Better Offline

Monologue: What The Hell Are The Hyperscalers Doing?

10 min
Jul 24, 20265 days ago
Listen to Episode
Summary

Host Ed Zitron delivers a scathing monologue arguing that hyperscalers like Google, Microsoft, and Amazon are engaged in a form of financial theatre, spending over $1.1 trillion in CapEx on AI infrastructure while masking stagnant organic cloud growth by recycling money through unprofitable AI labs like Anthropic and OpenAI. He presents data showing that without Anthropic, Google's cloud growth story collapses, and without OpenAI, Microsoft's cloud growth drops to a mere 8% year-over-year. Zitron warns of a looming 'subprime data center crisis' driven by over $500 billion in data center SPVs that require $1.6 trillion in annual compute revenue to be viable — more than double the entire global software market.

Insights
  • Google's dramatic cloud revenue growth (47.7% to 82.1% YoY over three quarters) appears almost entirely attributable to Anthropic's infrastructure spending, not organic enterprise AI adoption.
  • Microsoft's Intelligent Cloud segment growth would fall from 26% to approximately 8% YoY without OpenAI's $17.2 billion in Azure spending — revealing how dependent hyperscaler growth narratives are on a handful of AI lab customers.
  • Hyperscalers face a structural trap: cutting CapEx would unwind the entire AI investment trade, crashing GPU demand and exposing the lack of real end-market demand for AI compute.
  • The data center buildout is mathematically unsustainable — 190 gigawatts of planned capacity would require $1.6 trillion per year in compute revenue, against a global software market of under $800 billion.
  • Private credit funds backing data center SPVs are funded by pensions and insurance funds, meaning a demand shortfall could have systemic financial consequences beyond the tech sector.
Trends
Hyperscaler CapEx growth is outpacing cloud revenue growth, signalling diminishing returns on AI infrastructure investmentAI lab spending (OpenAI, Anthropic) is artificially inflating cloud revenue metrics, masking the end of the hypergrowth era for major cloud platformsA potential 'subprime data center' crisis is forming as speculative data center debt outpaces realistic AI compute demand projectionsMainstream media is beginning to shift from treating AI skeptics as contrarians to engaging them as credible analystsThe AI investment narrative is increasingly described as cult-like, with dissent treated as ideological heresy rather than legitimate analysisGPU and server supply chains (Nvidia, Broadcom, Quanta, Hon Hai) are exposed to significant demand risk if hyperscaler CapEx cycles turnEnterprise AI monetisation remains opaque, with hyperscalers refusing to break out AI-specific revenues — a sign of weak underlying adoptionOracle's financial position is flagged as deteriorating, suggesting legacy enterprise cloud players face compounding pressurePrivate credit markets are increasingly exposed to tech infrastructure risk via data center financing vehiclesThe gap between AI hype and measurable enterprise revenue is widening, increasing the probability of a market correction
Companies
Google
Criticised for raising CapEx to $195-205B while cloud growth appears driven by Anthropic, not organic AI demand.
Anthropic
Identified as the primary driver of Google Cloud's revenue growth via a multi-billion dollar TPU infrastructure deal.
Microsoft
Cloud growth shown to drop from 26% to ~8% YoY when OpenAI's $17.2B Azure spend is excluded.
OpenAI
Spent $17.2B on Microsoft Azure in 2025, representing 14.3% of Microsoft cloud revenue and 18% of its YoY growth.
Amazon
Mentioned as a hyperscaler contributing to the estimated $1.1 trillion collective AI CapEx spend.
Meta
Listed among hyperscalers whose GPU purchasing sustains the AI investment trade and server supply chains.
Nvidia
Flagged as exposed to demand risk if hyperscaler CapEx spending slows or reverses.
Broadcom
Cited as a supplier of high bandwidth memory for GPUs and TPUs, vulnerable to a CapEx pullback.
Quanta
Taiwanese server manufacturer identified as dependent on hyperscaler GPU orders to sustain its business.
Hon Hai
Taiwanese server manufacturer cited as part of the supply chain that would be hit by reduced hyperscaler spending.
Oracle
Flagged as a subject of an upcoming premium newsletter piece, with the host implying the company is in decline.
People
Ed Zitron
Host delivering the monologue, presenting financial analysis of hyperscaler AI CapEx and cloud revenue dynamics.
Sundar Pichai
Criticised for announcing $195-205B CapEx guidance while Google's AI revenue appears dependent on Anthropic.
Satya Nadella
Called out for Microsoft's cloud growth being heavily reliant on OpenAI spend, with host challenging him to appear on...
Dario Amodei
Referenced sarcastically as 'weird Dario', accused of avoiding the host's interview requests.
Sam Altman
Mockingly referred to as 'Clammy Samuel' and accused of refusing to speak to the host.
Larry Ellison
Mentioned dismissively in the context of Oracle's reported decline, subject of an upcoming premium newsletter.
Nick Suresh
Upcoming guest on Better Offline to discuss AI mania and its cult-like effect on global decision-making.
Quotes
"Hyperscalers are doing kayfabe to keep the crowd riled up, but everybody's slowly working out that the wrestlers aren't really hitting each other and everything is scripted."
Ed Zitron
"Without OpenAI, their cloud revenue growth was a mere 8% year over year. That's pathetic. That's abominable. That's barely keeping up with inflation. That's dogshit."
Ed Zitron
"For the 190 gigawatts of supposedly planned data centers, you'd need $1.6 trillion a year in annual compute revenue to make it worthwhile. And by the way, the global software market is under $800 billion a year."
Ed Zitron
"I deeply worry about what happens when the demand doesn't arrive, as these data center debt deals are often funded by private credit funds, which are in turn funded by both public and private pensions and insurance funds."
Ed Zitron
"Outside of the fourth quarter of 2024, Google's capital expenditures have outpaced cloud growth by at least 12% year over year for two goddamn years."
Ed Zitron
Full Transcript
7 Speakers
Speaker A

This is an iHeart podcast. Guaranteed Human

0:00

Speaker B

hey everyone, it's Kalpen. I'm inviting you to join the best sounding book club you've ever heard with my podcast, Hearsay, The Audible and iHeart Audiobook Club. Every episode I nerd out with amazing guests and dive into the best new audiobooks available on Audible. It's the book club for your ears. Listen to Irsay, the Audible and I Heart Audiobook Club on the iHeartRadio app or wherever you get your podcasts.

0:05

Speaker C

Thinking about selling your car? Start with a real offer worth real money from CarMax. Then take seven days to compare your offer and when you're ready, accept your offer and get paid. CarMax will even pick up your car. That's it. No haggling, no hassles, no pressure. Just a real offer worth real money from CarMax. Get all the details and a real offer today@carmax.com wanna drive car at home? Pickup not available in all locations. Distance restrictions and a fee may apply.

0:34

Speaker D

This is Chelsea Handler from Dear Chelsea, Traveling with Doug is one of my favorite things. But when you're on tour like I am, bouncing between 35 cities, feeding your dog can get complicated. And that's why I started bringing just fresh from Just Food for Dogs. It's real fresh dog food made with human grade ingredients that he's been obsessed with. But it's shelf stable, so I don't need a freezer or a cooler I or a whole travel operation just to feed my dog. Doug loves it and I feel good giving it to him. So go to justfoodfordogs.com and get 50% off your first order.

1:04

Speaker E

Amazon Health AI presents painful thoughts why

1:37

Speaker C

did I search the Internet for answers

1:42

Speaker F

to my cold sore problem? Now I'm stuck down a rabbit hole filled with images of alarmingly graphic sores

1:44

Speaker C

in various stages of ooze.

1:51

Speaker F

I can clear my search history, but I can never unsee that.

1:54

Speaker E

Don't go down the rabbit hole. Amazon Health AI gets you the right care fast. Healthcare just got less painful.

1:59

Speaker G

Superstition, fear, and jealousy. Welcome to this week's Better Offline Monologue. I'm your host, Ed Zitro. Better Offline as ever. Please subscribe to the newsletter. We're running a one year anniversary deal right now. The link will be in the episode notes and we're giving you 10 bucks off the annual rate for life. And even if you die and come back to life, I'll give it to you anyway. The deal runs through midnight on Sunday. But if you happen to hear this after this point, email me at ezetteroffline.com or ez betteroffline.com for the British and Canadians in the audience and I'll sort you out. I'm generous like that and I love my premium mates. Really some of the best analyst work out there. It's very deep. I've got the terminal going now, pulling all sorts of data. The many numbers I read every week are driving me truly insane. And with that in mind, today's monologue starts with a simple question. What the fuck are hyperscalers doing now? I realize I've asked this question maybe 100,000 times in the last two years, but it's begun to get so incredibly stupid that I'm actually a little exasperated. McKinsey spawn and Google CEO Sundar Pichai announced in Wednesday's earnings that Google would raise its year end capital expenditures to somewhere between $195 billion and $205 billion, after burning $80.6 billion in the first half of the year. And this suggests that Google will spend another $114 billion by the end of 2026. They're going to be doing like 55, 60 billion dollars a quarter. It's fucking insane. It's fucking stupid. I'm sorry for swearing, but it's absolutely deranged. What are you doing Sundar? You've already spent $288 billion in CapEx since 2022, and all you've got to show for it is Several also ran LLMs and one indeterminately large customer in Anthropic, who you've also invested $13 billion in. Now one answer might be that Anthropic helped juice Google Cloud's growth by 82% year over year in this quarter. But considering that Google doesn't actually break out, how much of that is from AI or from Anthropic or really from anything? It's difficult to say what exactly this other than turning its free cash flow negative for the first time since Google went public. While I can't say for sure how much Anthropic is spending, I do believe it's taking up the vast majority of Google's AI infrastructure and representing most of its cloud growth. Google Cloud was, up until fairly recently, the last year or so, a much smaller business than Microsoft Azure or Amazon Web Services have been growing somewhere between 22% and 30% year over year, hitting 33.9% year over year growth in the third quarter of 2025. That very same quarter, Google and Anthropic Announced a deal to expand its use of TPUs, which is Google's special GPUs in a contract worth, and I quote, tens of billions of dollars, mysteriously. Can't imagine how Google's next three quarters of cloud revenue growth showed 47.7%, 63.4% and 82.1% year over year growth. Whatever could that be? Where did that come from? I can't work it out. Can you, dear listener?

2:11

Speaker E

What?

5:19

Speaker G

What could it possibly. It's anthropic. It's obviously anthropic. Except in those same quarters, Google's capital expenditures went up 95%, 107% and 100% year over year, absorbing every dollar of cloud revenue and then some. In fact, Outside of the fourth quarter of 2024, Google's capital expenditures have outpaced cloud growth by at least 12% year over year for two goddamn years. And if your pushback is, well, Google spending that capex on something else. No, they're not. They're not. Maybe they got a couple Billy coming in from Google workspace and cucking their various subscribers with annoying things. It's coming from anthropic. Google is not making that money from Gemini. Your ass must be crazy if you think that's. That's the point. But all of this begs the question, when exactly does all of this pay off? How does it pay off? What does the payoff even look like? And why am I one of the few people asking these questions? Hyperscalers have now spent, I estimate, about $1.1 trillion. I haven't seen Amazon and Microsoft's quarters yet, but I'm going to guess they spent a good amount on capex. They now spent 1.1 trillion bloody dollars on capital expenditures. And when you ask them to tell you how much revenue they made from said expenditures, they look at you like that meme of Brendan Fraser from the whale. Anyway, here's a fun fact for you. Thanks to my own reporting on OpenAI's audited financials, we know that it spent $17.2 billion on Microsoft Azure in calendar year 2025, a year when Microsoft's Intelligent Cloud segment made $120.4 billion in revenue. The segment grew as a result, 26% year over year. I'm a curious little critter. So I ran the numbers and OpenAI's revenue represented 14.3% of all Microsoft's cloud revenues for that period. And to be clear, it's calendar year 2025. They have a fiscal quarter thing. But I went and worked it out. But this means that OpenAI's money represented 18% of the year over year growth of Microsoft's cloud. Without it, their cloud revenue growth was a mere 8% year over year. That's pathetic. That's abominable. That's barely keeping up with inflation. That's dogshit. Satya. Satya. You should be ashamed of yourself. Satya, come on my show so I can ask you very reasonable questions. I'll be nice, I swear you won't. You're scared just like the rest of them. Just like Clammy Samuel1 Clamule, your public relations people said you're too scared to speak to me. Well, they said you wouldn't speak to me and I know you're scared. Same with weird Dario. Where are you boys? You're too scared of better off lime. Anyway, enough of the Trump voice though. There is actually a reason that hyperscalers are still spending all that CAPEX, OpenAI and Anthropic allow hyperscalers to boost growth by feeding their own money back into their businesses. A short term solution to the long term problem of the end of the era of hypergrowth. One of the reasons they're so shy about AI revenues is that they know it'll become obvious where those revenues are coming from. Two unprofitable and unsustainable AI labs and a little trickle, little spurt from AI services they dump on their enterprise customers through confusing subscriptions, upsells and just adding AI to random things and being like look folks, don't we love this? Don't you all love this, you little shits. And hyperscalers also know that the second they cut their capex spend, the AI trade unwinds. If Microsoft, Google, Amazon and Meta aren't buying hundreds of billions of dollars of GPUs a year, that means Taiwanese server manufacturers like Quanta and Hon Hai won't sell as many serv servers and thus won't need as many GPUs. Broadcom and Nvidia won't need as much high bandwidth memory for those TPUs and GPUs and data center speculators across the world will suddenly realize that they conflated hyperscalers being stupid with their money with the existence of massive amounts of demand for AI compute. Put another way, hyperscalers are doing kayfabe to keep the crowd riled up, but everybody's slowly working out that the wrestlers aren't really hitting each other and everything is scripted. I also should add I can just anticipate I'm going to get an email from someone being like, well, actually this is good. The capital expenditures are good because it showed growth even though it's short term. That's a silly, silly perspective to take, considering they're feeding all the revenues back into the business in the capex for stuff that won't be useful in the future. Mathematics, my friend. Try it sometime. Anyway, it is a funny old time for me right now, because I feel like I'm on three or four podcasts or TV shows or interviews a week as the system appears to be trying to cover its arse by having me on to explain stuff I said two years ago, but a little more insistently and clearly than before. While the mainstream media seems hesitant to fully commit to the graveyard smash of the AI bubble, it feels as if there's a conscious attempt to represent my work as more than just the other side, or a hater, or a skeptic or a cynic, which is both edifying and enraging because we are now so far puffed past the point of no return. I mean, yeah, we really can't avoid what's coming. As I wrote in my newsletter this week, I think there's a subprime data center crisis brewing right now with over $500 billion in data center SPVs that can only ever make investors whole if there's hundreds of billions of dollars of AI compute demand in the next two years. And to be clear, when I say hundreds of billions of dollars demand, I mean a year. I mean literally a year. I ran the numbers and and for the what I think is 190 gigawatts of supposedly planned data centers, you'd need $1.6 trillion a year in annual compute revenue to make it worth worthwhile. And by the way, the global software market is under $800 billion a year, so don't know where that fucking money's coming from. And no one else seems to really want to think about it. And I deeply worry about what happens when the demand doesn't arrive, as these data center debt deals are often funded by private credit funds, which are in turn funded by both public and private pensions and insurance funds. Anyway, do read the free newsletter to learn more. It is free, this one. I have a premium one coming out today that will be about, well, the fact that Oracle's dying, which, I mean. Hey you in the back. Stop laughing. Stop it. Poor Larry Ellison. Poor. Who gives a fuck? In all seriousness though, it's going to be a banger. The free one's a banger. There's bangers everywhere. And I'll be back next week with my good friend Nick Suresh to talk about how AI mania is eviscerating global decision making. Creating a cult like environment where one's participation in AI is almost a religious act. And if you talk badly about AI, it's considered a carnal sin or a cardinal sin, one of those two. Sure one of you will correct me on this one. Nevertheless, it's a really he it's based on this incredible piece he wrote. I can't wait for you to read it. I can't wait for you to me talk to him. I don't know. Can't wait for you to consume my content. What a peculiar thing to say. Anyway, I'll catch you next week. Zitron out.

5:20

Speaker A

Think about it. When it comes to your own food, you can walk into a restaurant, see what's being made, and know exactly what you're getting. But with dog food, most brands keep that completely hidden. Just Food for Dogs does the opposite. They've built their entire brand around open kitchens. You can actually walk in and watch them prepare your dog's meals with real human grade ingredients like chicken, beef, carrots and peas cooked right there in front of you. No mystery, no behind the scenes, you're not allowed to see. That kind of transparency is rare in the pet food world, and it's a big reason they've become the number one vet recommended fresh dog food, earning trust from pet parents who want to feel confident in what they're feeding. When a brand is willing to show you exactly how your dog's food is made, it says a lot about the care, quality and standards behind every meal they produce. Nothing to hide, Everything to love, go to justfoodfordogs.com and get 50% off your first order.

12:23

Speaker F

You know what quality feels like. You can see it in the way a fabric moves, recognize it in a flawless fit, and appreciate it in the details that make our styles unique. It's the standard Coldwater Creek has honored for over 40 years, derived from a rich Mountain west heritage and designed for today in styles that are distinctively Coldwater Creek. For a wardrobe you can count on season after season, visit coldwatercreek.com shop new arrivals and save 15% on purchases. $75 or more with Code Iheart, a

13:23

Speaker E

burst pipe, a dead water heater, the AC calling it quits. Who do you call? Homeserve is an easy way to handle unexpected home repairs with plans covering stuff basic homeowners insurance usually won't. Instead of scrambling for a contractor, you make one call to get the repair process started. Join the millions of customers who trust HomeServe right now. Go to HomeServe.com podcast for 50% less your first year. That's HomeServe.com podcast savings compared to renewal price void in Florida.

13:54

Speaker F

The Hulu Original Series Furious is now streaming on Hulu and Hulu on Disney. Starring Emmy Rossam as Alice Black, Furious follows a rookie FBI agent on the hunt for Catherine, played by Lola Petticrew, a mysterious and calculating female serial killer. While Alice upholds justice, Catherine kills for it, terrorizing the rich and powerful men of New York in her pursuit of vengeance. With secrets that change everything, Alice discovers there's a thin line between hunter and prey. Watch the Hulu Original series Furious on July 27, streaming on Hulu and Hulu on Disney for bundle subscribers. Terms apply.

14:23

Speaker A

This is an iHeart podcast. Guaranteed Human.

15:02