Following earnings this week that saw tech giants like Microsoft and Amazon report aggressive AI spending plans, EZ Primary Research CEO Ed Zitron gives his more skeptical take on the industry. In speaking about OpenAI and Anthropic Zitron argues "everyone is buying into these stocks because they believe all of that CapEx is going towards diverse and spread out AI demand, when in fact, what it's actually doing is helping create infrastructure for two unprofitable, unsustainable companies."
--------
Watch Bloomberg Radio LIVE on YouTube
Weekdays 7am-6pm ET
Saturday & Sunday 7am-10am ET
WATCH HE...
What Ed Zitron said
Written from the verified transcript and checked against it. Every figure links to the moment it was said.
Ed Zitron argued that AI hyperscaler revenue is dangerously concentrated in OpenAI and Anthropic, citing UBS estimates that Google Cloud's revenue from these two firms will be 27% this year and 48% next year, over $124 billion annually. He said Microsoft's Intelligent Cloud growth in 2025 was 69% from OpenAI; without it, growth would have been just 8%. Zitron called the situation 'a lie' and 'a scandal.' He flagged OpenAI's IPO delay to 2027 as a flashpoint and noted that data centers take 12 to 36 months to build. He cited Sightline Climate data on 190GW of planned capacity, requiring over $1.6 trillion in annual revenue. Zitron disclosed OpenAI lost $20.9 billion in 2025 and received $800 million from SoftBank for 'Crystal Intelligence.' He criticized circular financing, where Microsoft, Google, and Amazon built infrastructure for these companies without payment, with one Microsoft executive estimating $100 billion in costs.
Key takeaways
Google Cloud revenue from OpenAI and Anthropic is projected at 27% this year and 48% next year, over $124 billion annually.
OpenAI lost $20.9 billion in 2025; over $800 million of its revenue came from SoftBank for 'Crystal Intelligence' with no evidence of activity.
Microsoft Intelligent Cloud growth in 2025 was 69% from OpenAI; without it, growth would be 8%, barely beating inflation.
OpenAI's IPO is delayed to 2027, which Zitron calls 'lethal' and a likely first domino to fall.
Planned data center capacity of 190GW would require over $1.6 trillion in annual revenue, which two unprofitable customers cannot sustain.
Numbers and commitments
Figure
What it refers to
Type
At
27%
Google Cloud revenue from OpenAI and Anthropic this year
Well, I think investors have to ask a question right now: what am I getting into when I invest in Microsoft, Google, and Amazon? So UBS estimates that 27% of Google Cloud's revenue this year will be OpenAI and Anthropic, increasing to over 48% next year. That is a remarkable amount of money. It's going to be over $124 billion next year. Everyone is buying into these stocks because they believe all of that capex is going towards diverse and spread-out demand, when in fact what it's actually doing is helping create infrastructure for two unprofitable, unsustainable companies.
Well, that was what I was saying. So Barclays actually says that this year 13% of all revenue will be from both OpenAI and Anthropic, and next year will be 18% — it's a much bigger business than Google Cloud. Now, just to be clear, when I was saying 27% this year and 48% next year for Google Cloud, I meant both Anthropic and OpenAI.
Most people don't know that OpenAI is a large customer of Google Cloud. It's not a well-known fact. But this was actually mentioned by UBS, by Stephen Ju.
Well, I have serious questions about that. So in the year 2025, according to my own reporting on OpenAI's numbers, 69% of the year-over-year growth of Microsoft's Intelligent Cloud segment was actually from OpenAI. Without that, it would have only grown 8% year over year, which is barely beating inflation. And so everyone is being sold what I consider kind of a lie. It's honestly kind of a scandal.
So this goes back to — I feel like we have companies, the circular financing, the circularity of it all, and kind of creating demand for their products. So when does it start? When does the... I asked this earlier with a guest: when does the party end, in your view?
Yeah. So with OpenAI's IPO, I think that could be one of the flashpoints. Remember, this company was meant to go public this year. The fact that a month or two ago... and now the New York Times has reported that they are considering — they are delaying until 2027. That's lethal for a number of people. But OpenAI and Anthropic need continual flows of capital. They do not pay their bills out of existing cash flow. So when anything happens to that cash, I think that's the first domino to fall.
But then again, there's also the overall problem of data centers just not getting built very fast — taking about 12 to 36 months, depending on how small or large the data center is that's actually being built. And the problem is that everyone believes that AI is coming out of cash flow, that AI is coming out of just this diverse revenue base, when it's really not — it's extremely narrow. The Information reported a few months ago that 89% of the largest AI companies' revenue comes just from OpenAI and Anthropic. It's heavily centralized.
Does it have to be centralized? To some extent this is expensive to do, in terms of data center buildout and so on and so forth. And what's going to make generative AI really, really good is having access to lots of information. So doesn't it have to be, to some extent, concentrated?
Zitron, E. (2026, July 31). Zitron: "Everyone Has Been Sold a Lie" on AI [Interview transcript]. Bloomberg Podcasts. CEOInterviews.AI. https://ceointerviews.ai/interview/3108862/
MLA
Ed Zitron. "Zitron: "Everyone Has Been Sold a Lie" on AI." Bloomberg Podcasts, 31 Jul. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/3108862/.
BibTeX
@misc{zitron2026_3108862,
author = {Ed Zitron},
title = {Zitron: "Everyone Has Been Sold a Lie" on AI},
howpublished = {Interview transcript, Bloomberg Podcasts. CEOInterviews.AI},
year = {2026},
month = {jul},
url = {https://ceointerviews.ai/interview/3108862/},
note = {Speaker-attributed transcript with timestamps}
}
