Anthropic Is Making $65 Billion a Year and Getting Ready to Go Public
Anthropic, the company behind the Claude AI assistant, is now bringing in money at a rate of more than $65 billion per year, according to a source familiar with the company's finances. That figure, reported in mid-August 2026, sets the stage for the company's planned stock market debut, which is being valued against a prediction that Anthropic will earn roughly $190 billion to $200 billion by 2028 (Reuters).
The growth has been fast and getting faster. Think of the revenue run rate as a speedometer reading — it takes how much money the company is making right now and projects it over a full year. In November 2025, Anthropic expected to reach about $26 billion the following year (Reuters). By April 2026, it had already passed $30 billion (Reuters). By late May, it hit $47 billion (AP News). The latest number, $65 billion, came just weeks before the company is expected to start pitching investors ahead of its stock market launch (Reuters).
Anthropic quietly filed paperwork for an initial public offering (IPO) in early June (Reuters). An IPO is when a private company first sells shares to the public. The company has also secured a loan of more than $10 billion ahead of that offering, according to a Bloomberg News report (Reuters).
The private funding rounds before the IPO have been enormous. In September 2025, Anthropic raised $13 billion from investors led by ICONIQ, giving the company a valuation of $183 billion (Anthropic). In February 2026, a $30 billion round led by GIC and Coatue valued it at $380 billion (Anthropic; Reuters). By late May, a $65 billion round led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital pushed the valuation to $965 billion (AP News; Anthropic). The WSJ reported in May that the valuation was nearing $1 trillion (WSJ). Microsoft and Nvidia together committed up to $15 billion in January (Bloomberg).
Much of that growth comes from Claude Code, a tool that helps programmers write software using AI. In February, Anthropic said Claude Code was generating $2.5 billion per year on its own (WSJ). By July, the WSJ reported that Anthropic had pulled ahead of its main rival, OpenAI, in both revenue and valuation, largely because of Claude Code's success (WSJ). In May, the WSJ reported Anthropic's revenue was on track to reach $50 billion by the end of that month (WSJ).
The gap with OpenAI is wide. OpenAI expects to spend about 14 times as much cash as Anthropic before it becomes profitable in 2030, according to a November 2025 WSJ report (WSJ).
Anthropic has also taken two notable steps beyond the financial numbers. In April, it launched Project Glasswing, an effort to protect the world's most important software systems as AI becomes more powerful (Anthropic). In June, while valued at roughly $1 trillion, the company warned that AI systems are getting close to being able to improve themselves without human help (WSJ).
The IPO valuation math deserves a closer look. A forecast of $190 billion to $200 billion in revenue by 2028, for a company making $65 billion today, means roughly tripling in two years. Anthropic has done that before — it went from under $30 billion in early 2026 to $65 billion by August. But keeping that pace going as the numbers get bigger, and as competition from OpenAI and free, open-source AI tools intensifies, is harder than doing it once. The $10 billion-plus loan the company took on before its IPO signals that Anthropic is still spending heavily even at these revenue levels. Whether it can actually turn a profit, not just grow revenue, is the question that remains unanswered.
The self-improvement warning adds another dimension. A company predicting $200 billion in revenue from AI products while also warning that those products may soon improve on their own is either being unusually honest or preparing the ground for a safety and regulation story that will help shape the IPO. In my view, both are probably true. The approach lets Anthropic present itself as the careful, responsible guardian of a powerful technology at the very moment it is asking public investors to value it at nearly a trillion dollars based on the assumption that the technology will keep growing just as fast.
On the positive side, what this makes possible is clear. A publicly traded Anthropic with $65 billion in annual revenue and a $10 billion-plus loan facility would have the money to develop more advanced AI and to pursue security projects like Glasswing at a level no private company could match. The IPO, if it reaches the expected valuation, would also give the public a real way to put a price on AI capability itself for the first time — not just on the data centers and servers that support it. The industry has not had that through three years of private deals, and it comes at a moment when, by Anthropic's own words, the technology is approaching a point where simple predictions may no longer hold.
