I can generate $20 billion in ARR this year too! I just need you to give me $100 billion and allow me to sell each of your dollars for 0.2 dollars.
I've been able to help setup cross app automation for my partner's business, remodel my house, plan a trip of Japan and assist with the cultural barrier, vibe code apps, technical support and so much more.
I think that there were some article here that claimed that even inference is done at loss - and talking about per subscriber. I think it was for their 200$ subscription.
In a way we will be in a deal with it situation soon where they will just impose metered models and not subscription.
1. It actually under performs Claude, Gemini and even some of the Grok models for accuracy with our use case of parsing PDFs and other rather arbitrarily formatted files.
Of course they are.
> As long as the inference is not done at a loss.
If making money on inference alone was possible, there would be a dozen different smaller providers who'd be taking the open weights models and offering that as service. But it seems that every provider is anchored at $20/month, so you can bet that none of them can go any lower.
Maybe, but arguably a major reason you can't make money on inference right now is that the useful life of models is too short, so you can't amortize the development costs across much time because there is so much investment in the field that everyone is developing new models (shortening useful life in a competitive market) and everyone is simultaneously driving up the costs of inputs needed for developing models (increasing the costs that have to be amortized over the short useful life). Perversely, the AI bubble popping and resolving those issues may make profitability much easier for the survivors that have strong revenue streams.
There are! Look through the provider list for some open model on https://openrouter.ai . For instance, DeepSeek 3.1 has a dozen providers. It would not make any sense to offer those below cost because you have neither moat nor branding.
- good tools for agentic workflows
- no tools for context management
- infrastructure for input token caching
Also, there are many providers of open source models with caching (Moonshot AI, Groq, DeepSeek, FireWorks AI, MiniMax): https://openrouter.ai/docs/guides/best-practices/prompt-cach...
Only the self-hosting diehards will bother with that. Those that want to compete with Claude Code, Gemini CLI, Codex et caterva will have to provide the whole package and do it a price point that is competitive even with low volumes - which is hard to do because the big LLM providers are all subsidizing their offerings.
OpenAI is a basket case:
- Too expensive and inconvenient to compete with commoditized, bundled assistants (from Google/ Microsoft/Apple)
- Too closed to compete with cheap, customizable open-source models
- Too dependent on partners
- Too late to establish its own platform lock-in
It echoes what happened to:
- Netscape (squeezed by Microsoft bundling + open protocols)
- BlackBerry (squeezed by Apple ecosystem + open Android OS)
- Dropbox (squeezed by iCloud, Google Drive, OneDrive + open tools like rclone)
When you live between giants and open-source, your margin collapses from both sides.
The WWW in the 1990s was an explosion of data. To the casual observer, the web-browser appeared to be the internet. But it wasn't and in itself could never make money (See Netscape). The internet was the data.
The people who build the infrastructure for the WWW (Worldcom, Nortel, Cisco, etc.) found the whole enterprise to be an extremely loss-making activity. Many of them failed.
Google succeeded because it provided an application layer of search that helped people to navigate the WWW and ultimately helped people make sense of it. It helped people to connect with businesses. Selling subtle advertising along the way is what made them successful.
Facebook did the same with social media. It allowed people to connect with other people and monetized that.
Over time, as they became more dominant, the advertising got less subtle and then the income really started to flow.
Salesforce is similar in that it helps businesses connect with and do business with each other. They just use a subscription model, rather than advertising. This works because the businesses that use it can see a direct link to it and their profitability.
Salesforce doesn't make a good product, and certainly not the best product. It doesn't matter, you don't need to if you can convince idiots with money to invest in you. And then the switching cost is too much, too late.
That business model is a dying one and all the software companies know it. That's why Microsoft has spent the last 15 years opening up their ecosystems. As automation increases, switching cost decreases. You cant rely on it.
It's worth that much to me in the time saved. But I'm a business owner, so I think the calculus might be quite different (since I can find ways to recoup those costs) from an individual, who pays out of their main income.
I outlined examples of how I used CC/AI a couple months ago [1]. Since then I've used it even more, to help reduce our cloud bills.
Normies literally see no difference between GPT and Claude, just that Claude is much more expensive and CEO is even more of a dummie than Altman.
The challenge is that if the numbers are accurate they need 5-10x to break even on inference compute costs, before getting into training costs and all the other actual overhead of running a company like compensation.
Will everyone be willing to pay 5-10x? Probably no.
Will half of users pay 10-20x? Or a quarter pay 20x++?
Or we end up with ads … which already seem to be in motion
That's not to say that there aren't many, like you, for whom $500 is a perfectly good deal, there's just not nearly enough for OpenAI to ever turn a profit.
But ya, OAI is clearly making a ton of revenue. That doesn't mean it's a good business, though. Giving them a 20 year horizon, shareholders will be very upset unless the firm can deliver about a trillion in profit, not revenue, to justify the 100B (so far) in investment, and that would barely beat the long term s&p 500 average return.
But Altman himself has said he'll need much more investment in the coming years. And even if OAI became profitable by jacking up prices and flooding gpt with ads, the underlying technology is so commodified, they'd never be able to achieve a high margin, assuming they can turn a profit at all.
I think there's something off with their plans right now: it's pretty clear at this point that they can't own the technological frontier, Google is just too close already and from a purely technological PoV they are much better suited to have the best tech in the medium term. (There's no moat and Google has way more data and compute available, and also tons of cash to burn without depending on external funding).
But ChatGPT is an insane brand and for most (free) customers I don't think model capabilities (aka “intelligence”) are that important. So if they stopped training frontier models right now and focus on driving their costs low by optimizing their inference compute budget while serving ads, they can make a lot of money from their user base.
But that would probably mean losing most of its paying customers over the long run (companies won't be buying mediocre token at a premium for long) and more importantly it would require abandoning the AGI bullshit narrative, which I'm not sure Altman is willing to do. (And even if he was, how to do that without collapsing from lack of liquidity due to investors feeling betrayed is an open question).
I mean, so was netscape.
The mass commoditization of the tech is rapidly driving AI to be a feature, not a product. And Google is very strongly positioned to take advantage of that. Microsoft too, and of course they have a relationship with OpenAI but that’s fraying.
Netscape, to a large degree, killed itself.
Not to say IE turned into anything good though. But it did have its hayday.
TPUs are cool, but the best leverage remains to reduce your (active) parameters count.
I don't think ads are that easy, because the hard part of ads isn't taking money and serving up ad slop, it's providing convincing tracking and analytics.
As soon as ad slop appears a lot of customers will run - not all, but enough to make monetisation problematic.
There isn't even a tenth of enough money if you group together all of advertising. Like, the entire industry. Ads is a bad, bad plan that wont work. Advertising is also extremely overvalued. And even at it's overvalued price tag, it's nowhere near enough.
Logically speaking, yes it is easy to switch between OAI and Gemini, or Coke and Pepsi. But brand loyalty is more about emotions (comfort, familiarity,..) rather logical reasoning.
If ads are so overpriced, how big is your short position on google? Also ads are extremely inefficient in terms of conversion. Ads rendered by an intelligent, personalized system will be OOM more efficient, negating most of the "overvalue".
I'm not saying they should serve ads. It's a terrible strategy for other reasons.
That’s possible because they’re immensely profitable.
Well those are obviously worse products.
> If ads are so overpriced, how big is your short position on google?
I hate hearing this stupid, stupid line.
Most companies are run by neanderthals with more money than brains. Companies burn money on advertising because why not? Making your product better is hard and takes time, advertising is the easiest thing you can do. Does it work? Not really, no, but you get extra business for as close to zero effort you can possibly get. Hit a wall? Just advertise more!
> Ads rendered by an intelligent, personalized system will be OOM more efficient, negating most of the "overvalue".
This is exactly what people said about personalized ads. "No you don't understand! It's not like a billboard!"
And that's true, but consumers are not fucking braindead, and there's also the laws of economics. If I have 50 bucks, I'm not spending 20 fucking dollars on your dumbass paint, no matter how much you advertise it. And that's not a me thing, that's a consumer thing. You can spend 1 quadrillion dollars advertising ferraris and guess what - you will STILL quickly saturate that market and hit a hard ceiling. Because consumer's can't afford it.
And that's not even touching on the fact that most of the metrics around advertisements are just obviously bullshit. How many human eyeballs are actually on ads? Much, much less than everyone thinks.
Yes, sure, we can build highly personalized ads. Whatever. But at the end of the day, consumers still have the exact same amount of disposable income as before. We have created Z E R O value, what we have done is consolidated it.
Hmm, what happens when markets consolidate too much? Well, I guess that would mean advertising becomes completely worthless, wouldn't it? What a conundrum! It's a good thing our markets haven't been consolidating for the past 70 years...
I think you underestimate how valuable being the top slot on google is. Just the other day i googled “bluetooth speaker” and bought the first result (an ad). One hour of that can net you millions of dollars. That’s why consumer brands bid more and more every year on digital advertising.
For many brands, yes, and they don't know it.
> I think you underestimate how valuable being the top slot on google is.
The more you advertise, the less valuable each ad space becomes. Consumers have a lot of money they have to dole out. Giving them more ads won't increase that pot of money - it will make your cut smaller and smaller as it's split across more brands.
> consumers have a lot of money that they dole out. More ads wont increase the cut of money
Consumer spending is not a fixed pie chart or a zero sum game. US consumer spending has grown from $14 to $19 trillion since 2020. $5 trillion in new pie!!
Your model of ads is: “I, a consumer, have decided to buy a bluetooth speaker, and the ads push and pull me towards particular brands”. But that’s not how ads work! Ads don’t just compete for fixed spending, they induce NEW spending. An ad can give a customer the idea of buying, and grow the market.
All that's telling you is the economy is not doing nearly as well as some of our metrics would have you believe.
Real wages are about the same as before, probably lower. Consumers are buying the same amount of stuff - no value has been created. Rather, the dollar has been devalued, much more than we're willing to let on.
There's real value, like actual physical goods, service and labor, and fake value. Fake value tries to proxy real value, but historically it's often way off.
Money is fake value. Stocks are even more fake value. It doesn't matter if your stock price is through the roof if you're not selling a product people want, for example. The product is the value, the stock price is people trying to approximate the value and future value.
> If you have a better understanding than the idiots in charge, then go be rich!
Doesn't work this way because most markets are dumb as rocks.
> If you have a better model for real value, you'll outcompete them.
Doesn't work this way because most markets are dumb as rocks.
Look, after a certain point you have to detach from what you're being told and look at the world around you.
Prime example: tobacco. For humanity, Tabacoo has a negative value. You should be getting paid to smoke. Why? Because it kills you, and that's very expensive.
But that's hard to measure, right? So we just sell the cigarettes and say their value is what they're sold for. But that's not their actual value.
Their actual value, in the real world, in your hands and in your lungs, is negative. That's not an opinion. That's objective. That's just what it is.
When you look around our markets, almost all products are like this to some degree. The value we're creating is not necessarily real value.
Ads are another prime example. Do they enrich the world? Do they help consumers? No. They have zero real value. They just move money around via manipulation. That's not my opinion. That's just the objective reality.
Eventually, the real world catches up to la la land. You can't just say "well do ads and you make money". When there's no more money to move around, then even our fake value estimates of ads approach zero.
Of course the tech savvy enterprises will use the best models. But the plumber down the road doesn't care whether she asks Gemini or ChatGPT about the sizing of some fittings.
Everyone is vastly, vastly overestimating advertising. Advertising is a side hustle, because the product is the main hustle.
I think your are the one vastly underestimate advertising.
Consumers can spend what they can spend. Not even 1 quadrillion dollars in advertising can change that. There is a hard, hard cap to the value of advertisement because of that. It's just how the thing works.
“Enough” for what exactly?
Furthermore, OpenAI has to make up for a ton of debt they are taking on. They've already lost $9B, and are planning on losing another $75B in the next 2 years. As such, they have a ton of digging to do to get themselves out of the massive hole they're digging.
First of all, your numbers a off by an order of magnitude at least: even GPT-5 can generate 1000 tokens for 1c, which is much more than a paragraph.
And then again that's why my entire argument revolved around the fact that OpenAI would need to stop aiming for the technological edge. Deepseek generates 25k tokens for a cent and it's still a gigantic model. I'd you use a model comparable in size to gpt-oss-120b you can even increase that up to 100-200k tokens per cent (going from 32GB worth of active parameters, 32B at q8 for Deepseek, to 4GB, 8B using MXFP4 for gpt-oss-120b). That would mean being able to serve more than 100 answers per cent spent on inference.
If they can serve .1c worth of ads per request, that's 90% gross margin for you.
- users want the best/smartest LLM
- the best performance for inference is found by spending more and more tokens (deep thinking)
- pricing is based on cost per token
Then the inference providers/hyperscalers will take all of the margin available to app makers (and then give it to Nvidia apparently). It is a bad business to be in, and not viable for OpenAI at their valuation.
I think they all have become sufficiently good for most people to stick to what they are used to (especially in terms of tone/“personality” + the memory shared between conversations).
At youtube's ad income rate (~$13/year), the current (but growing) ~800 million chatgpt users would add ~$10 billion. At facebook's rate (~$40-50/year) $32-40 billion. Potentially, an assistant would be more integrated into your life than either of those two.
The "audience retention" is the key question, not the profitability if they maintain their current audience. I've been surprised how many non-technical people I know don't want to try other models. "ChatGPT knows me".
The field is too young to know what will keep users, but there are definitely things that plausibly could create a lock-in effect. I mentioned one ("ChatGPT knows me") which could grow over time as people have shared more of themselves with ChatGPT. There's also pilots of multi-person chats, and the social elements in Sora. Some people already feel compelled to stick to the "person" they're comfortable talking to. The chance of OpenAI finding something isn't zero.
A lot of YT's growth at the time was word of mouth and brand among the population, which is currently ChatGPT's position.
Altman knows this and why he called code red. If OpenAI hasn't produce a fully new model in 1.5 years, how much longer can they hang on before people will turn to alternatives that are technically better? How long before they could feasibly put out a new model if they are having issues in pre-training?
Their brand is not ok based on what I've heard, certainly no moat
Still feels like ChatGPT is synonymous with the current wave of generative ai
Even if they aren’t the market lead and it’s main offering is being commodified
A lot of "normal people" are learning fast about ChatGPT alternatives now. Gemini in particular is getting a lot of mainstream buzz. Things like this [1] with 14k likes are happening everyday on social. Marc Benioff's love for Gemini broke through into the mainstream also.
[1] https://x.com/kimmonismus/status/1995900344224907500 [2] https://x.com/Benioff/status/1992726929204760661
But a lot of HN users use gmail, which has the same model. And there are plenty of paid email providers which seem far less popular (I use one). Ads didn't end up being a problem for most people provided they were kept independent of the content itself.
2. I’ve never seen ads on the Gmail webapp (It sure does data collection)
Ads could fund more quota or bigger models for users who don't wish to pay (and/or just make it more sustainable)
Google will almost certainly be doing this with Gemini, and if ChatGPT can't offer as much it leaves an easy reason for people to switch.
2. It does have ads in the default interface, though they're quite unobtrusive. You might also have a blocker. But yes, I suspect their size allows them to provide it mildly "at a loss" to support their ads elsewhere.
How can a model achieve this kind of stickiness? By "knowing you"? I don't think that's the same at all. Personally, one of the reasons I prefer Claude is that it doesn't pretend to know me. I can control the context better.
YouTube was ambitious for its time. "In 2007, YouTube consumed as much bandwidth as the entire Internet had in 2000" but they weren't believed to start breaking even until 2015.
^1 and free users are a large majority!
The same isn't true about ChatGPT.
Anthropic and Google provides a similar product, and switching to a better/cheaper platform is fairly easy as it only depends on you and not on others (content creators or friends) doing the same.
Inference is cheap because the final model, despite its size, is ridiculously less resource intensive to use than it is to produce.
ChatGPT in its latest form isn't bad by any means, but it is falling behind. And that requires significant overhead, both to train and to iterate on model architecture. It is often a variable cost as well.
And this is definitely not happening. They are covering training costs with investors money, and they can't really stop it without their competitors catching up
The question is, does OpenAI get value out of the exchange?
You touched on it ever so briefly: “as long as inference is not done at a loss”. That is it, isn’t it? Or more generally, As long as OpenAI is making money . But they are not.
There’s the rub.
It’s not only about whether you think giving them your money is a good exchange. It needs to be a good exchange for both sides, for the business to be viable.
I think that’s what they’re saying. OpenAI is selling you a $1 product for $0.2
Tokens are too cheap right now and nobody is working on a path to dial up the cost
This is literally what OpenAI is doing. They are bleeding cash, i.e. spending more than they earn. How useful it is to you is not relevant in the context of the sustainability. You know what is also super useful to some people? Private yachts and jets. It does not mean they are good for the society as a whole. But even leaving out the hollistic view for a moment - their business model is not sustainable unless they manage to convince the politics to declare them national infrastructure or something like that, and have taxpayers continue to finance them, which is what they already probed for in the last months. Out of interest, why would you want ChatGPT plan your trip to Japan? Isn't planning it yourself a part of the excitement?
you could have done all of this without a chatbot.
It’s an especially good analogy if there is no plausible path to positive gross margin (e.g. the old MoviePass) which I think is even less likely to be true for OpenAI.
obviously the nature of OpenAIs revenue is very different than selling $1 for $0.2 because their customers are buying an actual service, not anything with resale value or obviously fungible for $
For example: free shipping at Amazon does not have resale value and is not obviously fungible, but everyone understands they are eating a cost that otherwise would be borne by their customers. The suggestion is that OpenAI is doing similar, though it is harder to tease out because their books are opaque.
As for profits, I haven't looked recently, but IIRC profits are mostly:
1. AWS
2. Prime membership fees
The latter drives loyalty and therefore volume and predictability, which allows Amazon to e.g. operate their own mini-UPS in the quest to make money on most parcels. They also rolled back free shipping on everything over the years and use it more surgically and with minimum order sizes.
The value of an LLM isn't an LLM. That's entirely 100% fungible. The value is exclusively what it produces.
If other people can produce the same thing, your LLM value approaches 0.
If you hope that ChatGPT will be worthless because the underlying technology will commodify, then you are naive and will be disappointed.
If that logic made sense, why has it never happened before? Servers and computers have been commodified for decades! Salesforce is just a database, social media is just a relational database, Uber is just a GPS wrapper, AWS is just a server.
People pay money, setup subscriptions, and download apps to solve a problem, and once they solve that problem they rarely switch. ChatGPT is the fifth most visited website in the world! Facebook and Deepseek making opensource models means you can make your own ChatGPT, just like you can make your own Google, and nobody will use it, just like nobody uses the dozens of “better” search engines out there.
The problem is: suppose Google has an equivalent model (they do, but if you disagree, just pretend). Suppose they do. What then is OpenAI offering that makes its product more intriguing? Nothing. They have a chat interface. An intern can make a chat interface.
> ChatGPT is the fifth most visited website in the world!
To me, this is absolutely worthless information. That DOES NOT mean that ChatGPT is in the clear and nobody else will overtake them.
Your analogies really paint the picture here aptly. Salesforce is not just a database, it's a lot of stuff on top of it. AWS is not just a server, it's a lot of stuff on top of it. Uber is not just a GPS wrapper, it's a taxi service. That's a different thing.
ChatGPT... is just a model. What they add on top approaches zero. Because that's just how the technology works. It takes text and gives it to a model and then spits out the output. What more can you add onto that system, removing the model? Make it easier to input text? Make it easier to get output? Well that's truly trivial to do, and I would argue ChatGPT isn't even in the top 10 when it comes to that. Today.
Nobody wants the same thing but cheaper, or the same thing but marginally better. You either solve the problem first, or you lose. The first site to ever threaten the dominance of google.com is chatgpt.com! Why? Because it’s NOT just “google but better”, it’s an entirely new thing.
> To me, this is absolutely worthless information. That DOES NOT mean that ChatGPT is in the clear and nobody else will overtake them
Do you think chatgpt.com will be worse the 5th most visited website 5 years from now? I’ll gladly take that bet, let’s do $100, i’ll even give you 2:1 odds. Do you think openAI will be bankrupt in <10 years? Let’s bet $1000, hell I’ll give you 10:1 odds.
Chatgpt.com alone is clearly at least as valuable as instagram.com, soon to be as valuable as google.com, and long term more than either.