“It is not a moral good to be correct about how much AI can do.” Ed Zitron · The Tech Report
September 2026
Zitron had spent the segment arguing that a thin slice of heavy customers carries almost all AI revenue, and that boosters answer that argument by demanding you admit ChatGPT is better. This is his objection to the demand itself: it treats a paid software product as a position on right and wrong. He goes on to say they want the argument kept there so nobody looks at who is actually paying.
Transcript
It's just that easy. It's 1% now, now it'll be 20%. But that's kind of the thing. It's you have this contingent of customers who are spending lots of money who are likely the largest burden on your infrastructure, which means you can no longer suggest, oh, it's like the gym model where there are a bunch of seats that people buy, they don't use very much, but you're still pulling in the revenue. It means that most of your revenue comes from very OPEX heavy customers. And yes, when those go away, the demand for those GPUs you're renting out also goes away. And these are all take or pay agreements when it comes to the hyperscaler. So with Microsoft Core, we've et al., they are take or pay, which means you have bought this capacity whether you use it or not. You must buy the, you can't just say, oh, I'll go on the meter like an electricity bill. It's no, you say, I'm going to, I'm going to spend. I don't know if it's, I truly, I don't have insight into. Azure or Core Weave contracts, but it's going to be you are renting X critical IT load for Y months or Y years. So that doesn't matter if you have demand for all of those megawatts. It means you have rented them, you will pay for them. This is a big deal. And this groundbreaker piece I mentioned, I'm sure people have seen it, makes this point where this is the equivalent of during the great financial crisis when all of the adjustable rate mortgages switched. So to explain, so in the great financial crisis, a bunch of people got something called a teaser rate. Teaser rate was: okay, your first year is 1% or 4% or something quite low. Then after the teaser rate goes away, suddenly your mortgage goes up to 5%, 6%, 7%, in some cases, 10%. That happened, I think, 2007 was when that big wave hit. In this case, the teaser rate going away is going to be when you sign a compute contract, you go, okay, here's a little bit of money up front to secure that. Thank you kindly. And they go and build the data center. Next year, a bunch of compute capacity is going to come online, or these contracts are just going to start. And suddenly Anthropic and OpenAI are going to have to pony up a bunch of money. This is off the top of my head, rough estimates from UBS, Was Vargo, Al. I think we're talking at least $130 billion worth of compute contracts that they'll have to start paying for. Billions with Torwe, billions with Cerebrus. There's all sorts of money that's going to have to go out the door. Right now, they don't have to spend that money. Next year, it's going to hit them like a tidal wave. And so if they don't have the customer demand to use that compute capacity, they won't have the money to pay for that compute capacity. So they'll have to raise money either way. But it's all, there is a wall coming that they're barreling towards at high speed. And the reason that everyone's not freaking out, the reason that people are not even particularly alarmed is because right now everything's going fine. Nvidia sells GPUs. I mean, Broadcom missed on estimates for next quarter's guidance. I'm sure everyone can find a rationalization for that. My rationalization is the money isn't coming in. Also, the fact that Broadcom's largest customers are going to be Anthropic and OpenAI, that's also bad. So we just have all of these consequences waiting around the corner and everyone just saying, it's not a big deal. We'll deal with it. Like, stop thinking. Stop asking questions. Nothing bad ever happens.
I mean, so if we take that 80% from 1% statistic and we kind of extrapolate it out, what does it mean for the narrative that AI is this inevitable thing that has such broad demand and such broad use cases that you got to get on board or get in the bin?
It means that whatever AI is, you can put to the side. Most people are not paying for it. It means however useful you may think it is, and I do not think it's that useful at scale. Most people won't pay for it. Most companies, most enterprises, which is where the most of your revenue is going to come from. Sam Altman said in an interview with Alex Heath that once or twice involved journalism over a course of more than an hour, said this point where actually we have more enterprise customers than consumers right now. That's bad, man, because your enterprise customers are not spending that much. This is the only way this works. AI is not a consumer product. They're really trying to make it one. If enterprises won't pay, consumers will definitely not pay. So however useful you may think this is, and I really don't think it is. I think it makes the easy things easier, the hard things harder. It helped me fix a Minecraft thing for my kid. And for that, I am grateful. But again, it's like, how much did it cost to do that? I keep these things, by the way. I pay for like the 20 bucks subscriptions just to take a look occasionally to make sure something big hasn't changed. They can do some computer stuff if you allow them to run rabshot over your computer, which I don't trust. The point is, put aside what I say, most people won't pay for it. Most enterprises will not pay for it. They have signed $1.2 trillion worth of compute commitments based on the idea that everyone will pay for it, when really about 900 companies will pay for it, or a few hundred companies, an indeterminate amount of hundreds of companies will pay for this. And those companies are the only ones that will pay significant money. And we're talking millions, tens of millions, hundreds of millions. But again, within that subset of data, we're AI companies that are subsidizing their own users. So it's the answer is, whatever bigness this might have, it's not big in a revenue generating sense. And all of the achievements or improvements you've seen are because they've just crammed as much money into it. What other industries could have done incredible things if you sunk a trillion dollars into them? That is the question to ask. It is a. A ridiculous thing right now where AI boosters will come to you and they will talk to you like a religious fanatic and they will say, Admit it! Admit that ChatGPT is better! Admit to me, you must concede that ChatGPT is better. First of all, none of these idiots will concede a single damn thing about the finances, so screw them. Second of all, this is software you pay for. This is not morals. It is not a moral good to be correct about how much AI can do. On top of that, they're not correct on it, but it's not like liking or disliking a software product you pay money for is not a ideological thing. It's stupid to see it that way. And they want it to be like that so they can push this through so that you'll all stop thinking about one thing, which is that most people will not pay for AI. Even though it's got better in the last year, most people will not pay for it. Most people will pay for a subsidized subscription. But companies, so companies with real budgets, much larger budgets than the average household, than the average person, when given the chance, in fact, only given one option, which is to pay on the API rate, they mostly don't. They mostly don't. Most 1% of customers hold up everything else. And it's quite worrying, honestly, because they very clearly been telling us all how powerful and amazing this is and how widely adopted it is. Who cares if it's not really making money? And I mean, when I say making money as well, sorry, I can't believe I said this. These aren't profitable companies either. Even if they were profitable, this would be horrible. Even if they were like 30% gross margins with training, this would be horrible. Because it means that, just on a fundamental level, it means that it is impossible to grow from here. Like it's just, how do you make more companies spend more money on AI? The AI itself needs to give them real value, real tangible ROI. As we've been over, no one has proven that it has ROI. The companies spending a lot of money are probably just doing a bunch of coding. I've heard from large companies that are using it. I've talked to them. It is just people throwing everything in there. It's being like, how do I do document?
You know that feeling. Too many updates? Too many meetings? Too many things falling through the cracks. Monday.com was built for that gap. The AI work platform designed from the ground up for people and agents to work side by side and deliver more together. Drafting the brief, flagging risks, running a competitive scan, your agent can do it all. And the best part, anyone can build one. Even you. Try it. Create your first Monday agent today at Monday.com.
A BetterHelp ad. Matching with my therapist was very quick and easy. I just filled out a couple of questions. Before you know it, I was matched with a therapist. I just felt like a weight just was lifted off of my shoulders. I felt like it was a good match from the first time we talked. ButterHelp is there to meet people at their needs, giving them a space to feel safe. There's a stigma about therapy. But with BetterHelp, I need people where they are. Wherever you are, that's where BetterHelp begins. Visit betterhelp.com/slash random podcast to get started.
Pay testimonials. Results may vary.
More from The Tech Report
