EZ Primary Research Chief Warns of Concentration Risk in Hyperscaler Cloud Revenue
UBS estimates suggest OpenAI and Anthropic could drive nearly half of Google Cloud’s revenue next year, raising questions about infrastructure sustainability despite strong analyst ratings.
Ed Zitron, chief executive of EZ Primary Research, appeared on a Bloomberg Businessweek segment on 7 August 2026 to argue that the explosive revenue growth powering Google Cloud, AWS, and Azure is dangerously concentrated in just two unprofitable companies: OpenAI and Anthropic. Citing UBS estimates, Zitron noted that 27% of Google Cloud’s revenue this year comes from these two firms, rising to 48% next year, which would total over $124 billion.
Zitron’s thesis suggests that what appears to be broad enterprise demand for AI compute is actually circular financing concentrated in two private entities that do not pay their bills from existing cash flow. He highlighted that 69% of Microsoft Intelligent Cloud’s year-over-year growth in 2025 was driven by OpenAI alone, and without it, the segment would have grown just 8%. At AWS, Barclays puts exposure to those same two customers at 13% this year, rising to 18% next year.
The segment posted $24.768 billion in Q2 2026 revenue, growing 82% year-over-year, its fifth consecutive quarter of acceleration. Alphabet spent $44.924 billion on capex in the quarter, more than double the prior year, and raised roughly $70 billion through combined equity and debt financing. Free cash flow turned negative at -$5.855 billion, and the buyback was suspended.
Zitron then attacked the demand side, citing OpenAI losses of $20.9 billion in 2025 and flagging that over $800 million of OpenAI's revenue came from SoftBank's "Crystal Intelligence" program. Scaling that concern industry-wide, he referenced Sightline Climate projections that data centres will require over $1.6 trillion in annual revenue to sustain. Two customers cannot backfill that hole, Zitron argued, especially when Anthropic and OpenAI cannot afford anything.
The rhetorical peak of Zitron's segment was a comparison to Enron. With OpenAI's IPO reportedly delayed to 2027, Zitron argued executives at the hyperscalers have a fiduciary responsibility to shareholders that may be getting overlooked amid the AI infrastructure race.
Despite these warnings, market analysts remain broadly constructive, with Alphabet carrying 58 Buy ratings and an average target of $428.04. GOOGL is up 80.8% over the past year, AMZN up 23.01%, and MSFT down 3.23% over the same one-year window.
The bull case, that Gemini, Copilot, and Bedrock are seeding genuine enterprise demand well beyond two labs, is real and reflected in Microsoft's commercial RPO of $678 billion, up 84% year-over-year, and over 30 million paid Microsoft 365 Copilot seats. Zitron's warning is a contrarian argument, clearly his opinion, and worth weighing against those data points rather than treating as a verdict.
