs with tons of personal information, and then also trust it to advise them. Give this behavior a few more years to continue to normalize and product recommendations from AI will be as trusted as those from a close friends. This is the holy grail of marketing.
I was having dinner with some friends and one asked "Why doesn't Claude link to Amazon when recommending a book? Couldn't they make a ton in affiliate links?" My response was that I suspect Anthropic would rather pass on that easy revenue to build trust so that one day they can recommend and sell the book to you.
And, because everything about LLMs is closed and private, I suspect we won't even know when this is happening. There's a world where you ask an LLM for a recipe, it provides all the ingredients for your meal from paid sponsors, then schedules to have them delivered to your door bypassing Amazon all together.
All of this can be achieved with just adding layers on to what AI already is today.
> It also estimates that LLM companies will capture 2% of the digital advertising market, which seems kind of low to me. There will be challenges in capturing it and challenges with user trust, but it seems super promising because it will likely be harder to block and has a lot of intent context that should make it like search advertising++. And for context, search advertising is 40% of digital ad revenue.
Yeah, I don't like that estimate. It's either way too low, or much too high. Like, I've seen no sign of OpenAI building an ads team or product, which they'd need to do soon if it's going to contribute meaningful revenue by 2030.
> Like, I've seen no sign of OpenAI building an ads team or product
You just haven't been paying attention. They hired Fidji Simo to lead applications in may, she led monetization/ads at facebook for a decade and have been staffing up aggressively with pros.
Reading between the lines in interview with wired last week[0], they're about to go all in with ads across the board, not just the free version. Start with free, expand everywhere. The monetization opportunities in chatgpt are going to make what google offers with adwords look quaint, and every CMO/performance marketer is going to go in head first. 2% is tiny IMO.
[0] - https://archive.is/n4DxY
https://openai.com/careers/growth-paid-marketing-platform-en...
Is that role not exactly what you mention?
At least the description is not at all about building an adtech platform inside OpenAI, it's about optimizing their marketing spend (which being a big brand, makes sense).
There are a bunch of people from FB at OpenAI, so they could staff an adtech team internally I think, but I also think they might not be looking at ads yet, with having "higher" ambitions (at least not the typical ads machine ala FB/Google). Also if they really needed to monetize, I bet they could wire up Meta ads platform to buy on ChatGPT, saving themselves a decade of building a solid buying platform for marketers.
> There are a bunch of people from FB at OpenAI, so they could staff an adtech team internally I think
Well they have Fidji, so she could definitely recruit enough people to make it work.
> with having "higher" ambitions (at least not the typical ads machine ala FB/Google)
Everyone has higher ambitions till the bills come due. Instagram was once going to only have thoughtfully artisan brand content and now it's just DR (like every other place on the Internet).
> At least the description is not at all about building an adtech platform inside OpenAI, it's about optimizing their marketing spend (which being a big brand, makes sense).
The job description has both, suggesting that they're hedging their bets. They want someone to build attribution systems which is both wildly, wildly ambitious and not necessary unless they want to sell ads.
> I bet they could wire up Meta ads platform to buy on ChatGPT, saving themselves a decade of building a solid buying platform for marketers.
Wouldn't work. The Meta ads system is so tuned for feed based ranking that I suspect they wouldn't gain much from this approach.
> it’s not at all about building an affect platform inside OpenAI.
Directly from posting: “building the technical infrastructure behind OpenAI’s paid marketing platform”
no. this role is for running ads campaigns at scale (on google, meta, etc) to grow openai users. its at a large enough scale it's called "platform" but it would be internal use only.
> Your role will include projects such as developing campaign management tools, integrating with major ad platforms, building real-time attribution and reporting pipelines, and enabling experimentation frameworks to optimize our objectives.
Actually yes (I did mean to check again but I hadn't seen evidence of this before).
I do think that this seems odd, looks like they're hiring an IC to build some of this stuff, which seems odd as I would have expected them to be hiring multiple teams.
That being said, the earliest they could start making decent money from this is 2028, and if we don't see them hire a real sales team by next March then it's more likely to be 2030 or so.
That's a marketing role, not a product role.
https://chatgpt.com/merchants/
Thanks for calling this out. Here is a better comparison. Before Google was founded, the market for online search advertising was negligible. But the global market for all advertising media spend was on the order of 400B (NYT 1998). Today, Google's advertising revenue is around 260B / year or about 60% of the entire global advertising spend circa 1998.
If you think of openAI like a new google, as in a new category-defining primary channel for consumers to search and discover products. Well, 2% does seem pretty low.
>Today, Google's advertising revenue is around 260B / year or about 60% of the entire global advertising spend circa 1998.
Or about 30% of the global advertising spend circa 2024.
I wonder if there is an upper bound on what portion of the economy can be advertising. At some point it must become saturated. People can only consume so much marketing.
But that occurred with a new form of media that people now use in more of their time than back before Google. It implies AI is growth in time spent. I think the trend is more likely that AI will replace other media.
There's also a possible scenario where the online ads market around search engines gets completely disrupted and the only remaining avenues for ad spending are around content delivery systems (social media, youtube, streaming, webpages, etc.). All other discovery happens within chatbots and they just get a revenue share whenever a chatbot refers a user to a particular product. I think ChatGPT is soon going to roll out this feature where you can do walmart shopping without leaving the chat.
Your revenue share concept sounds passive. I suspect advertisers will also be able to pay for placement.
Shopping within Alexa never made sense. I'm not sure I'll want to do it via ChatGPT.
Maybe they're thinking they can build a universal store with search over every store? Like a "Google Shopping" type experience?
Google, Meta and Microsoft have AI search as well, so OAI with no ad product or real time bidding platform isn't going to just walk in and take their market.
2% is optimistic in my opinion.
Google, Meta and Microsoft would have to compete on demand, i.e. users of the chat product. Not saying they won't manage, but I don't think the competition is about ad tech infrastructure as much as it is about eyeballs.
It might take Microsoft's Bing share, but Google and Meta pioneered the application of slot machine variable-reward mechanics to Facebook, Instagram and Youtube, so it would take a lot more than competing on demand to challenge them.
> And for context, search advertising is 40% of digital ad revenue.
But all the search companies have their own AI so how would OAI make money in this sector?
Several ways, although I'm not sure whether the below will happen:
1. Paid ads - ChatGPT could offer paid listings at the top of its answers, just like Google does when it provides a results page. Not all people will necessarily leave Google/Gemini for future search queries, but some of the money that used to go to Google/Bing could now go to OpenAI.
2. Behavioral targeting based on past ChatGPT queries. If you have been asking about headache remedies, you might see ads for painkillers - both within ChatGPT and as display ads across the web.
3. Affiliate / commission revenue - if you've asked for product recommendations, at least some might be affiliate links.
The revenue from the above likely wouldn't cover all costs based on their current expenditure. But it would help a bit - particularly for monetizing free users.
Plus, I'm sure there will be new advertising models that emerge in time. If an advertiser could say "I can offer $30 per new customer" and let AI figure out how to get them and send a bill, that's very different to someone setting up an ad campaign - which involves everything from audience selection and creative, to bid management and conversion rate optimization.
So I don't necessarily disagree with your suggestions, but that is just not a $1T company you're describing. That's basically a X/Twitter size company, and most agree that $44B was overpaying.
It's not that OpenAI hasn't created something impressive, it just came at to high a price. We're talking space program money, but without all the neat technologies that came along as a result. OpenAI more or less develop ONE technology, no related product or technologies are spun out of the program. To top it all off, the thing they built, apparently not that hard to replicate.
You have to take into consideration the source. FT is part of the Anthropic circle of media outlets and financial ties. It benefits them to create a draft of support to OpenAI competition, primarily Anthropic, but they(FT) also have deep ties to Google and the adtech regime.
They benefit from slowing and attacking OpenAI because there's no clear purpose for these centralized media platforms except as feeds for AI, and even then, social media and independents are higher quality sources and filters. Independents are often making more money doing their own journalism directly than the 9 to 5 office drones the big outlets are running. Print media has been on the decline for almost 3 decades now, and AI is just the latest asteroid impact, so they're desperate to stay relevant and profitable.
They're not dead yet, and they're using lawsuits and backroom deals to insert themselves into the ecosystem wherever they can.
This stuff boils down to heavily biased industry propaganda, subtly propping up their allies, overtly bashing and degrading their opponents. Maybe this will be the decade the old media institutions finally wither up and die. New media already captures more than 90% of the available attention in the market. There will be one last feeding frenzy as they bilk the boomers as hard as possible, but boomers are on their last hurrah, and they'll be the last generation for whom TV ads are meaningfully relevant.
Newspapers, broadcast TV, and radio are dead, long live the media. I, for one, welcome our new AI overlords.
All of which is great theory without any kind of evidence? Whereas the evidence pretty clearly shows OpenAI is losing tons of money and the revenue is not on track to recover it?
Tapping into AdTech is extremely hard, as it's hard driven by network effects. What you mean is "displaying ads inside OpenAI products" then, yes, achievable, but that's a miniscule part of targeted Ad markets - 2% is actually very optimistic. Otherwise, they can sell literally 0 products to existing players as they all have already established "AI" toolsets to help them for ad generation and targeting.
Query: LibraGPT, create a plan for my trip to Italia
Response: Book a car at <totally not an ad> and it will be waiting for you at arrival terminal, drive to Napoli and stay at <totally not an ad> with an amazing view. There's an amazing <totally not an ad> place that serves grandma's favorite carbonara! Do you want me to make the bookings with a totally not fake 20% discount?
I'm traveling like this all the time already, I don't understand why it's hard for people to understand that ad placement is actually easier for chat than search
> all the time already
What are you actually saying? You're already using chatbots that are embedding non-disclosed paid endorsements? And you like that?
> ad placement is actually easier for chat
Can you point to, I don't know, anything to back this up?
> that ad placement is actually easier for chat than search
Yes, but the reason why people are turning to chatgpt is because the time to actual info that _I want_ is much much lower.
The point of advertising is to displace the thing that you actually want with something they are paying the company to promote.
You can handwave about personalization, but do you want adtech people having access to your life's context?
But who wants that? And you're going to say that's exactly what a travel agent does, selling me stuff so he can get a kickback. But when stuff goes wrong, I'll yell at the travel agent so he has some incentive to curate his ads.
How to speedrun massive penalties and disgorgement from FTC.
I guess we'll just put that in the "Cost of Goods Sold" bucket.
I'm not aware of any FTC rule that would preempt this sort of product as long as it met the endorsement disclosure rules (16 CFR Part 255), same as paid influencers do today.
What are you imagining they run afoul of?
https://www.ftc.gov/business-guidance/resources/ftcs-endorse...
https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B...
friendzis's example showed a plausible way to generate revenue by inserting paid placements into the chat bot response without disclosures by pretending they are just honest, organic suggestions.
Right. That's not a novel idea, and this is a well-trod area of concern. That's why these FTC rules have been around for many years.
edit: to be clear, I am saying that in the absence of clear disclosures, that would run afoul of current FTC rules. And historically they have been quick to react to novel ways of misleading consumers.
bribe the artificial idiot to con the user, brilliant !
[dead]
That seems a bit risky for when the car isn't waiting for you at the terminal.
At least with an ad it's obvious a separate company is involved. If you do all the payment through OpenAI it seems to leave them open to liability.
> If you do all the payment through OpenAI it seems to leave them open to liability.
Booking, airbnb, rentalcars, etc all seem to be doing pretty fine regulatory wise.
Could be as simple as referral link commission like all those totally-not-content-farm travel blogs.
Travel sites, VPNs and insurance all pay quite handsomely (compared to say amazon links on cooking sites)
If ChatGPT shows ads, I'll switch to Claude or Gemini or DeepSeek.
I expect all hosted model providers will serve ads (regular, marked advertisements, no need for them to pretend not to, people don't care) once the first provider takes the lid off on the whole thing and it proves to be making money. There's no point in differentiating as the one hosted model with no ads because it only attracts a few people, the same way existing Google search and YouTube alternatives that respect the user are niche. Only offline, self hosted models will be ad free (in my estimation).
Assuming you know it's an ad. Ads in answers will generate a ton of revenue and you'll never know if that Hilton really is the best hotel or if they just paid the most.
This isn't a realistic concern unless FTC rules changed substantially from where they are today (see my other comment on this post for links). Sponsored link disclosures would be in place.
Everything else aside, it's simply not worth it for them to try to skirt these rules because the majority of their users (or Google's) simply don't care if something is paid placement or not, provided it meets their needs.
That's only true if you can demonstrate a substantial percentage of people would be unaware of it. The reason influencers have to disclose is some but not all take endorsement money. It would be pretty easy for OpenAI to say it was common knowledge or to bury disclosures in the fine print of the services and not every time it happened
The US federal government is now a mob-style organization. The laws, rules, and regulations that are written down are only applicable as far as Trump and those around him want them to be. Loyalty to the boss is the only inviolable rule.
In other words, if they want to put ads into chat, they just need to be perceived as well aligned to Trump to avoid any actual punishment.
They're all already trained on ads, and it would be silly to think advertisers aren't going to optimize for this.
The OpenAI pitch for “publishing partnerships” (basically buying bias and placement) leaked last year.
Do you think it would show ads, or just prioritize content based on who has paid for ad placement?
Seems like ad targeting might be a tough sell here though, it’d basically have to be “trust me bro”. Like - I want to advertise coca-cola when people ask about terraforming deserts? I think I wouldn’t be either surprised by amazing success or terrifying failure.
Perplexity actually did search with references linked to websites they could relate in a graph and even that only made them like $27k.
I think the problem is on Facebook and Google you can build an actual graph because content is a thing (a url, video link etc). It will be much harder to I think convert my philosophical musings into active insights.
So few people understand how advertising on the internet works and that is I guess why Google and Meta basically print money.
Even here the idea that it’s as simple as “just sell ads” is utterly laughable and yet it’s literally the mechanism by which most of the internet operates.
FT is really losing it. Used to be reliable with quality takes. Now mostly following in line with the spectating takers.
In what sense? They are asking the questions that investment managers would be asking, like: "where the fuck is your revenue going to come from"
and "200 billion, when your revenue is 12, is the market you are targeting actually big enough to support that"
It is to the point of yellow journalism. They know that the "OpenAI is going to go belly up in a week!" take is going to be popular with AI skeptics, which includes a large number of HN viewers. This thread shot up to the top of the front page almost immediately. All of that adds to the chances of roping in more subscribers.
[dead]
It's FT Alphaville, which means it's only supposed to be a blog-style comment on the HSBC report
FT (like all media) does not sell news, they sell ads.
I think people are severely underestimating ChatGPT as a way to make money other than subscriptions. I also think people are underestimating the branding power ChatGPT has already. All my friends have ChatGPT on their phone. None of them except me has Gemini or Claude app.
This doesn't account for OpenAI's other ambitions such as Sora app.
Hey Sam Altman or OpenAI employee, if you are reading this, I think you should buy the North American version of TikTok if the opportunity presents itself. The future of short videos will be heavily AI generated/assisted. Combine Tiktok's audience with your Sora tools and ChatGPT data and you got yourself a true Instagram competitor immediately. If the $14b sales price of US Tiktok is real, that's an absolute bargain in the grand scheme of things.
> Meta has 3.5 billion users and projected ~$200b revenue in 2025
Meta makes about $200B on ads, Google makes about $235B on ads. Advertising is roughly 1.5% of the total GDP of the US and hasn't changed in 20+ years. So what you have is a big ass pie with a few players fighting for it that barely grows every year.
OpenAI has to somehow:
1. Compete directly with Google Gemini and Meta's Llama for a piece of users pie with a product that has very little differentiator (functionality and technically speaking).
2. Have to prove to advertisers that their single dollar ad purchase on OpenAI is categorically worth more than any other channel.
3. Have enough forward capital to continue purchasing capital-intense hardware purchases.
4. Having enough capital to weather any potential economic headwinds.
I know where my bet is.
OpenAI has branding power, a clear product focused mindset, focused attention, and moves faster than Google and Meta. Nearly 1 billion users in 3 years is no joke.
We're on Hacker News. Y Combinator literally teaches their companies that they can beat incumbents using focus and speed.
My bet is on OpenAI. When they IPO, I can easily see them with $1 trillion in valuation and raise the a record amount of money in an IPO.
If Meta and Google don't see OpenAI and LLMs as an existential threat, they wouldn't invest so much. I think AI has that potential to completely disrupt Google and Meta because it fundamentally changes the way people behave. It's a paradigm shift. It isn't just playing the same game.
You’re betting on OpenAI a company that has literally never sold an online ad against the two kings of online advertising…cool.
They have to compete against entities who really know what they are doing.
> Meta has 3.5 billion users and projected ~$200b revenue in 2025.
Meta has WhatsApp, Instagram, and Facebook to account for that.
OpenAI has ChatGPT (not a social platform).
It seems to me you're comparing apples and oranges here.
Anyways, check this out: https://openai.com/index/group-chats-in-chatgpt/
I don't think so. 1 billion users and a clear intention to deliver ads with an immense amount of data on users. That's a clear threat to both Meta and Google.PS. That's why Meta and Google are all in on AI. OpenAI is an existential threat to both in my humble opinion.
> You didn't state reasons why not being a social platform matters here.
There's nothing to pointlessly waste your time on. You open it to do a thing, you either do the thing or get frustrated or leave. Social networks are designed to waste your time even when they outlive their usefulness, therefore they can serve you more ads.
You could argue Google is the same as ChatGPT in that regard, but that's why Google has Adsense in almost any search result you click on.
As for your group chats feature argument, anyone can make a social network, that's the easy part. Getting friend groups to switch is the more difficult part.
> PS. That's why Meta and Google are all in on AI. OpenAI is an existential threat to both in my humble opinion.
They're all in on AI because that's what their investors want them to do to "not be left behind". Meta was all in Metaverse. And on a cryptocurrency before that (Diem). And on Free Basics before that. The fact that none of those succeeded didn't hurt them at all precisely because they had an infinite money glitch known as ads.
They can afford to waste amounts of money equivalent to a yearly budget of a small country, ChatGPT can't.
> You didn't state reasons why not being a social platform matters here.
The network effects matter so much more for a social platform than a chat bot. The switching costs for a user are much lower, so users can move to a different one much easier.
How sticky will chat bots prove to be in the long term? Will OpenAI be able to maintain a lead in the space in the long term, the way Google was over Bing? It's possible, but it's also pretty easy to imagine other providers being competitive and a landscape where users move between different LLMs more fluidly
Another reason why social media matters is that people actually spend their free time in all those feeds. On the other hand, using LLMs is much more oriented towards specific utilities. Having 1B users who visit you once a day to ask for an email proofread will not make you profitable.
But, like search, it captures intent to buy much better, while looking at feeds for internatinment does not. Adwords worked because of that, they could capture ad revenue on queries that lead to sales. The amount of extra context in AI chat is even better, as is the ability to steer the conversation to different options.
No one spends their free time on Google search either. Didn't stop Google from being even more profitable than Meta.
I think Google is the much bigger threat. I've more or elss stopped using ChatGPT now, it's easier to just type the question directly into Google and get the response from their AI rather than navigating to chatgpt first. Anecdotal but I don't see anything long term keeping people on that site.
I think the play here for OpenAI is they will eventually acquire reddit and that will be their first intro into a social platform.
Then they will have a social platform that they will continue to use to mine AI training data from + a source of ad revenue.
> It seems to me you're comparing apples and oranges here.
apart from those oranges have ~100bn a year to spend on rnd and still make a profit, where as openai doesn't
So yes, it is apples to oranges. but its reality.
Google makes over a billion of its ad Revenue from search. Intent works.
But I think Open AI is not a slam dunk for Ads. Gemini and AI mode will compete for the same budget, and Google's Ad machine is polished.
I think eventually you will buy Ads for Open AI in Google's marketing platforms, just like most people buy bing ads in Google.
OpenAI knows my intent better than Google.
I'm telling it nearly everything from my work problems to health problems to love life problems to product research, traveling plans, etc.
Problem is the day Google puts a similarly priced and more powerful chatbot similar to ChatGPT and powered by Gemini most users will switch without a second thought, and Google will get all that data about their personal life too. OpenAI having a free tier with powerful models is unsustainable, brand loyalty is a joke at this stage, people will not bother going to ChatGPT if the search bar in their browser starts a discussion with Google. I’m sure big players like Google know OpenAI is a temporary scheme to get free money, they are surfing along now but they are playing the long game.
That's different than intent.
Your intent is the immediate need: how do I fix this leaky faucet?
Your user profile (love life problems) is generally not useful there.
Google and Meta know your intent without you even telling them...
You do not understand how online ads work. Please just stop.
Maybe, but OpenAI only get the data you decide to type into ChatGPT in a vibe therapy session or whatever, and they have to burn a load of GPU time to keep you feeding it more slop.
Google just passively collects email and browsing history, much better data for targeting ads and way less cost to run.
I wonder if meta is a poor comparison for advertising because they're users tend to spend more time on their products doom scrolling, as opposed to something like google, where you get your answer right away and move on.
ChatGPT is a hybrid between Google and Meta. People use it for product and service search and research. People also use it as a companion - especially young people.
It's becoming social as well: https://openai.com/index/group-chats-in-chatgpt/
I can reset my advertisement profile by creating a new account on ChatGPT, for Meta platforms (Facebook, Instagram and WhatsApp) this is not the case.
>I think you should buy the North American version of TikTok if the opportunity presents itself. The future of short videos will be heavily AI generated/assisted.
I will have whatever you're smoking. If a social media platform literally proves the dead internet theory, it's not making any money.
> short videos will be heavily AI generated/assisted.
Low effort input. Low effort consumption. What a depressing vision of the suture. This is why I don't use social media.
I actually believe we are speed running into this. I have seen way too many people watching AI generated videos and scrolling through them.
For context, this is close to all venture capital spending in the US in 2024 ($215B): https://nvca.org/press_releases/nvca-releases-2025-yearbook-...
I just realized that other industries are way larger than AI. Assuming they capture the entire advertising market, only $390 Billion was spent in the US last year. Compare that to health care, where 4.3 Trillion was spent in the US last year, or commercial banking's revenue of 1.5 Trillion, commercial real estate's 1.5 Trillion, gasoline stations' 1.10 Trillion, etc. What's amazing is, despite the fact that AI isn't making much money, is taking on considerable debt, and isn't even assured to be all that useful, one third of the stock market is now just AI crap. The economy is going to collapse because of a small, brand-new industry. This... shouldn't be possible.
I know right it's like people are buying future value rather than based on current revenues or something.
How long did Apple keep going up following the smartphone revolution?
Build a system which can read MRI images, now AI took over the whole MRI Analysis sector.
Let AI diagnose the avg cold, now AI took over the household / local doctor for 90% of cases.
Use AI for support, now AI took over the call-center business.
AI can already code. It might not be perfect, it might not be always good but NO ONE assumed that 2025 some matrix multiplication can mimick another human being so well that you can write with it and it will produce working code at all.
thats the hype, thats the market of ai.
And in parallel we get robotic too. Only possible because of this AI thing. Because robotic with ML is so much better than whatever we had before. Now you can talk to a robot, the robot can move, the robot can plan actions. All of these robots use some type of ML in the background. Segment Anything is possible because of AI.
Thats the reason why this 'AI crap' is so hyped.
I think they are missing a few very significant revenue sources that likely going to grow beyond just simple enterprise user accounts for Chat GPT.
I'll call out two of them.
Image, video, and other content generation is going to become more important and companies will be spending on that. We've seen some impressive improvements there. IMHO near term a lot of that stuff might start showing up in advertising and news content, the whole media industry is going to be a massive consumer of this stuff. And there's going to be a lot of competition for the really high quality models that can be run at scale. Five years until 2030 is a lot of time for some pretty serious improvements to land.
Another area that the article skips over is agentic tools. Those are showing a lot of promise right now. Agentic coding tools are just the tip of the iceberg here. A lot of these tools are going to be using APIs. So API revenue is a source of revenue. There are applications across the entire IT industry. SAAS, legacy software, productivity tools, etc.
Yes 207B is a lot of money. And there's no guarantee that OpenAI comes out on top "winning" all these markets of course and we can argue about how big these markets will be. But OpenAI does have a good starting position and some street credibility here. It's a big bet on revenue and potential here. But so is betting against all that and dismissing things. And there's a lot of middle ground here.
> Image, video, and other content generation is going to become more important and companies will be spending on that. We've seen some impressive improvements there.
They are assuming that people _want_ automated content enough to pay for it.
Like at the moment its great because its essentially free, but paying >200billion?
Are we assuming that we can automate the equivalent of instagram, netflix/disney and whatsapp and make advertising revenue?
I think the unmentioned pivot is into robotics, which seems to have actual tangible value (ie automation of things that are hard to do now)
but still 200billion in r&d
Everything professionally produced you currently see uses a lot of CGI, and you can’t even tell.
Essentially what they are doing is cheaper, more accessible CGI. And in the same way at it is now, you are not going to be able to tell it was used in expensive productions and will be able to see it in the cheap productions.
Actually, I sense a mounting "AI fatigue". Angry comments under AI contents. Social media accounts that proclaim to be "AI Free". I was thinking today that the killer app for AI could be a filter that automatically exclude all AI content.
I agree. Back in the 90's there was plenty of skepticism, along with some mockery, surrounding the internet. But you never saw headlines like "Lawyer caught using the internet" or "Artist busted using the World Wide Web".
With AI, it's different.
There's a huge vegan market segment but meat industry still plods along
[dead]
Google are experimenting with different filters that intentionally blur the lines between real and fake video in their shorts. A creator I follow who makes human-in-shot videos suddenly looked fake, like some animation filter was applied. I commented on how bad it looked, and someone said YT are doing it to people's videos as some sort of "trial". The effects were to make the human look more AI generated. The only use for this I can surmise is to make AI content even harder to detect.
It is very obviously there. People are not simply accepting the AI-bro justification that one day they won't notice the difference so it's all fine.
Ordinary people are understandably either exhausted by or angry at AI content, because it is relentless and misleading, and because they know instinctively that the people who want this to succeed also want to destroy employment and concentrate profits in the hands of a few Silicon Valley sociopaths.
> They are assuming that people _want_ automated content enough to pay for it.
I’m assuming gen ai will supplement so instead of ten artists producing content you’ll have three orchestrating and correcting but the output will be indistinguishable from what it was before. Just now the margin is much higher.
I never think of direct to consumer sales in this context.
Anecdotally I've seen lots of anti AI posts on instagram from gen-z accounts, and on YouTube ive seen "ai not used to make this music" in the description.
Also anecdotal, but I do think there's gen z anti-ai sentiment, these kids are joining a world where they will never own a house and have bleak prospects and now even art is something they can't do if AI takes off, so the market might not be there
Seems like older people love AI art, people over 50, which is a sizeable market don't get me wrong, but in 10-15 years idk
I would assume that generative AI will be able to make endless hours of personalized reality TV content that will hit a surprisingly large demographic. And then it can use recommendation algorithms to distribute that reality TV AI slop to all the people that might or might not want it in order to turn it profitable. Evidence: all the god-awful foreign reality TV on Netflix currently.
It's a bit harder to replace Disney at this point, but give it 20 years. I'm bullish on AI slop on par with The Apple Dumpling Gang or Herbie Goes to Monte Carlo within a decade or so. Evidence: The rapid evolution of AI generated video in the past few years extrapolated into the future.
I keep thinking about how, 20 years, 3D printers became accessible for tech/price, and there was a lot of talk about how we'd all just print what we need (pretty much like in Stephenson's book Diamond Age), and you'd get rich selling digital patterns for material goods.
Instead, we got the elevation of the handmade, the verifiably human created, typified by the rise of Etsy. The last 20 years have been a boom time for artists and craftspeople.
I keep seeing AI slop and thinking that all this will do is make verifiably human created content more valuable by comparison, while generative AI content will seem lowbrow and not worth the cost to make it.
Creative people are on the whole a lot better at keeping AI slop out of art and craft fairs than they have been with the lazier output of 3D printers, CNC, laser engravers and off-the-shelf resin mould art.
It is as if the relatively unspoken feelings about the downsides of technologies as a gateway to art have been rapidly refined to deal with AI (and of course, even the CNC and laser engraver people have common cause).
But I think it is fair to say that if they feel success, there will be a growing pushback against the use of 3D printers, eufyMake resin printing and CNC in a niche where hand tools used to be the norm.
And speaking even as someone who has niche product ideas that will be entirely 3D printed/CNC cut/engraved, I don't really disagree with it. I am mostly not that kind of creative person (putting aside experimental photography techniques) and I see no reason why they shouldn't push back.
The reality is that "craft" fairs are an odd mix of people who spent a lot of time refining their art and selling things that are the work of hours of expressive creativity and effort, and table upon table of glittery resin mould art and vinyl cut stencil output stuck on off-the-shelf products. I think AI might help people refine their feelings about this stuff they once felt bad/incorrect/unkind about excluding.
It's a bit like the way the art photography market is rediscovering things like carbon printing, photosensitive etching and experimental cyanotype, and getting a lot more choosy about inkjet-printed DSLR output.
Yeah, I have a hard time believing that there’s a massive demand for AI-generated videos and images. Like, why would the news industry want to generate images and videos with AI? It’s not news. The advertising industry maybe, but even then it’s probably not your top brands that go full in on it. If you see that all of Apple’s adverts are generated with AI, it’ll probably lower your brand perception.
We already put more importance on handmade goods vs. factory-made, even if the latter is cheaper and better quality. I have my doubts about humanity collectively embracing content generated from prompts by black boxes.
> Yeah, I have a hard time believing that there’s a massive demand for AI-generated videos and images. Like, why would the news industry want to generate images and videos with AI? It’s not news. The advertising industry maybe, but even then it’s probably not your top brands that go full in on it. If you see that all of Apple’s adverts are generated with AI, it’ll probably lower your brand perception.
(disclaimer: I don't work in ad and don't know more about it than the next person)
Whether the end product (the ad) is AI-generated or not is almost irrelevant. The whole production chain will likely be AIfied: to produce one ad you need to go through many concepts, gather reference images/videos, make prototypes, iterate on all that, and probably a ton of other things that I don't know about... The final ad is 1 image/video, but there's been dozens/hundreds of other images/videos produced in this process. Whether the final ad is AI-generated or not, AI will almost certainly (for better or worse...) have a major place in the production chain.
The cost of operating television studios and paying related staff, including on-air talent, is probably significant. I can easily see major news networks turning to AI-generated newsreaders. TikTok is already full of AI voiceovers; seems like a short leap, to me.
Is the cost of on-air talent so great that you want to replace recognisable faces of your network with a generic voiceover?
Personally, I consider TikTok very different to news networks. TikTok is also primarily vertical video. Are news networks going to do that too?
I don't quite follow this point. Master Chief is recognizable. So is Lara Croft. So is Darth Vader's voice. Networks could easily develop virtual personalities with distinctive, bankable, appealing characteristics.
They wouldn't have off-air scandals, require insurance, pensions, teams of wardrobe and makeup artists, security details; They wouldn't need to travel. And that is just the on-air talent. You can replace thousands of tv studios all over the world with a handful of workstations and compute power.
And why haven't they? Master Chief has been around since 2001, Lara Croft since 1996 and Darth Vader since 1977. The technology has been around for ages, and as far as I know, no networks have opted for virtual anchors.
Just from where you are pulling the data that on-air personalities are too expensive?
I don't have a good answer for why they haven't already. I have wondered about the possibility of doing this for 10 years or more.
"The data that on-air personalities are too expensive?" It doesn't seem to me , for the purposes of this conversation, that identification of a cost center required a quantitative analysis. The cost of human talent is non-zero, presumably large enough to merit scrutiny, and unpredictable; that is sufficient, to my way of thinking. So is the cost of the equipment and infrastructure to capture and transmit video image of that human talent, and the humans who maintain and operate that infrastructure.
We've seen several decades of human cost-reduction initiatives, across multiple industries and fields of endeavor, so I'm taking that as evidence that if there is a cost that can be reduced or removed, someone somewhere is looking at doing so. Everything from assembly-line automation, to switching to email over inter-office memos and mailrooms, to the abandonment of fixed-benefit pensions, to self-service kiosks in fast-food restaurants, demonstrates that costs will be cut where they can be cut.
> We already put more importance on handmade goods vs. factory-made, even if the latter is cheaper and better quality
I guess if 'importance' means sentimentality, however, the "factory-made things" industry certainly makes the vast majority of the money to be made in manufacturing. Handmade is a niche. I think the argument is that whether people prefer it or not, AI-generated content will have a major place in the marketplace for content because of its natural advantages. In other words, a movie heavily generated with AI won't win any Oscars but there is a world where such things would make a lot of money and some of that money would flow to OpenAI.
I have a hard time understanding why I would want to waste my time watching something generated by AI when they clearly didn't value the time spent making it.
I have a limited lifespan, I'm not going to use that to consume slop.
That's generally been my principle too. If it wasn't worth making by a human, why is it worth consuming by one?
> IMHO near term a lot of that stuff might start showing up in advertising and news content, the whole media industry is going to be a massive consumer of this stuff.
One of the candidate for mayor of New York ran a whole AI video ad campaign.
Half of the real estate listings have AI remodeled pictures.
Probably a quarter of the printed ads I see each day are AI generated.
It will only increase, but it's already here.
Real estate agents only use AI because it's cheaper than paying $70 CAD per image for human-made images. There's a cap on how valuable the AI generated images are.
The last time I sold a house, a professional photographer took some beautiful photos, gave the realtor like 60-80 and he picked 40 for the listing. He absorbed the cost of that with his commission. No way in hell did he pay $2800-5600 (I'm also in Canada) for those photos.
I'm referring to virtual staging, not photography
I think actually selling media is going to be tough. 1: you have to beat the current price point so people will use it, 2: it can’t be too cheap or you can’t make money. That mad market price point makes it hard to get right. The more content the fewer eyeballs, meaning the less value, meaning people willing to pay less. If it makes making movies a lot cheaper, then there’s a lot less money to be made.
The thing about tech is it has to scale to meet tech valuations, and there’s a reason instagram basically gives content creation tools away for free.
The main thread through the case laid out has a fundamental flaw "assumes a spherical cow in a vaccuum" . There are cheaper/faster more open competitors out there doing as much as OAI and getting integrated in real product pipelines across the planet.
Its not a Blueray vs HD-DVD situation (winner takes all and competition dies on the vine) or early Google vs Yahoo Search i would posit its more like bitkeeper vs git - and will probably eveolve into a git/hg/bazaar type situation. Costs will only keep dropping and the easiest/cost effective options will get faster uptake in the backend dev world where it matters.If im integrating a LLM and the whole field can kinda do the equivalent capex/opex will push the decision.
But there's fierce competition since no moat. How will generated images and videos avoid being a stock photo situation?
This sounds very similar to the Tesla stock value hype train. It’s all a gamble on the promise of tech that isn’t quite there yet.
What I don't understand is how Elon will ever compete with Chinese humanoid robot pricing and output. He's known for not caring about privacy, so there's no real difference between buying a Tesla one or a Chinese one. Both will look and listen into your home and transfer as much data as possible to their true owners.
He congratulates Jeff Bezos for New Glenn, and rightly so, but he's absolutely silent about the Chinese robots. One has to admit that they're already ahead of the US. This comes from a German who well recognizes how far behind Germany is in comparison to the US, in regards to any digital technology.
Oh that's easy: as soon as China gets good enough to compete at any market seen as vital, "the west" either makes up some nebulous national security threat to cut them off or imposes high enough import taxes to make the home-made versions cheaper in comparison. The same way "the west" has done it with cars, phones, and social media in recent years.
There is no free market competition between "the west" and China, they're two completely separate markets for a reason. You're only allowed to cross that isle in a limited capacity, as soon as you get big enough to be seen as a legit threat, you're getting cut off. Works both ways, Tesla can sell in China, but they can't beat BYD. Apple can sell iPhones, but only because their market share is not big enough to matter.
I don’t know if this is ridiculous, but I’m curious if access to LLMs will one day be priced like the Bloomberg Terminal or something. Where access for one user is like 20,000 dollars. Maybe less than that, but like 5k per person.
Seems crazy by most software standards, but when Bloomberg became a software only program (they stopped selling physical terminals) and people were shocked when they paid almost nothing for excel but then so much for the second tool they needed as traders.
Yet it still is priced so high and people pay.
The difference is that Bloomberg Terminals were always expensive, and so people expected to pay. LLMs are basically free (subsidized) at this point, and people are very sensitive to large price increases.
Sure and I’m sure there would be a huge shock, but simple economics would dictate that if that’s the true equilibrium of price for LLMs to be economical, then it would have to get to the price eventually
I think there could be a few directions. Consumer level LLM usage will become free. Corporate-grade LLM use will cost a lot of money. Maybe the highest grade LLM use will be regulated and only done through government labs.
What’s a high grade LLM though in such a competitive environment? And if China releases more high grade open source models, that pricing model is f8cked.
One interesting thing I heard someone say about LLM’s is this could be a people’s innovation, basically something so low margin it actually provides more value to “the people” than billionaires.
that $18,000-$24,000 bloomberg terminal had a 300bps modem
it took longer to generate a page of content or get a complete answer than a free LLM takes on an i5 CPU.
its missing live market data and news but incorporating that with something akin to openrouter would be trivial.
chinese open source models disagree with you
There's good quality LLMs you can run for free, today, so no.
The trend is going the opposite way, intelligence as cheap as possible.
The top two comments mention that capturing, at minimum, 2% of the digital advertising market seems low. I disagree. I see that as an enormous hill to climb.
The LLMs will create new content, but they aren't creating new business channels in the advertising industry. As an example even once Google achieved search domination they still didn't have this. They had purchase a lot of things to make that happen like Urchin, Adscape, DoubleClick, YouTube, and a lot more.
It's really hard for start ups to compete for VC money: "Hello Mr. VC I'm going to burn your money for buzzword" just no longer works, first openAi has an industrial buzz word generator, second their money burning plan scales much better than my money burning plan.
Yeah but they aren't a diversified portfolio of money burning plans and that's the real secret to responsible investing. I think Warren Buffet said that.
TechCrunch: Anthropic is projecting positive cash flow by 2028, while OpenAI is expecting sizable losses, with cash burn reaching $14 billion in 2026
This should be a easy barrier for OpenAI, cgpt has only been public for 3 or so years and they are already close to mediating most of the global economy and human decision making
AI credits will likely replace current money anyway
Raising $207B is plausible vs the dozens of Nuclear Power plants they also need.
No, they can use the ideas from The Matrix for power delivery.
Yeah, that's the "so it can continue to lose money" part.
ChatGPT just launched "shopping research."[0]
Hideous idea as it is, I fully expect they break even in 2026.
[0]: https://openai.com/index/chatgpt-shopping-research/
Google is already integrating this directly in search https://blog.google/products/shopping/agentic-checkout-holid...
I tried it yesterday by comparing the specs of a few low-budget desktops for my father. It worked well and saved me a lot of time and the hassle of comparing multiple tabs on a PC.
I am a bit worried about the feature where it calls the shops for inventory checking. That's the whole point of having a website. Now we are going to have expensive AI that calls other AI answering machines and no value will be added. Meanwhile, it will become even more difficult to talk to a human when necessary.
Honestly, this is huge for people like me who tend to over-research and over-think the hell out of product choices. "Find me a top-fill warm-mist humidifier that looks nice, is competitively priced against similar products, and is available from a retailer in $city_name. Now watch for it to go on sale and lmk."
If they can figure out how to get the right kickbacks/referrals without compromising user trust and really nail the search and aggregation of data this could be a real money-maker.
Trusting AI with your shopping is very short sighted.
Lol what a terrible idea. Why not just hand every decision you'll ever make to AI?
Nobody needs critical thinking or anything. Just have AI do it so you save $3 and 4 minutes.
Why would I want to spend 1-2h researching humidifiers if I can spend that time in any other way, and still end up with a humidifier that fits my needs first try?
This kind of task is perfect for AI in a way that doesn't take away too much from the human experience. I'll keep my art, but shopping can die off.
Because you end up with a $1 value piece of crap that someone spent considerable time optimizing and faking reviews for LLMs instead of on the product. Basically in the medium term this strategy will get you temu stuff
How often do you buy the first result on an Amazon search? Because that's delegating your labour, isn't it? Surely the best products are getting to the top, right? Well no, they're being paid to get to the top. An LLM that has in-app shopping is gonna be the same thing
for the same problems with amazon, youre relying on a computer to tell you what to buy, which is very shortly going to be infested with promoted products and adverts instead of genuine advice. The AI implementers will poison the responses in the name of advertising, of this i have zero doubt in my mind.
10+ years ago you could in fact just pick the best-reviewed product on Amazon at a certain price point and have a great experience! God help you if you tried that today.
Well, you can always do the same thing that an LLM would: open SEO spam ranking sites "best humidifiers 2025", filled with referral links to Amazon or other sellers, which basically copy product descriptions and assign rankings that aren't based on any tests or real data.
Compared to what, reading Amazon reviews? Google site:reddit.com?
Fundamentally, is it really that different from being persuaded by an advertisement or trusting what the marketing says on the box?
> Just have AI do it so you save $3 and 4 minutes.
Maybe I am deeply suboptimal, but typically this kind of decision takes me far more than 4 minutes.
For this to be useful they need up to date information, so it just Googles shit and reads Reddit comments. I just don't see how that is likely to be any better than Googling shit and reading Reddit comments yourself.
If they had some direct feed of quality product information it could be interesting. But who would trust that to be impartial?
It’s better in that I don’t have to waste my time reading Google and Reddit myself, but can let a robot do it.
Do you buy the first item that pops up on Amazon for a search that you've made? Because that's letting the robot do it for you.
If the answer is "no because that's an ad", well, how do you know that the output from ChatGPT isn't all just products that have bought their rank in the results?
Project Farm solves the trust problem with methodology + video documentation and the monetization problem with affiliate links for every product tested.
I can Google and read but it takes a lot of time
> If they can figure out how to get the right kickbacks/referrals without compromising user trust
i'm trying to envision a situation in which the former doesn't cancel out the latter but i'm having a pretty hard time doing that. it seems inevitable that these LLM services will just become another way to deliver advertised content to users.
> If they can figure out how to get the right kickbacks/referrals without compromising user trust and really nail the search and aggregation of data this could be a real money-maker.
As another commenter points out, "not compromising user trust" seems at odds with "money-maker" in the long-term. Surely Google and other large tech companies have demonstrated that to you at this point? I don't understand why so many people think OpenAI or any of them will be any different?
Meh. Google has something better cooking.
https://www.theverge.com/news/819431/google-shopping-ai-gemi...
I'm sure they'll have no issues raising the funds. If they run into trouble, they can always ask ChatGPT for advice.
LLMs have high utility for coding. The PMF is so strong that even with just coding i feel they will see an ARPU expansion beyond conservative assumptions.
There are adjacancies in white collar work like financial analysis that they will go after. All these will capture high ARPU usage.
Consumer is not their only path to revenue but it is probably the easiest to model. The enterprise play to automate and accelerate some white collar workers is a clear target not reflected here.
but with coding, the users are mercenary and will instantly switch to a model that is better. Less sticky compared to the consumer product
They can get all the enterprise contracts they want but I don't think it will be enough
They've committed to spend an amount equal to the entire US defense budget just in October.
I swear on the crap I just pulled out of my ass that I'll spend 100 quadrillion dollars. Imaginary ones of course. Now pump my stock.
Seriously though, massive recession is on the horizon, no one is going to be spending squat.
I think you mean the US WAR budget! /s
I heard this about Amazon for 20 years
Amazon raised like 10million. People complained about lack of dividends but that was at least money earned
You misunderstand the Amazon story. Amazon did two pre-IPO funding rounds: <1m angel round and an 8m Series A led by Kleiner Perkins. That's it. In contrast, OAI has raised almost 60 billion in 10 funding rounds, and if you RtFA you can see the estimates are they have to raise hundreds of billions more.
For the IPO itself Amazon sold 3 million shares at $18, for a raise of 54m (from the IPO alone they had enough cash to pay off every investor up to that point). In July 2001, in the heart of the dotcom crash, they raised 100m by selling equity to undisclosed investors, and of course they have been using shares as part of their compensation packages for a very long time, but that's about it for Amazon's entire equity raises.
They did raise 15b in a bond issuance a few weeks ago, their first bonds issued since 2022, with the money going to several things but mostly AI. However, since bond payoffs are very different from selling equity this is a very different play from what OAI is doing. Amazon will never pay more than a fixed amount for that money, capped upside to the bond-holders.
The reason this is different is that Amazon has largely run either a small profit or a small loss, quarter after quarter, because they take their profits and instead of recording it, putting it in a bank, or dividend-ing it to shareholders they put it into building datacenters and warehouses and software and the like. But because of that enormous cash generation they have only rarely tapped outside investors, either in bonds or equity markets. OAI is not generating near enough cash to fund their operations, so they have been selling equity in absolutely enormous quantities- they have already raised more cash pre-IPO than any company in history and outside estimates like this one from HSBC call for them to blow past the amount they've already raised. This is fundamentally very very different.
Amazon had a product on day 1... particularly one that was unique and made a lot of sense for the moment it was introduced.
OpenAI doesn’t have a product? Have we existed in the same reality for the last 3 years? Something something fastest grown user base in the history of tech
Selling books online was certainly profitable, but I'm not sure about unique. Amazon's big success is that they had no particular ties to any existing publisher so they didn't have the corporate headwinds of "this will kill our brick and mortar stores and their distribution systems!".
No, it was unique. Being able to browse such an extensive collection of books and order them from your computer was mind-blowing.
Amazon wasn't committing to larger spending in one year than the US defense budget.
2025 US defense budget is $849.8 billion[1].
Had a look and:
>In total, OpenAI aims to invest approximately $1.4 trillion in computing infrastructure – encompassing Google Cloud, Nvidia chips, and data center expansions.
Huh yeah fair. That's more than the yearly defense budget. Absurd. Though I'm sure it's not _yearly_
- [1] https://en.wikipedia.org/wiki/Military_budget_of_the_United_...
The sovereign wealth funds and billionaires need something to do with all their available cash. It's no fun just watching the number increase anymore, that got boring quite a while ago.
That is such an insane bet though! How many trillions can they have itching in their pockets
To add to the valid counterpoints in the existing replies, Amazon was cash flow positive for most of those 20 years, so they were investing their own money back into things that could generate even more revenue (and profit!) for the company.
OpenAI is still losing money much faster than it can make it and is planning to accelerate those losses indefinitely.
I think the most likely outcome is that it turns into something like Uber, where they continue to lose money waiting for a major technological leap (truly unassisted and reliable AI in this case, fully self-driving cars in the case of Uber) and then pivot a bit to a largely unnecessary and poorly executed business model that people reluctantly use for the most part (with some eager advocates) but makes some money.
People still said this about Uber as of last year, few realizing that they're generating $10B in profit a year now
Uber is also more or less a monopoly in many place. They took over the taxi industry and now drivers make less money than before.
OpenAI has multiple competitors, who all build their LLMs for less money.
And unlike taxi drivers / companies, these competitors have much deeper pockets.
It is not hard to imagine Google being able to outlast OpenAI for a decade and it is hard to imagine OpenAI being able to survive for more than another couple of years given that.
Is there any link directly to the report? I am unable to find any in this article or a number of others which seem to just copy the same information here.
I think beyond the number of crazy assumptions (no Google taking market share in the consumer market?? only 2% of digital advertising expected to be captured by OpenAI?) it is hard to nail down which levers could move which might make this funding hole disappear.
That’s 207b / 8b people / 4 years / 365 days or 1.17 cents a person.
If ChatGPT is delivering that, they should have no problem raising money.
That if competition doesn't under-cut you and you convert at least part of the free users in the process.
And with near every level of hardware getting some kind of AI acceleration a lof of the uses might even be local
I don’t think LLM leader is going to come out of this AI race on top. OpenAI is like Yahoo! in my view. They were there in the beginning and led by default. Someone somewhere will be the Google of this era. It won’t be LLM but the next step. Leaps and bounds better than LLM. I also think we won’t use even 1/100th of the energy needs projected right now.
The reason for this is that leverage is extremely expensive and punished in the current macro regime. OpenAI will lose money faster if they scale up. That's because the Fed can capitulate all it wants, but real interest rates and real yields won't move. The market is saying "pay me".
Previously: https://news.ycombinator.com/item?id=46054092 | 11 hours ago | 44 points | 19 comments
Google earned $264.59 billion in ad revenue in 2024 alone. OpenAI can figure it out.
You already can't trust llm output. How will you trust it when its actively steering you based on profit motive?
Do you trust the result of your Google searches?
Who does?
Not really but it would be nice if we could have high quality non manipulative information :/
I have two words: 1) government contracts 2) ad dollars
That are 4 words. ;-) Other then that yes, goverment contracts, meaning offloading the costs to the regular people. I think the US goes all in with AI, trying to get world dominant and controlling the market. But I think this is a mistake and the US will get into trouble with this.
The third might be 'AI Device'
It is not losing money, the whole industry haven't figured out how to generate revenue yet. it is the starting phase of LLM like we experienced in .com bubble. Is it a bubble at this point? Yes. Can it generate revenue in the future? Yes. Will openAI survive in this bubble? Maybe.
You seem to be complicating your thinking here.
They are spending more money than they are bringing in. This means they are losing money.
It is quite literally losing money at an unsustainable rate. They need a path to profitability otherwise this is a massive boondoggle for all of the investors
Unfortunately “all of the investors” is in reality the whole world’s markets due to the disgusting top heavy nature of the SP500 currently.
I have been talking to AI a lot about what portfolios will survive that crash. :)
And what is the conclusion, what will survive?
we are still at early age. think it as human, the early age is burning money. when grow older, the ability of making money increases.
Compared with .com bubble. most .com dies, and the one who survived are GOOGLE, AMAZON, Tencent etc.
Yes but you need money to survive until the later stages
Yeah, financial bubble != useless technology. Maybe coding agents do cost $50 per month in the long term, but I might just pay that for entertainment and personal stuff. Like, I don't even try to vibe code my job, but in the evenings having a cool slop generator is good times.
Right now I use a Chinese vibe code plan, really good value.
[dead]
Can we just stop putting the "I" behind the "A"? Thanks.
They aren’t going to get new equity funding they are going to go public so they can start issuing bonds.
Meanwhile Google does not have to raise money and they're profitable whether AI works or not.
The most interesting point about OpenAI I have heard lately is they are literally trying to make themselves too big to fail. If they go down so does everyone else, which explains all those strange deals with everybody and the comment from their (cfo?) about being backstopped by the gov.
If OpenAI fails, they could potentially bring in their downfall the major cloud providers who invested in hardware for them, expecting that it would pay-off over time.
Only Oracle went into debt to fund this expansion, and may well die. The rest of the cloud oriented mag7 used cash, can afford to write it off, and will continue being monopolies unimpeded.
I don't see a world where there is such a catastrophic failure, unless someone comes up with a significantly more efficient architecture.
We're barely scratching the surface of the utility of LLMs with today's models. They aren't more pervasive because of their costs today, but what happens if they drop another order of magnitude with the current capabilities?
'Potentially' but we are nowhere close to this. Hyperscalers print a _lot_ of money they can afford to lose. Even Nvidia wouldn't be in too much trouble yet. (The pure LLM companies are already toast, IMHO).
Oh no not the Cloud!
This seems accurate, and plausibly the only way out. The biggest issue I see here is that in this case... the greed might topple a government
Hot take: a flagship silicon valley startup built on hype and overzealous ambition crashing and burning in 2026 is exactly what the industry needs right now.
> If they go down so does everyone else
What does that even mean?
Let's say OpenAI signed a commitment contract that they agreed to spend XXX USD in your company over N years. You invest in infrastructure, your contractors sometimes take loans, the construction companies take loans. Countries / Funds lend money to such companies (example: Saudi Arabia fund), these funds themselves raised debt, it can quickly spiral down.
If OpenAI fails, wouldn't their customers just move to other AI providers? So the total hardware demand by AI companies wouldn't dramatically change over a reasonable time frame?
If OpenAI fails, the assumed case is overcapitalisation relative to demand. In such a world, if the reason OpenAI has no liquidity is because there's insufficient demand, there's no customers to move across.
What happens if the OpenAI failure is because of lack of demand from their customers and the market in general? None of the AI providers or hardware providers will survive, no?
Similar to certain banks in 2007/2008, the idea would be “so much investment is tied to one company that if that company went bankrupt, it could have consequences for the broader economy”
The thing is, it is not 2007/2008 any more. The US government is holding record amounts of debt and countries around the world are now trying to become independent of it. This includes its bond markets on which the dollar relies upon to give it its reserve currency status, which in turn is what gives it its power to print money and bail industries out. If something happens that requires Big Tech to be bailed out and international bond holders decide the US is no longer reliable, it could very well end up triggering the collapse of the US dollar as the world's reserve currency and the downfall of the US as we know it.
The stock market will lose faith in AI companies, which will crash the stock price of Google, Microsoft, Oracle, Nvidia and CoreWeave. Investors will lose billions, many of those investors are pension funds. Any AI projects that aren't already profitable will shutdown.
And, because AI is currently what prevents the US economy from being in a recession (at least that what some people speculate), the US economy will stumble, which means that everyone else will to.
Is it unclear? Compared to other times a "too big to fail" industry failed?
If OpenAI crashes, for example funding stops, they go broke, fall behind, nobody buys anything, then all the money they invested for data centers or demand they created for NVIDIA chips and compute collapses. That creates surplus of hardware, causes lots of construction/buildout / stockup orders to get cancelled, and the whole thing ripples as suppliers and construction and data center providers etc etc suddenly lose a ton of anticipated profits.
Share prices drop as people dump to protect their portfolios, anticipating dips in the prices because share prices will drop as people dump to protect their portfolios (I'm not kidding).
Given that the big 7 AI companies are basically _all_ of the market growth lately, it doesn't even take a serious panic / paranoia episode to see the market itself stagnate or significantly regress, as people pull from anything AI related, and then pull from the market itself anticipating the market will fall.
It's a fairly standard playbook at this point.
That's different from saying that boeing is too big to fail for example. The US can't accept to lose its only major commercial aircraft manufacturer. Or Intel for similar reasons.
But what you're describing is about keeping the AI bubble from popping. Can a bubble really be too big too fail?
In terms of the stock market since without AI thw US would be in a recession.
People crying about the revenue gap constantly forget that OpenAI still hasn't turned on the ads, porn, and gambling. Trust, they will turn it on eventually.
Try using the shopping feature they just launched. It's literally just jamming ads into your chat session.
https://openai.com/index/chatgpt-shopping-research/
This is what they hired Fidji for, and they are going down the same dark path she took Meta down.
Everything’s for sale, and they will amplify whatever == revenue.
> Alphabet annual revenue for 2024 was $350.018B, a 13.87% increase from 2023.
So, they need to "just" get 20% of the market from Google to break even...
Which is silly because the whole point of asking an AI what induction stovetop to buy is so I don't have to do "shopping research"
Why would OpenAI have any sort of advantage in these industries compared to the established players?
it's pretty sobering to think that the so-called harbingers of SkyNet AGI have to fall back to mafia-era revenue streams like vice to convince shareholders that their money wasn't wasted.
I think the ads will be turned on, inevitably.
I hope the porn and gambling aren't turned on, but they will be. Probably under spin off companies to shield the brand, but using the tech.
I am suspect that ads will be able to cover the cost of inference, much less model building, salaries, product development, etc. The ad industry is built upon _and priced for_ websites that return results for virtually zero cost. That isn't going to be true for rendering AI.
They're already turning on the porn: https://www.bbc.com/news/articles/cpd2qv58yl5o
if I had my druthers, we'd leave the ads but get the porn and gambling.
two out of three of those require me to want to do them for them to affect my life.
You are more attracted to ads than porn?!
Define "attracted". I believe porn and gambling both do incredible amounts of damage to society. Ads probably do some damage to society, though I can recognize some benefits in being able to advertise and market products and services. I don't believe there is any benefit to the other two.
HSBC, the firm who did the analysis in the article, took into account projected revenues matching user numbers of 44% of the world's non-Chinese population.
With zero market share for Google.
Living in AImaginationland must be nice. C’mon kids, let’s all sing the AImagination song: AImagination, AImagination, AIiiimagination, …
44% for OpenAI wouldn't mean 0% for Google, but considering that even Apple only has about 30% of the global smartphone marketshare, OpenAI getting there seems unlikely.
It would be plausible but very optimistic projection if there was no competition that is currently crushing OpenAI
Oh, I agree it's implausible. But GP said
> People crying about the revenue gap constantly forget that OpenAI still hasn't turned on the ads, porn, and gambling
But quite the opposite, HSBC assumed that they will have a virtual global monopoly on AI, and even under those projections they will still need to take on hundreds of billions of debt. I'm sure if they get there getting access to that debt will be easier than I'm assuming currently.
Do you believe so ? (like real question, not sarcastic here)
I think that a company with a global monopoly on a thinking machine that replaces wide swaths of current internet business categories, whose user base makes up most of the world outside of the GFW, would be very valuable on the public market and would probably have access to as much debt as they want, even if it had negative consequences for the world economy. Even the 2nd or 3rd place company in that race will be quite profitable. It's not impossible, just implausible.
They may not be shipping good enough products, but on the flip side it still feels like they have almost no competition outside of coding and image gen. The EOL termination of 4-o should be some evidence of this.
OTOH, why don't they ship good enough products? To me all of OpenAIs recent investments strongly suggest they hit a dead end with their current LLM approach. After all, if they knew the path ahead for GPT looks great, why don't they invest into training the next big thing instead of doing datacenters with the intention of renting them out?
Ok, probably true, but it’s still a pretty far fall from “we’re just about to deliver AGI which will put us in the driver seat of the entire global economy.” Which was the core value prop of their $trillion investment pitch. Like, there are way cheaper ways to deliver ads, porn, and gambling than training and operating huge LLMs.
I think you're right, however that only defeats themselves in the long run. AI is something that can be run locally, while search, shopping, movies, social, etc. cannot. And once the ads are baked into the product, you will need a force like the EU to remove them else shareholders will riot. Ads will be the perfect weapon to shoot themselves in the foot.
The cost per generation is still too expensive, they would not profit from ads and very few people pay for porn. I don't see how they'd profit from gambling either.
Ads/porn I get but how does genAI powered gambling work?
"ChatGPT, here's 5 dollars. Please roll a fair die. did i win?"
This is truly the most stupid timeline.
> ads, porn, and gambling
We already have those at home without OpenAI.
Also, the competition would be ruthless.
Sure, because there's one single ad network, one single place where you can gamble and one single porn site in the entire planet ...
That is exactly the parent's point, that while those things could bring in revenue, the competition in those industries is already incredibly fierce. And that's before potential new entrants into the space on the AI front. Even in the best of cases I don't think anyone is gonna be able to pull anything other than middling margins.
They've already got (way!) more users than those other sites, though.
But yeah, maybe they should give up because of the imaginary obstacles brought up here, lol.
>potential new entrants into the space on the AI front
Are these "potential new entrants into the space on the AI front" in the room with us right now?
I'm not saying they should or will give up on these markets, just that their margins aren't likely to be great.
There's already so much AI porn right now. There's no reason to believe that OpenAI will be any better or able to command higher margins than their competitors. They also have the problem of being under a ton of public and regulatory scrutiny so the highest paying clientele, the people with niche unsavory fetishizes, won't be able to be serviced by them.
And that's ignoring the fact that getting into this space at all is pure HN speculation and unlikely to ever happen anyway.
This is absolutely what techbro MBAs raised on case studies of Apple and Toyota are likely to think. Gamblers and porn users do not have any sort of brand loyalty and are likely to be hostile to the idea of "vertical integration" and "360 degree views of the customer".
I fully expect all three to happen.
"Google is excluded entirely"
(Imo) They will turn to all of these(especially to porn and gambling) when the core model of "enhance your life" will slowly fade away. The "academia space", the teens/boomers demographics, all of those will stop using OpenAI at scale if they're bombarded with vices (porn, gambling, etc).
Ads & referrals are already in the works, and people are generally tolerant of those. But, as with any company, appearances matter. ChatGPT will definitely lose users at the slightest possibility of having non-sanitized content served to more morally sensible groups.
there will be differentiated cohorts
Wouldn't that entirely kill the AGI delulu narrative (for people who still believe in it)
Delusion isn't swayed by facts. There are more than enough examples of this in the world today.
How to read FT for free
If you sign-up for a free account you can read FT Alphaville, which is the blog-style section of the FT containing articles like this one
https://news.ycombinator.com/item?id=46058125
"If you owe the cloud computing company a hundred dollars, it's your problem, but if you owe the cloud computing company 207 billion dollars..."
Too big to fail. It's essentially 08 part 2 with the expected irr.
OpenAI deals will likely end up with government backing in the next 12 months. Then we’ll all be on the hook for it.
But this is not a bank, or an airline, or a real estate giant.
If OpenAI goes bankrupt, what happens? People won’t be able to write their precious slop oh no and serious professionals will just switch to any other LLM provider
Users might simply switch to Gemini or Claude but partnered companies might go bankrupt or see their stock tank.
This is why consensus will never produce innovation.
OpenAI has many plausible monetization avenues. (Whether it will execute on them -- a fair question. But acting like they just don't exist is low IQ.)
Take ONE example: online shopping.
Online shopping is $6-7 trillion per year, growing 7% annually.
Suppose ChatGPT captures 10% of that with a 5% fee. (Marketplaces like Amazon and Walmart charge 15%. Ebay is 5%.) That's $200B over five years.
TL;DR - financial analysts look at current charts and project them foward by 5 years and go "wow the numbers look bad".
Sure OpenAI may well be bleeding money into the 2030s, or may even go bust completely depending on how pessimistic you are, but the analysis completely skips:
- They are building their own data centers, and will be less reliant on renting compute from Microsoft and Amazon over time.
- Once the AI bubble has subsided costs for GPU purchases and rentals will decrease significantly. Plus there will be more advancements and competition in the space (e.g. Google TPUs) and Nvidia will no longer be able to name their own price.
- We will write more efficient software for training and inference.
- Once user growth is tapped out OpenAI will no longer need to have the overly generous free tier that they do today. And if they decide to turn up the advertising faucet these users could bring in a ton more revenue than in the projection. Thinking that every AI company combined will capture only 2% of the total digital advertising market is ridiculous. AI apps are already challenging social media for scrolling time.
Basically, the entire space is evolving so rapidly that it's pointless to make a projection with the assumption that the landscape isn't going to change from here on.
AI is/will be the mother of all bubbles.
What about the usual capitalism point of view?
If their business isn't sustainable they should go bankrupt, and close shop.
Capitalism makes more sense if it's thought about purely in terms of individual self interest as supposed to something that leads to an efficient market on aggregate.
The CEOs making big calls across the economy have already negotiated golden parachutes in the event of their failure.
The financiers and lawyers getting a chunk of each bond deal they close have every incentive to raise more than what's actually needed. Investment funds flush with ZIRP dollars have every incentive to plow it back into investments to show that "the money is at work".
Yeah self interest is how we end with world we have today at the edge of destruction through various mechanisms.
Boom and bust hype cycles have always been a feature of capitalism, especially for advanced technology. Perhaps the only way to know the limit is to overshoot.
And then take the whole economy with it when blows up.
Not just the economy. We usually deal with these busts and crisis collectively through the public sector as well.
There's also a hidden opportunity cost in regards to hype cycles. So much energy, attention and money flows into the hype, while other businesses or entire sectors get overlooked and underappreciated.
Except the 7 firms make up >30% of all market growth, and move more capital than entire countries. Yes, they are hemorrhaging cash with every new customer, and burdening taxpayers by eating community electrical and water infrastructure.
In terms of business, it is not sustainable:
https://www.youtube.com/watch?v=t-8TDOFqkQA
The hype-cycle is nothing new =3
"Memoirs of extraordinary popular delusions and the madness of crowds" (Charles Mackay, 1852)
https://www.gutenberg.org/files/24518/24518-h/24518-h.htm
“We’re gonna need more fuel for this fire!”
A bit like a socialist country.
Nope, workers don't own the means of production -- this is the dictatorship of the bourgeoisie, baby.
But it's happening in the biggest capitalist country in a capitalist world
Not at all. This is capitalism.
Bailouts are not capitalism
They are if they happen under capitalist states, which is what the US is.
So if something is declared "capitalist", it maintains that state in perpetuity, regardless of the actions it takes in reality?
No, but it is a materialist analysis of capitalism in practice, rather than unrealized idealism
Queues are not socialism?
No, this is Sparta
Something something true capitalism has never been tried before
In unconstrained socialism, a small group of powerful individuals use government control of corporations for their personal benefit.
In unconstrained capitalism, a small group of powerful individuals use corporate control of government for their personal benefit.
Clearly, these outcomes are completely different.
"unconstrained socialism" is kind of an oxymoron, isn't it?
Exactly - private profits, public losses, is quintessential neoliberal capitalism.
Not when the government steps in because OpenAI, NVIDIA, Microsoft, etc are all going down together. There is so much pseudo-money-laundering going on between those companies. I would be shocked if the government (e.g. 2008 style socialism for the rich) steps in to make them “too big to fail”.
Microsoft and Nvidia aren't going down.
Their stock prices might go down but they're not going down.
Stock prices going down is the reason AIG, GM and Chrysler received bailouts.
Debt and an inability to pay it by selling stock is why they received bailouts.
Nvidia and Microsoft stand to make less money, not be in debt.
still capitalism
> as ubiquitous and useful as Microsoft 365
What is this saying? Is this sarcastic?
I don't know anybody with a Microsoft 365 subscription.
I suppose the cloud storage is nice, but you can do much better; Google gives you double that for the same price ($99/yr).
Businesses, a lot of people have it through work.
If they want to come talk to me I can help them out of this. Usually the smart guys are too busy talking to hear an offer though.
It's so sad how much money leaders will effortlessly pump into something like this, when we still have existential threats of climate change, incurable diseases, poverty, housing, and so on.
Meanwhile ungodly amounts of money are being used so some boomer can generate a AI video of a baby riding a puppy.