The Trillion Dollar AI Race

Business Onion · 2:57

In June 2026, two companies filed to go public within a week of each other. Both build artificial intelligence. Both are worth close to a trillion dollars. And one of them only exists because its founders walked out of the other. In 2021, a group of senior researchers left OpenAI — led by siblings Dario and Daniela Amodei — over a disagreement about how to build something this powerful: whether safety had to come first. They founded Anthropic and built it as a public-benefit corporation. For years it was the smaller, quieter one. Then Claude became the tool businesses reached for, especially to write software — and its revenue went from about a billion dollars a year to a run-rate near forty-seven billion in about eighteen months. In May 2026 Anthropic raised sixty-five billion dollars in a single round — the largest in history — at a $965B valuation, passing OpenAI. Then both reached for the public markets: Anthropic filed first on June 1st; OpenAI followed a week later. (Neither has sold a share yet — these are confidential drafts.) Why go public at all? Frontier AI burns tens of billions a year with profit years away, and public markets are the only pools deep enough to fund it. Both built unusual structures — OpenAI's nonprofit kept control; Anthropic put a trust of mission guardians above its investors — to take the money without surrendering the mission.