Anthropic is preparing for an initial public offering that could arrive as early as October, having reached a $965 billion valuation in its latest funding round and trading at $1.2 trillion on secondary markets. The numbers have turned early equity grants into extraordinary personal fortunes: a mid-level technical staff member who joined in 2024 with $1.3 million in equity now holds a stake worth an estimated $72 million. The seven co-founders will each be worth more than $15 billion, and financial advisers report that thousands of employees—from research scientists to a cafeteria chef with a net worth above $10 million—are confronting wealth that is, in the words of one former OpenAI employee, “way beyond what I even know what to do with.”

The trajectory from a pre-revenue startup valued at $4.1 billion in early 2023 to a company on the cusp of a public listing has been fuelled by surging demand for its AI coding tools and large language models. Menlo Ventures, which broke its own rules to invest $1 billion across multiple rounds when Anthropic had no public model and no revenue, now sees that stake valued at roughly $14 billion. Partner Matt Murphy described “heartburn” inside the firm over the initial decision but said he was convinced by chief executive Dario Amodei’s vision of building a formidable challenger to OpenAI, much as Lyft emerged alongside Uber. The firm has since raised $3 billion in new funding, more than double its previous haul.

The wealth concentration is already distorting San Francisco’s property market. Redfin estimates that post-IPO, Anthropic and OpenAI employees could collectively purchase nearly a third of all homes in the metro area, stoking anxiety about pricing out non-tech residents. At the same time, the capabilities of Anthropic’s models are drawing sharp warnings from Washington and Wall Street. JPMorgan Chase chief executive Jamie Dimon likened access to the Mythos model to “giving ballistic missiles to individuals,” noting that the tool has demonstrated an ability to uncover thousands of software vulnerabilities. In June, US authorities ordered Anthropic to restrict foreign nationals’ access to its top Fable 5 and Mythos 5 models over jailbreak risks; the restrictions were lifted after the company implemented new safeguards.

Anthropic is now in talks with banks to open additional credit lines worth billions of dollars, a common pre-IPO step, and investor meetings are expected within weeks. The company has already filed a confidential IPO application with US regulators. Its listing would likely beat rival OpenAI, which is targeting 2027, while Chinese competitor DeepSeek is also eyeing a public offering. The next concrete milestone is the start of the investor roadshow, which will test whether public markets are ready to absorb a valuation that has multiplied more than 200-fold in two years.