OpenAI raised $122 billion. It is valued at $852 billion. It is preparing to go public. And its own investors are privately telling the Financial Times that the company is, their word, "deeply unfocused."

"You have ChatGPT, a 1 billion-user business growing 50-100 per cent a year, what are you doing talking about enterprise and code?" one early OpenAI investor told the FT. That is not the kind of quote that appears in a puff piece. That is a shareholder who is worried.

The concern is simple math. To justify OpenAI's current valuation, investors need to underwrite an IPO valuation of at least $1.2 trillion. One investor who backed both OpenAI and Anthropic told the FT that is getting harder to defend, especially when Anthropic, most recently valued at $380 billion, is growing faster and costs less than half the price to buy into.

The Focus Problem

In the past few months, OpenAI has killed Sora, launched a $100/month Pro plan for coders, started building a super app that combines ChatGPT, Codex, and its browser, bought a media company, bought a finance startup, announced plans for conversational ads worth $100 billion by 2030, and dumped Microsoft for Amazon as its preferred cloud partner.

Any one of those moves makes sense individually. Together, they paint a picture of a company lurching between identities. Is OpenAI a consumer product company? An enterprise platform? A coding tool? An ad business? A media conglomerate? The answer appears to be: yes, all of them, simultaneously, while preparing for the most scrutinized IPO in technology history.

Meanwhile, Anthropic has one message: we build the most capable and safest AI. It tripled its revenue to $30 billion. It passed OpenAI in enterprise adoption, with 30.6% of US businesses now paying for Anthropic. It put a pharma CEO on its board. Every move reinforces the same story.

No Man's Land

The dual-backer investor told the FT that OpenAI was in danger of being left "in no man's land." That phrase should terrify OpenAI's leadership. It means the company is not clearly the best at anything. Not the best consumer product (ChatGPT is losing users for the first time). Not the best enterprise play (Anthropic is winning that race). Not the best coding tool (Claude Code is eating Codex's lunch). Not the cheapest (Google is giving models away free).

OpenAI's CFO Sarah Friar pushed back hard: "The suggestion that investors are not supportive of our strategy defies the facts. Our raise, the largest in history, was oversubscribed, completed in record time and backed by a broad set of global investors." She is not wrong. The money is real. But the questions are also real.

The IPO Clock

OpenAI is targeting a Q4 2026 IPO. Anthropic is also preparing to go public. SpaceX (now merged with xAI) is filing for what would be the largest IPO in history at $1.75 trillion. Wall Street does not have enough capital to absorb all three at once.

That creates a beauty contest. Institutional investors will compare OpenAI at $852 billion (needing $1.2 trillion at IPO) against Anthropic at $380 billion (growing faster, cleaner narrative, stronger enterprise position). The math is not complicated. If you can buy the company winning the enterprise race for less than half the price of the company having an identity crisis, the choice makes itself.

PYMNTS CEO Karen Webster made an important point: this race will not be determined by product releases alone, but by whether platforms understand they are competing for user habits, not features. The consumer who opens ChatGPT before her coffee is not switching because of a benchmark score. She is switching because another model earns a specific role in her routine.

OpenAI still has 1 billion users. That is an extraordinary moat. But moats erode fast when investors start using words like "deeply unfocused" and "no man's land" on the record.

First reported by the Financial Times. Additional reporting from PYMNTS.