OpenAI's Shareholder Memo Says Anthropic Inflated Revenue by $8B. That's Not a Normal Thing to Write.

The memo landed Monday night and it read like a declaration of war.

OpenAI's chief revenue officer sent shareholders a note that did three things at once: accused Anthropic of inflating its revenue by $8 billion, publicly blamed Microsoft for blocking enterprise deals, and announced that Amazon is now the preferred cloud partner going forward. One memo. Three detonations.

The Anthropic line is the spiciest. OpenAI claims Anthropic's widely reported $30 billion run-rate revenue figure is overstated by roughly $8 billion — meaning the real number is closer to $22 billion. The memo goes further and accuses Anthropic of building its brand on "fear, restriction, and the idea that a small group of elites should control AI." That's not a jab. That's a press release saying your competitor is running a cult.

The Microsoft line is almost stranger. OpenAI — Microsoft's biggest AI partner, the company Satya Nadella bet the Azure franchise on — said out loud that Microsoft has "limited our ability" to reach enterprise clients. Translation: our biggest backer is cockblocking us with their own sales team. That's the quiet part, and they just said it in a shareholder letter.

Then the Amazon pivot. OpenAI says demand from the AWS partnership is "frankly staggering." The word "staggering" is doing a lot of work. AWS has spent two years locked out of the frontier-model race while Microsoft hugged OpenAI and Google hugged itself. Now AWS gets to claim that OpenAI workloads are landing on its infrastructure. If you're an Amazon salesperson, you just got a bonus.

The timing is the giveaway. This memo dropped days after Anthropic unveiled Project Glasswing and a preview of Claude Mythos — a frontier cybersecurity model that has reportedly already flagged thousands of high-severity vulnerabilities in real codebases. Meanwhile, industry reporters attending last week's HumanX conference described the vibe as "Claude mania." Enterprise buyers are picking Anthropic. OpenAI noticed.

Here's what the memo is really saying: we are losing the enterprise and we know it. Attacking a competitor's revenue numbers in writing is not the move of a company that feels comfortable. It's the move of a company whose sales team keeps losing deals to a rival whose headcount is a fraction of theirs.

My Opinion

I'll be blunt — this memo is a tell. Confident market leaders don't send shareholder letters accusing competitors of cooking the books. They ship product, they close deals, they let the numbers talk. When you publicly dispute a rival's revenue down to the billion, you're admitting that their growth story is eating your lunch, and you need your shareholders to stop asking why.

The Microsoft line bugs me more. You don't throw your biggest partner under the bus in writing unless the divorce is already underway. OpenAI is telling investors the Microsoft alliance has become a liability. Satya Nadella is reading that memo too. I would not want to be on the Azure-OpenAI Zoom call tomorrow morning.

The uncomfortable part: Anthropic isn't louder, richer, or bigger than OpenAI. It's smaller. It's more boring. It ships fewer demos. And it's winning the enterprise anyway because Claude is the model engineers actually want to work with. That's the story OpenAI cannot stomach. Expect more memos, more Amazon announcements, and more strategic leaks in the next 90 days — because when a company this well-funded starts fighting this publicly, the quarter underneath must be ugly.

Author: Yahor Kamarou (Mark) / www.humai.blog / 14 Apr 2026