Something unusual is happening in the secondary market for AI company shares. OpenAI employees and early investors are quietly selling their stakes at discounts of 15 to 20 percent below the last official valuation. Meanwhile, Anthropic cannot close its funding rounds fast enough because demand keeps exceeding supply.

According to the LA Times and multiple secondary market trackers, the shift began accelerating after OpenAI announced its for-profit conversion plans. The move spooked a specific type of investor: the ones who bought in believing OpenAI's nonprofit mission was a feature, not a bug. Those investors are now heading for the exits.

Anthropic is the direct beneficiary. The company's revenue doubled in two months to $19 billion annualized, Claude Code is the hottest developer tool on the market, and institutional investors are lining up with checkbooks open. Coatue Management has reportedly told limited partners that Anthropic could be worth $2 trillion within three years.

The Trust Gap

This is not just about financials. It is about trust. OpenAI has spent the last year lurching from crisis to crisis: the board coup, the for-profit pivot, Sora getting killed, the super app gamble, and now a $122 billion raise that values the company at $852 billion despite burning $14 billion a year. Each move individually makes business sense. Together, they paint a picture of a company that does not know what it wants to be.

Anthropic, by contrast, has a boringly consistent story: build safe AI, sell it to enterprises, do not run ads, do not chase consumers. It is the anti-OpenAI pitch, and right now, the money prefers boring.

The secondary market is often the first place you see a sentiment shift before it shows up in headlines. Right now, it is saying something very clear: the smart money thinks Anthropic is the safer bet. Whether that lasts depends on whether OpenAI can turn its $852 billion valuation into something that actually makes money. So far, the math says no.