Better Offline - Monologue: AI's Compute Demand Story Is A Lie

Episode Date: May 1, 2026

In this week's Better Offline monologue, Ed Zitron runs through how 70-80% of Amazon and Microsoft’s AI revenue and capacity is taken up by OpenAI and Anthropic, and how capacity constraints across the industry are a result of hyperscalers sacrificing their infrastructure to the least-profitable startups of all time.Premium newsletter out later today - AI’s Demand Story Is A Lie - save $10 off a year of my premium newsletter: https://edzitronswheresyouredatghostio.outpost.pub/public/promo-subscription/gzqwkv54e1 YOU CAN NOW BUY BETTER OFFLINE MERCH! Go to https://cottonbureau.com/people/better-offline and use code FREE99 for free shipping on orders of $99 or more. --- LINKS: https://www.tinyurl.com/betterofflinelinks Newsletter: https://www.wheresyoured.at/ Reddit: https://www.reddit.com/r/BetterOffline/  Discord: chat.wheresyoured.at Ed's Socials: https://twitter.com/edzitron https://www.instagram.com/edzitron https://bsky.app/profile/edzitron.com https://www.threads.net/@edzitron Email Me: [e-mail] omnystudio.com/listener for privacy information.

This is an IHeart podcast. Guaranteed human. Run a business and not thinking about podcasting. Think again. More Americans listen to podcasts than adds supported streaming music from Spotify and Pandora. And as the number one podcaster, IHearts twice as large as the next two combined. Learn how podcasting can help your business. Call 844-844-I-Hart.

Joy is essential and it's also elusive. But now, there's a new and exciting way to start your journey toward a more joyful existence. Joy 101. It's a new podcast hosted by me, Hoda Kotby. If you're craving inspiration to maximize your joy, tune into these candid, uplifting, and moving on-air chats. Open your free IHeart Radio app. Search Joy 101 and listen now.

Joy 101 with Hoda Kotby is presented by CVS. Hey, this is Chuck from Stuff You Should Know, and we're submitting our most sciencey episodes for your peer review with our new Stuff You Should Know doing science playlist. Out now. You want to know about Occam's Razor? Simplest explanation is usually the right one?

We got you covered. Wondered what chaos theory is ever since the first time you saw Jurassic Park. Well, come on down. So distill a nice pot of tea, everybody, turn down the gas on your Bunsen burner, and slip into your most comfortable lab coat and listen to the stuff you should know doing science playlist on the IHeart Radio app,

Apple Podcasts, or wherever you get your podcasts. June is Black Music Month, and on the Drink Chams podcast, we're speaking with the hottest names in the culture, like Sway Lee. Do you realize how legendary you are? I appreciate that. I'd be seeing it, but I'm like, man,

I still got, like, so much more to do. Like, Prince, he dropped like 30 albums. We dropped, like, five right now. Like, that's the rate we gotta be going. Yeah, that's a good attitude. No matter the era, Drink Chams brings you the biggest names and the most unfiltered conversations. Listen to Drink Chams from the Black Effect Podcast Network

on the IHeart Radio app, Apple Podcasts, or wherever you get your podcast. Allzone Media. Hello and welcome to this week's Better Offline monologue. I'm your host, Ed Zittron. It's earning season and a quarterly jaunt through the psychosis of financial analysts that have worked for years to prop up an AI industry with questionable returns, spending now over $750 billion in this year, I believe, on CAPEX for, well, no real return on investment.

In fact, all of them seem to be losing money. I know, I know, though, some of you are going to say, but these businesses had great quarters, but you'll know it's that said quarterly growth does not actually seem to be coming from AI when you read their results, where they actually share the AI ROIC return on invested capital. So while Meta and Google refused to actually explain their AI returns, because that would require admitting how bad they were, Microsoft revealed that it had $37 billion in AI revenue run rate, about $3.08 billion a month or so, and Amazon had $15 billion annualized or around one $1.25 billion a month.

Do not be fooled. These are bad numbers. Based on my own reporting, OpenAI spent $3.648 billion on inference in the third quarter of 2025 on Microsoft Azure, or around $14.4 billion on an annualized basis. While I only add

half of the fourth quarter numbers, I estimate based on them that OpenAI's annualized spend hit over $18.5 billion, or around $4.6 billion a quarter, but they ended 2025, and that's not accounting for things like or a two or a launch of its Codex coding platform. In total, this puts OpenAI an estimated $13 billion of spend on Azure, just on inference in 2025,

with billions of dollars more on training models. Now, Microsoft also accounted for 67% of Coreweaves $5.15 billion in 2025 revenue, or around $3.45 billion. And as all of that is used by OpenAI, I really think you need to think very, very, deeply about, well, how much AI demand there really is outside of two very large fail suns. I'm also confident that CoreWeave's computer is used for training OpenAI's models, as CoreWeve's announcement related to its direct deal with OpenAI specifically said it was contracted, and I quote, to power the training of OpenAI's most advanced next generation

models, and said capacity per semaphore was only available because Microsoft declined to extend its current agreement with CoreWeave to use that compute for OpenAI. Altogether, that puts OpenAI spend on Microsoft's services at over $18 billion in 2025, and it's easy to see how that could grow to over $24 billion on an annualized basis in the last quarter, or around $2 billion a month. Microsoft is OpenAI's primary cloud provider, and I estimate that OpenAI represents around 70% of its AI revenue, if not more, while taking up the majority of its AI infrastructure.

Otherwise, Microsoft's 20 million co-pilot 365 subscribers likely pay no more, than $7 billion a year. I'll get to my feelings about that in a bit. Now, as far as Amazon goes, things get a lot grimmer. In early April, per Reuters, Andy Jassy, their CEO, admitted that its cloud business's AI revenue run rate was more than $15 billion in the first quarter of 2026, which translates to about $1.25 billion in monthly revenue, or roughly 0.419% of the $298.3 billion in CapEx, Amazon spent so far on the AI bubble, or around 25% of the $5 billion Amazon just invested in Anthropic last week. I think it's also reasonable to assume that a large part, if not a majority, of that revenue comes from Anthropic. Per my reporting last year, Anthropics spent $518.9 million

on Amazon Web Services in September 2025, at a time when it had around $7 billion in annualized revenue, a figure that increased by 500%, if you believe Anthropic, which I don't know if I do, to $30 billion in annualized revenue since as of April. $518.9 million is about $6.2 billion in annualized spend, and I think it's fair to assume that its spend will have at least doubled to $12 billion in annualized AI revenue for Amazon, or around 80% of the AI revenue. I also want to address this continual claim

that capacity constraints in AI are proof of some incredible demand. In reality, the reason that there's so little available computers, is that the majority of it is taken up by either Open AI, Anthropic, or hyperscalers running their own services with the dregs of what's left behind. Think of it like seeing a very, very large man on the bus, taking up four seats. There are absolutely capacity issues on this bus, but they're not a result of everybody wanting to ride it. Now, Epoch AI, who I have had my differences with, estimate, and I'm using these numbers

because there's fuck all out there as far as compute capacity goes, they estimate that Microsoft had around 2.02 gigawatts of compute capacity available at the end of 2025, a time when OpenAI said it had access to around 1.9 gigawatts of compute. As Microsoft is OpenAI's primary cloud provider and Oracle had yet to stand up more than 100 megawatts of Stargate Abilene, I think it's safe to assume that Open AI takes up 80 to 90% of Microsoft's available capacity. Similarly, EPOQ estimates that by end of 2025, Amazon had access to an estimated 1.67 giga watts of capacity, and I estimate over 75% is taken up by Anthropic. Amazon's Anthropic dedicated project Rainier Data Center accounted for 500 megawatts by the end

of that year, around a third of that capacity. And considering Anthropic spend, I think it's reasonable to believe it's taking up the vast majority of what's available. Of course there are capacity issues. You're selling all the capacity to one fucking company. Epok also estimates that Google had around 2.95 gigawatts. A remarkable figure that also relies on literally doubling its capacity since the beginning of 2025. Nevertheless, it likely uses a lot of its GPUs for Anthropic too, but it's hard to estimate how much, as we have very little insight into Anthropic spend on Google Cloud, outside of its promise to use, and I quote, up to 1 million TPUs with over a gigawatt of capacity coming online in 26 from the end of last year,

and a month ago, when it said it would use, and I quote, multiple gigawatts of next generation TPU capacity starting in 2027. I also think it's fair to assume that large swathes of Google's computer infrastructure is taken up by Google Gemini being shoved into literally every single part of Google's products. It's everywhere. It's like a fucking weed. It's horrible. Google also refuses to break out its AI revenues, but considering it owns around 13% of Anthropic, I also think it's fair to assume a large capacity amount goes to them. All right, you know what? I tried to be nice. I tried to keep my cool, but this whole situation's beginning to piss me off. I'm tired of hearing people online, do this victory lap thing about how AI demand is there when it isn't.

And when I started these modologues, I was told I'd be allowed to rent, and I haven't really let myself off the chain, so I think it's time to do so. The story of massive AI demand is a lie. Nearly a trillion dollars, if not more, annihilated to create the largest circle jerk of all time. Venture capitalists and hypers feed money to open AI and anthropics so that venture capitalists can feed money to startups to feed money back to open AI and Anthropic to feed back into hyperscalers for GPU compute, at which point they turn to Jensen Huang of Nvidia and buy more GPUs. While it might seem tempting to credit these men as geniuses for creating companies specifically to feed themselves revenue, well, but to keep up with this k-fabe of doing AI to

substantiate the build-out means that they've had to massively over-commit to the bit, even though the only two meaningful businesses in AI appear to be anthropic and open AI, and that's only because they're effectively intellectual honeypots for the entire industry. Outside of those two, the only other competitive AI businesses are those of Amazon, Microsoft, and Google, two of which have an annualized revenue of around 6% of their capital expenditures so far. Google's AI business is so booming that it refuses to break it out, and while it nebulously claims that AI is creating growth, it's not really clear how and they're vague about it because analysts and the media are ready to swallow the narrative as long as number go up. That's why Google doesn't break out their AI revenues by the fucking way.

Amazon and Microsoft had their hands forced by the markets after their stocks tumbled and fucked up by sharing their AI revenues. Amazon's $298.3 billion in CAPEX has successfully created an AI business that, and more than a quarter of a way to a trillion dollars has successfully managed to make $1.25 billion a month. That's fucking pathetic. If we had investors with IQs above room temperature, they'd run Andy Jassy out of Arlington like fucking Shrek. But this is the AI bubble, where everybody is wrong because once everybody admits what's actually happening, they're going to have to admit they all sounded insane for years. $1.25 billion a month.

After nearly $300 billion of capex, unbelievable. What a joke. And like I said, more than 75%, if not 80% or more of that, revenue is from fucking Anthropic, who Amazon just agreed to fund up to another $25 billion. At some point, it's just Microsoft Google and Amazon handing themselves money or handing it to Nvidia to justify why they've burned over $800 billion fucking dollars, creating some of the most wretched businesses of all time. I'm sorry, wow, Satya, you managed to get up to 20 million paying Microsoft 365 co-pilot subscriptions. 600 million. 600 million.

billion dollars a month in revenue, not profit, and all it took was you investing more than $13 billion in Open AI, forcing large language models into every one of your products in a way that borders on harassment, and about, what, 240-something billion dollars in CAPEX, as well as laying off tens of thousands of people and savaging the Xbox brand. Whoopty fucking shit, man, you should be ashamed of yourself. Pride month, Toronto. Pride is an opportunity for you to create your own space, to celebrate your existence. IHart Radio is proud to be an official sponsor of Pride Toronto Festival and we won't stop.

Celebrate Pride. Turn up the love and listen to IHart Pride Canada. Your 24-7 radio stream and the only playlist you need for your Toronto Pride celebrations. Pride is so great because it gives a whole bunch of people this visibility that they've never had before. We have a ton to celebrate Toronto. Happy Pride. IHart Radio. Run a business and not thinking about podcasting, think again. More Americans listen to podcasts than ads supported streaming music from Spotify and Pandora.

And as the number one podcaster, IHearts twice as large as the next two combined. So whatever your customers listen to, they'll hear your message. Plus, only IHeart can extend your message to audiences across broadcast radio. Think podcasting can help your business. Think IHeart. Streaming, radio, and podcasting. Call 844-844-I-Hart to get started. That's 844-Ehart.

Hey, I'm Hoda-Cotby host of the podcast, Joy 101. One with Hoda Kotby. Together, we're going to have meaningful conversations with the world's most fascinating people. Like when actress Olivia Munn shared how she overcame fierce health challenges. I've gone through breast cancer and then helped my mother through breast cancer. And that was more difficult. There's a lot of people who understand postpartner depression. I was not prepared for postpartum anxiety.

Listen to Joy 101 with Hoda Kotby on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts. Here at the Happiness Lab, we are serving up. some hot takes for the summer. Big ideas that just might reshape how you think about your well-being. Like, we've been thinking about the loneliness epidemic all wrong. You can be lonely in a marriage.

You can be lonely at a party. I don't think loneliness is actually about solitude. Loneliness is about something much bigger. Or that we should get rid of small talk altogether. We talk about current events. We talk about what you do for a living, but not do you love what you do for a living. Is this your dream job?

Or that the mental health crisis isn't what we think it is. and that kids today are doing better than we assume. It was really disorienting for us as researchers to be so wrong about our hypothesis. We are so scared that we are going to underreact to a severe challenge that we tend to overreact. For more surprising ideas backed by psychological science, check out our new series, Happiness Hot Takes. Listen to the Happiness Lab with me, Dr. Laurie Santos, on the IHeart Radio app, Apple Podcast, or wherever you get your podcasts. Open AI is, in and of itself, a kind of psychosis generator. It was the first thing in a long time that felt like a new thing since the iPhone for the people that entirely obsess over growth.

It was the panacea for the entire tech industry, creating a new way for business idiots to spend money on infrastructure, a new thing for consultants to scam people with, a new series of things to be an expert in as a journalist, a blogger, or just the scam artist, all wrapped up in something that could also be a consumer product, an enterprise software product, and a new kind of API to attach to other enterprise software products and then charge more money. In theory, OpenAI success would lift everything at once, hardware, software, and even adjacent fields. It promised to both democratize access to creating software, while also heavily reinforcing existing power structures to the point that every dollar inevitably ended up in the magnificent Seven's pocket or asshole. The problem is that the system needed to actually work one

day. It needed to eventually make more money than it cost. Every single one of these companies is talking about AI nonstop, and not one of them can show you a profit. The only thing they can do is tell lies of omission by saying AI helped boost everything. And when you ask for specifics, well, the results are either tepid or so secretive you'd think they were hiding a dead body. The only reason Google, Amazon and Microsoft are being tolerated that their current access is because their non-AI segments continue to grow through endless price increases and in shittification, and its other business units, by which I mean Open AI and Anthropic, are yet to run out of money to feed them. Sorry, by the way, I just don't know what Meta's doing. I haven't talked about it very much because

I don't think Meta knows what Meta is doing. Every so often they bury a fact in one of their blogs about they saw a 3% increase in something related to AI. Then they promised to burn another $170 billion in a year, and again, we can't really tell why. They also lost another $4 billion on reality labs in the quarter, by the way. There should be a legitimate criminal inquiry into where this money is going. This is very strange. I think they're at like $80, $86 billion, and all we have to show for it is the Metaverse and Pervert glasses. Meanwhile, SpaceX and Mr. Musk is rushing to have the strangest and largest IPO of all time. All as daily stories leak about billions of dollars of losses and whatever the fuck that deal with Cursor is, apparently SpaceX will buy it for $60 billion

or pay it $10 billion to fuck off. I think what actually happens is a secret third thing. SpaceX funds it for a bit, then there's a falling out between Musk and Michael Truel of Cursor, and then the company either rushes another acquisition or just dies. Remember, Elon Musk killed Cursor's funding round that was in place by doing this acquisition, and he can't buy this company before SpaceX goes public. Elon Musk took fucking Open AI to court. Do you think he'll care about killing Cursor? Who's going to sue them? The remains?

It's just a fucking rapper company. I'm so fucking tired of it. Anyway, that lawsuit with OpenA.I., it's a real alien versus predator situation. And if I'm honest, I found the whole thing a little boring. A duo of dullards, shoulder-barging each other to see who can run a company that neither of them can really describe, because neither of them do anything other than pontificate and take credit for other people's work.

too dumb bitches going, exactly. If this breaks open AI, I'll be surprised. But if it does, it would be extremely fitting that fucking Elon Musk would accidentally destroy the AI industry, like Mr. Bean sitting down on a button that launches a nuke. If I'm wrong here, it'd be very funny. I'm just not giving it much hope. Nevertheless, this entire industry is only made possible by the global K-Fabe circular economy of taking every single sign is good for AI and ignoring every single possible.

glaring warning sign and the hopes they'll go away. That is the entire strategy. That's all it is. It's just people being like, well, the bad things don't matter. They do matter. Last week, Microsoft said it's shifting GitHub copilot to token-based billing, meaning that people will have to pay the actual cost of their models. This is effectively killing the product as people know it and invalidates every single story about its growth, revenue or otherwise, ever written. I don't see anyone on writing those, by the way. No one gives a fuck. And to give you some context about the scale of how big GitHub co-pilot was or is, until they stop subsidizing people's compute,

because right now when you use GitHub Copilot, Microsoft's just paying the models for you with them. Anyway, GitHub Copilot is the second largest customer of Anthropics models and was only that large because it was subsidizing the compute spend of its customers. A lot of that money spent on Anthropic. This is bad. And really, that's the only way to build any kind of AI business. Google and Amazon realized that AI,

revenues are contingent on the continued survival of Anthropic, and Amazon and Microsoft's revenues are contingent on Open AI and Anthropic. They know that if these companies die, they're going to lose billions of dollars in revenue, but also they have to compete with them for fear that they'll be seen as falling behind their horrible progeny, and thus they're incinerating their brands and endlessly pontificating about the power of AI while spending nearly a trillion dollars on CAPEX, almost entirely to make sure their competition, which is also their customer, which is also their welfare recipient doesn't die because if they die, they lose a bunch of revenue, but they need to spend way more money than they'll ever make to justify that. Yeah, it's a mess and a mistake,

and eventually one of these companies is going to grow tired of it. Microsoft was already billions under the analyst estimates for CAPEX. They're moving to token-based billing on GitHub copilot. They claim to invest in Anthropic in February, but didn't mention it in their 10-Q, their earnings, in any way, shape or form, which means no money was sent. At some point, these fuck nuts are going to be forced to reckon with what they're doing. Until then, we'll have increasingly more frenzied and ejaculatory statements about AI demand that fail to match with reality. I truly think that it's going to be like this, if not crazier, until one day when the music suddenly stops. Somebody's going to blink. Somebody is going to take a step back and give everybody else

permission to stop too. Maybe perplexity, lovable, rip-lit or cognition dice. Maybe Microsoft shifting GitHub co-pilot to token-based billing on June 1st inspires others like Anthropic to follow suit. Maybe AI token austerity begins at Microsoft, Meta, or any other large company burning 10 plus percent of their head count on AI tokens. Maybe Nvidia fails to inspire the markets in just the right way, or the fact that data centers are not opening fast enough to a fully digested 2025's GPUs finally catches up with the economic mismatch that Jensen Huang always beats and raises expectations. And that really is the strangest thing. At the current rate of sales, it's taking more than six months to install a single quarter's worth of GPU sales. At this point, it's

obvious that there are warehouses at these fucking things. It just isn't obvious whether they're in ones owned by hyperscalers or the Taiwanese ODMs, original design manufacturers like quantum computing in Foxcon that build their servers. And last quarter, it looked like quantum computing's inventories were growing by billions of dollars. Is that good? Anyway, none of this makes sense. It hasn't from the beginning.

It's the largest bubble in history and has reached such an intellectual and financial scale that many have taken science on it in a way that will be completely impossible to walk back if they're wrong. As things deteriorate, expect them to cling to their mythologies tighter and become more and more agitated and make vulgar threats and weird comments that don't really make sense, but nevertheless quite verbose. And really, we've never seen anything like this in our lives. You realize that open AI doesn't make any sense, right? It can't survive without eternal subsidies. None of these businesses can. Every single AI business you see is unprofitable,

and none of them have a path to break even, let alone sustainability. And to make matters worse, it appears the only way to create demand was to literally fund it yourself. These men believe they've created perpetual energy. What they've actually done is shit their pants and set their houses on fire. Anyway, catch you next week. Joy is essential and it's also elusive, but now there's a new and exciting way to start your journey toward a more joyful existence. Joy 101.

It's a new podcast hosted by me, Hoda Kotby. If you're craving inspiration to maximize your journey, tune into these candid, uplifting, and moving on-air chats. Open your free IHeart Radio app. Search Joy 101 and listen now. Joy 101 with Hoda Kotby is presented by CVS. Hey, this is Chuck from Stuff You Should Know, and we're submitting our most sciencey episodes for your peer review with our new stuff you should know doing science playlist. Out now. You want to know about Occam's Razor? Simplest explanation is usually the right one? We got you covered.

wondered what chaos theory is ever since the first time you saw Jurassic Park well come on down. So distill a nice pot of tea, everybody, turn down the gas on your Bunsen burner and slip into your most comfortable lab coat and listen to the stuff you should know doing science playlist on the IHeart Radio app, Apple

podcasts, or wherever you get your podcasts. June is Black Music Month and on the Drink Chams podcast we're speaking with the hottest names in the culture like Sway Lee. Do you realize how legendary you are? I appreciate that I'd be seeing it, but I'm like, man, I still got, like, so much more to do.

Like, Prince, he dropped, like, 30 albums. We dropped, like, five right now. Like, that's the rate we got to be going. Yep, that's a good attitude. No matter the era, Drink Chams brings you the biggest names and the most unfiltered conversations. Listen to Drink Chams from the Black Effect Podcast Network on the Iheart radio app, Apple Podcasts, or wherever you get your podcast. Hey, it's Alec Baldwin. This season on my podcast, here's the thing.

I talk to composer Mark Shaman. the hang. It's the pleasure of hanging out with the people that you're with. You know, Rob and I was always a great hang. And director Morgan Neville. Film School teaches you all the wrong things about making documentary. What do you want to say? Documentary is all about your ear. What do you hear? I feel like my job is listening really, really hard. Listen to Here's the Thing on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts. This is an IHeart podcast. Guaranteed Human. Thank you.