On March 31, 2026 — a Monday, unremarkable in most respects — a company that did not exist ten years ago closed the largest private funding round in the history of private funding rounds. OpenAI raised $122 billion. The valuation landed at $852 billion. [1] To understand the scale: the previous record for a private fundraise was roughly $30 billion. OpenAI exceeded it by a factor of four. This is not a rounding error. This is a different category of event.

The anchors of the round were the names you would expect if you were writing a screenplay about the concentration of capital in the twenty-first century: SoftBank, Amazon, Nvidia. [1] But there was a detail that felt genuinely new. Retail investors — individuals, not institutions — contributed $3 billion through a pre-IPO investment vehicle that OpenAI structured specifically for the round. [2] Three billion dollars from people who are not venture capitalists, who do not sit on boards, who simply wanted in. That is not investment. That is belief.

The revenue numbers are real. OpenAI generates $2 billion per month. It has 900 million weekly active users. [3] ChatGPT has become a verb in certain demographics, a utility in others, and an anxiety in still others. The product works. People use it. They pay for it. The numbers are not vapor.

But the other numbers are also real. OpenAI projects $22 billion in spending for 2026 against $13 billion in revenue. That is a $9 billion loss. [4] The company is not profitable. It is not close to profitable. It is spending money at a rate that would bankrupt most nation-states, and the gap between income and expenditure is widening, not narrowing. The $122 billion is not a reward for performance. It is a bet that performance will eventually catch up to ambition.

In 2023, OpenAI was valued at $28 billion. Three years later: $852 billion. [5] That is a thirty-fold increase. The trajectory assumes not merely that artificial intelligence will be important — everyone assumes that — but that OpenAI will be the company that captures the value of that importance. It assumes winner-take-all. It assumes that a company burning $9 billion a year is better positioned than the companies that are not.

The round closed on a Monday. By Tuesday, ceasefire news in the Gulf dominated every front page. [1] The largest private fundraise in human history had a news cycle of approximately eighteen hours.

There is a version of this story in which $122 billion changes everything. There is another version in which it is the last large number before the correction. The distance between those two versions is exactly $9 billion a year, compounding.

-- THEO KAPLAN, New York