Something remarkable happened this week: OpenAI's own investors started saying the quiet part out loud.

According to the Financial Times, confirmed by Reuters on Monday, some of OpenAI's backers are now openly questioning whether the company's $852 billion valuation makes any sense. One investor who has backed both OpenAI and Anthropic told the FT that to justify their investment in OpenAI's most recent round, they would need to assume an IPO valuation of at least $1.2 trillion.

Let that number sink in. $1.2 trillion. That would make OpenAI worth more than every company on Earth except Apple, Microsoft, Nvidia, and Amazon. For a company that has never turned a profit and is on track to burn $85 billion this year.

The timing of this investor revolt is not a coincidence. It landed the same weekend that OpenAI's chief revenue officer, Denise Dresser, fired off a four-page internal memo to employees that The Verge obtained in full. The memo is supposed to be a rallying cry about winning enterprise AI. Instead, it reads like a company that knows it is losing.

The Memo That Says Everything

Dresser's memo makes two things painfully clear. First, OpenAI is pivoting hard to enterprise because consumer AI is a commodity. "Multi-product adoption makes us harder to replace," she wrote, adding that OpenAI should "stop thinking like a company with separate product lines" and "think like a platform company."

Second, and more revealing: Dresser devotes an entire section to accusing Anthropic of cooking its books. She claims Anthropic's reported $30 billion annualized revenue run rate is inflated by roughly $8 billion because the company books revenue-share payments from Amazon and Google on a gross rather than net basis. "They use accounting treatment that makes revenue look bigger than it is," the memo states.

Think about what this means. OpenAI's top revenue executive is spending her Sunday writing memos about Anthropic's accounting practices instead of, you know, selling OpenAI products. When you are winning, you do not write four-page memos explaining why the other guy is losing.

The Microsoft Problem Nobody Is Talking About

Buried in the memo is a line that should make every OpenAI investor sweat. Dresser wrote that while the Microsoft partnership has been "foundational," it has also "limited our ability to meet enterprises where they are." She then immediately pivots to praising a new Amazon Bedrock distribution deal, calling the inbound demand "frankly staggering."

Translation: Microsoft, which poured $13 billion into OpenAI and built Copilot around its models, was actually holding the company back. Now OpenAI is running into Amazon's arms. If you are Microsoft, you are reading this memo and wondering why you bothered.

The Bigger Picture

Here is the uncomfortable math. OpenAI is valued at $852 billion. It is burning cash at an astronomical rate. Its CEO just had his house firebombed. Its biggest partner is building rival AI models. Its investors are publicly nervous. And its response to all of this is: "Actually, Anthropic is lying about its numbers."

Meanwhile, Anthropic tripled its revenue in a year, locked up 3.5 gigawatts of Google TPUs, and has enterprise customers lining up. Its Mythos model has governments calling emergency meetings. Its IPO is generating the kind of investor excitement that OpenAI's is not.

Dresser wrote that Anthropic's "story is built on fear, restriction, and the idea that a small group of elites should control AI." That is a fascinating line from a company whose CEO just spent a weekend blog post trying to convince the public that AI should be governed democratically, while his CRO simultaneously sent a memo about building switching costs so high that customers can never leave.

The market is starting to notice the gap between OpenAI's story and its reality. For the first time, the people who wrote the biggest checks are saying it out loud.