Ed Zitron: Big Tech AI Boom Is Built on a Scandalous Lie

Speaking on Bloomberg Businessweek Daily, EZ Primary Research CEO Ed Zitron delivered a scathing critique of the current artificial intelligence market structure. According to Zitron, Wall Street investors buying into hyperscaler tech stocks under the impression that massive capital expenditure reflects broad commercial adoption are being sold a profound misrepresentation. In reality, Zitron argues, cloud giants are pouring billions into infrastructure to subsidize two heavily unprofitable, unsustainable startups: OpenAI and Anthropic.

Who Is Ed Zitron

Ed Zitron is a prominent technology analyst, reporter, and the founder of EZ Primary Research. Known for his rigorous investigative work into financial filings and corporate claims, Zitron has become one of the most vocal critics of Silicon Valley hypecycles. His detailed analysis of internal numbers and audited balance sheets has consistently challenged tech industry narratives, particularly surrounding generative AI unit economics.

The Concentrated Cloud Revenue Fallacy

Wall Street has pushed stock valuations higher for Alphabet Inc. (NASDAQ:GOOGL), Microsoft (NASDAQ:MSFT), and Amazon.com Inc. (NASDAQ:AMZN) on the promise that compute capital expenditures serve a vast array of corporate buyers. Zitron pointed to institutional research from UBS showing that 27% of Google Cloud revenue this year, climbing to over 48% next year, comes from OpenAI and Anthropic combined, accounting for well over $124 billion in projected revenue.

The full discussion can be found on Bloomberg Podcast's YouTube channel.

A similar dependency appears at Amazon Web Services. Barclays research indicates that 13% of AWS revenue this year and 18% next year will derive solely from OpenAI and Anthropic. Zitron stressed that most market participants remain completely unaware that OpenAI operates as a major customer for Google Cloud alongside Microsoft Azure.

"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 is actually doing is helping create infrastructure for two unprofitable, unsustainable companies." - Ed Zitron

Microsoft Intelligent Cloud Growth Masked by OpenAI

Zitron revealed data from his reporting on OpenAI audited financials to highlight how deeply hyperscaler optics rely on a single client. For calendar year 2025, 69% of the year over year growth in Microsoft Intelligent Cloud segment came directly from OpenAI. Strip away OpenAI spending, and Microsoft cloud segment grew at just 8% year over year, barely keeping pace with inflation.

"Everyone is being sold what I consider kind of a lie. It is honestly kind of a scandal." - Ed Zitron

This tight dependency creates a structural loop. Hyperscalers provide capital or compute subsidies to startups, which the startups immediately turn around and record as cloud spend with those same hyperscalers. This circular financing artificially inflates cloud growth rates while giving public market investors a false sense of organic market expansion.

The Mechanics of Circular Financing

Zitron outlined how Big Tech firms structured these arrangements to manufacture their own biggest buyers. Hyperscalers built $70 billion to $100 billion worth of compute infrastructure for OpenAI and Anthropic, granting them compute access that no other market entrants could ever afford.

Zitron cited an example involving hardware supply chains: Broadcom sells TPUs to Google, which sells them to Anthropic, only for Google to rent compute capacity back through Anthropic. While this double counting looks lucrative on paper, it relies entirely on startups that burn cash at unprecedented rates.

Zitron drew explicit comparisons to corporate history, noting that this setup mimics high-profile financial collapses where executive teams created artificial demand loops to satisfy quarterly equity expectations.

OpenAI Cash Burn and SoftBank Revenue Questions

Examining audited financial disclosures that Zitron originally reported for the Financial Times, he highlighted the deep structural losses inside leading AI model builders. OpenAI lost $20.9 billion in 2025, with cash drain accelerating rather than contracting.

Furthermore, Zitron highlighted that over $800 million of OpenAI recognized revenue originated from SoftBank under a program termed Crystal Intelligence. Zitron noted that despite extensive digging, he found no evidence of operational deliverables associated with this initiative, despite SoftBank holding a major equity stake without board representation.

StartupHub.ai data shows OpenAI maintains a rating score of 84/100, supported by VERIFIED financials including $100B raised in 2026 at a post-money valuation of $850B. By comparison, Google Cloud holds a StartupHub score of 17/100. Among close competitors tracked by StartupHub.ai, Anthropic scores 76/100, Arena scores 71/100, Prometheus scores 64/100, Subquadratic scores 60/100, and Locai Labs scores 50/100.

The $1.6 Trillion Data Center ROI Impossibility

The core problem for hyperscaler CapEx is the sheer mathematical imbalance between data center construction costs and real-world software revenues. Zitron cited data from Sightline Climate indicating that 190 gigawatts of data center capacity are currently planned or under construction globally.

Factoring in a power usage effectiveness (PUE) rating of 1.3 at standard energy costs, data centers at that scale require more than $1.6 trillion in annual recurring revenue just to service capital expenditures and operation costs. Zitron emphasized that two venture-backed customers, even spending $400 billion annually combined, cannot satisfy that revenue requirement.

"OpenAI and Anthropic have to grow so very large to make AI pay off, because otherwise there just is not demand for compute at scale." - Ed Zitron

With OpenAI pushing its expected IPO timeline out to 2027 and data centers taking 12 to 36 months to build, Zitron warns that any disruption in venture capital flows to these primary model vendors will trigger a domino effect across public cloud stocks.