EZ Primary Research chief executive Ed Zitron used a Bloomberg interview published on July 31, 2026 to argue that the capital expenditure wave lifting technology stocks rests on two loss-making customers, citing UBS estimates that OpenAI and Anthropic will account for 27% of Google Cloud revenue this year and more than 48% next year.

The conversation, released on the Bloomberg Podcasts YouTube channel after a week in which Microsoft, Amazon, Meta, and Alphabet reported quarterly results, drew more than 420,000 views by August 2. "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," Zitron said, according to the interview.

The claim lands on numbers that investors rarely see broken out. According to UBS estimates cited by Zitron, OpenAIand Anthropic together will generate more than $124 billion of Google Cloud revenue next year, with Anthropic alone contributing $76 billion in 2027. Zitron noted that OpenAI's position as a large Google Cloud customer remains little known, a detail he attributed to UBS analyst Stephen Ju.

Two customers behind the cloud growth

Microsoft shows a similar pattern, according to the interview. Zitron cited Barclays estimates placing the two AI companies at 13% of revenue this year and 18% next year for a cloud operation he described as a much bigger business than Google Cloud. His own reporting on OpenAI's finances produced a sharper figure. According to Zitron, 69% of the year-over-year growth in the Microsoft Intelligent Cloud segment during 2025 came from OpenAI. Without that single customer, the segment would have grown 8%, which he characterized as barely beating inflation.

"Everyone is being sold what I consider kind of a lie. It's honestly kind of a scandal," Zitron said during the interview.

The concentration extends beyond the hyperscalers. According to reporting by The Information referenced in the conversation, 89% of revenue at the largest AI companies comes from OpenAI and Anthropic alone. When a host asked whether the industry must be concentrated to some degree, given the cost of building capable models, Zitron clarified that his concern targets the concentration of revenue in two companies rather than the concentration of compute itself.

The scale of the projected payments raises its own question. "How is Anthropic going to afford that? They burned tens of billions of dollars," Zitron said, referring to the $76 billion UBS estimate for 2027.

OpenAI's balance sheet and a delayed listing

Zitron brought direct knowledge of OpenAI's accounts to the discussion, having reported the company's audited financials for the Financial Times. "It's a company just burning cash. They lost $20.9 billion in 2025," he said. More than $800 million of OpenAI's revenue that year came from SoftBank for a program called Crystal Intelligence, according to Zitron, who said he could find no evidence of activity connected to it. SoftBank holds a large shareholding in OpenAI without board seats, he added.

The listing timeline compounds the pressure. OpenAI had been expected to go public this year, but the New York Times reported the company is considering a delay until 2027, according to the interview. "That's lethal for a number of people," Zitron said.

The structural problem, in his telling, sits in the cash cycle. "OpenAI and Anthropic need continual flows of capital. They do not pay their bills out of existing cash flow," Zitron said. Any interruption to that capital, he argued, becomes the first domino.

Circular structures and owned infrastructure

Asked whether the companies running the hyperscalers failed at due diligence, Zitron offered a different reading. The platforms "did the due diligence in the sense that they said, we are going to create our largest customers and we're going to own large parts of them," he said. The arrangement extends to hardware. According to Zitron, Broadcom sells tensor processing units to Google, the chips are then sold to Anthropic and rented back to Anthropic through Google. "Google gets to double up on revenue," he said.

The funding picture follows the same loop. OpenAI and Anthropic have raised between $200 billion and $300 billion, according to Zitron, yet the effective total runs higher because Microsoft, Google, and Amazon built their infrastructure for them. He referenced testimony from the trial between Sam Altman and Elon Musk, in which a Microsoft executive put that infrastructure cost at $100 billion, before settling on a figure of roughly $70 billion to $80 billion of capacity the AI companies never had to pay for. The exchange turned briefly to Enron and WorldCom as reference points for fiduciary responsibility, with the explicit caveat that no comparable conduct was being alleged.

PPC Land coverage documents how deep the ownership ties run. Microsoft and Nvidia committed a combined $15 billion to Anthropic in November 2025, valuing the company at approximately $350 billion, while Anthropic committed to purchasing $30 billion of Azure compute capacity - a $3.2 billion gain on that stake lifted Microsoft's net income 31% in the quarter reported on July 29, 2026. The customer-concentration risk Zitron describes has already reached credit ratings elsewhere: S&P Global Ratings cut Oracle to BBB- in July 2026, noting that OpenAI accounts for roughly half of the $638 billion in remaining performance obligations Oracle carries.

The data center mathematics

Zitron then worked through the demand side of the buildout. Sightline Climate identified about 190 gigawatts of data center capacity in planning or construction as of February, according to figures he cited. Applying a power usage effectiveness rating of 1.3 - the ratio between total facility power and the power reaching computing equipment - and $12 million per megawatt, Zitron calculated that the facilities would require more than $1.6 trillion in annual revenue to justify themselves. "Having two customers is not going to do that," he said. Even the heaviest spender could not close the gap; he estimated OpenAI would need to spend $400 billion a year, funding he does not believe the company will secure.

Construction timelines add friction. Data centers take 12 to 36 months to build depending on size, according to the interview, slower than the market narrative assumes.

The transformation of the platforms themselves troubles Zitron as much as the arithmetic. Amazon, Google, and Microsoft have shifted from cash-generating businesses with low asset intensity into what he called "bulbous, GPU filled asset mongers," filled with semi-built data centers serving two or three customers. Meta invests in AI as well but does not yet sell compute capacity, he noted. For his thesis to hold, Zitron argued, nothing dramatic needs to happen: OpenAI and Anthropic would simply have to grow to an implausible size to make the capacity pay off, because otherwise demand for compute at scale does not exist.

Why this matters for the marketing community

The companies named throughout the interview are the companies that sell most of the world's advertising, and the spending Zitron questions runs through the earnings reports marketers watch each quarter. Alphabet raised its 2026 capital expenditure guidance to $195 billion to $205 billion on July 22, 2026, reported negative free cash flow of $5.9 billion for the quarter, and now carries $98.2 billion in long-term debt, up from roughly $16 billion a year earlier. The company had already raised approximately $85 billion in equity in June 2026 to fund infrastructure it describes as supply-constrained against demand, a framing that sits at the opposite pole from Zitron's reading. Meta reported quarterly capital expenditures of $31.08 billion on July 30, 2026, against $17.01 billion a year earlier, with free cash flow falling to $784 million. Microsoft's capital expenditures rose 70% to $41.0 billion in its June quarter, with calendar 2026 spending expected around $175 billion. Advertising revenue funds a substantial share of those budgets, which means ad businesses now bankroll infrastructure whose demand case rests, in Zitron's account, on two unprofitable tenants.

The exposure runs in the other direction too. The same AWS infrastructure serving Claude and GPT-5 powers programmatic bidding, creative generation, and Amazon's Rufus shopping assistant, so the operational continuity of AI ad tooling depends on the economics Zitron disputes. OpenAI itself has entered the advertising market that would need to fund it, with projections of $102 billion in advertising revenue aimed at Google's $224 billion search business. Concentration compounds the stakes: OpenAI, Google, and Anthropic held more than 84% of the AI agent market as of May 2026, leaving marketers who build on these systems dependent on a narrow set of suppliers whose financing Zitron considers unsustainable.

The productivity question raised near the end of the interview cuts closest to marketing employment. One host framed the best case for the spending as productivity gains in which fewer people do more work, noting serious implications for the labor force if the bet succeeds and for the stocks if it fails. The marketing industry has already placed that bet with its own money. Agency leaders named AI their top investment priority for the second consecutive year, with 77.7% of vice presidents and above planning to increase AI spending, while 39.9% of agencies conducted layoffs within the preceding 12 months. AI sales and marketing investment reached $3.7 billion globally in the first part of 2026. Whether tokens sold by two cash-burning model companies can substitute for salaried expertise remains contested territory: one analysis argues cheaper marketing work expands total demand for marketers rather than shrinking it, while task-level measurement shows models covering a fifth of job tasks without revealing whether those are the tasks that matter. If Zitron is right about the financing, the industry restructuring its workforce around these systems has tied its labor model to companies that, in his words, do not pay their bills out of existing cash flow.

Timeline

Summary

Who: Ed Zitron, chief executive of EZ Primary Research, speaking on Bloomberg, with claims involving OpenAI, Anthropic, Google, Microsoft, Amazon, Meta, SoftBank, Broadcom, and estimates from UBS, Barclays, Sightline Climate, and The Information.

What: Zitron argued that AI capital expenditure builds infrastructure for two unprofitable companies, citing UBS estimates that OpenAI and Anthropic will supply 27% of Google Cloud revenue this year and more than 48% next year, worth over $124 billion, alongside OpenAI's $20.9 billion loss in 2025 and a calculation that planned data centers require more than $1.6 trillion in annual revenue.

When: The interview was published on July 31, 2026, following the week in which Microsoft, Meta, Amazon, and Alphabet reported quarterly results, and had drawn more than 420,000 views by August 2, 2026.

Where: The interview took place in the Bloomberg Interactive Brokers studio and was distributed globally through the Bloomberg Podcasts YouTube channel.

Why: The revenue concentration matters to the marketing community because advertising income funds the hyperscaler infrastructure in question, AI advertising tools run on that same infrastructure, and agencies restructuring their workforces around AI have tied their labor models to two companies that, according to Zitron, do not pay their bills out of existing cash flow.