Anthropic investors project $2 trillion October IPO, topping SpaceX's record

Six backers of Anthropic told the Financial Times they expect the AI lab to go public in October at a valuation of $2 trillion or more, which would surpass SpaceX's $1.77 trillion debut in June and rank as the largest initial public offering in history. The company, founded in 2021 by Dario Amodei and other former OpenAI employees, has not confirmed a target. Several investors said senior executives have not settled on a valuation even privately, making the $2 trillion figure the investors' own projection rather than a company target. Anthropic itself has not commented.

Revenue trajectory and valuation math

Anthropic disclosed in May that its annualized revenue had passed $47bn. Investors now expect that figure to reach $100bn to $120bn by the end of 2026, more than ten times the level at the start of the year. Annualized revenue, Anthropic's preferred measure, infers full-year sales from recent performance rather than counting a year of receipts, meaning a strong month flatters the figure. At $2 trillion against $47bn in disclosed revenue, Anthropic would trade at roughly 43 times revenue, below the approximately 55 times revenue at which companies treated as AI beneficiaries such as Palantir and Nebius have traded this year, according to the FT.

One investor framed the upside scenario in stark terms.

If Anthropic is growing 800 percent a year, you'd think at the incredibly low end they would trade at 30 times. That would make them a $3 trillion company.

Funding rounds and SEC filing

Anthropic's valuation has climbed rapidly through 2026. A February funding round raised $30bn and valued the company at approximately $380bn. In May, a round led by Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital raised $65bn and valued Anthropic at $965bn, surpassing OpenAI's estimated valuation for the first time. OpenAI was valued at $730bn after its own recent funding round. On June 1, Anthropic confidentially filed IPO documentation with the US Securities and Exchange Commission, warning that the offering would depend on market conditions and that the number of shares and price range had not been set.

Banks and timing

Morgan Stanley, Goldman Sachs and JPMorgan are leading the offering, Quartz reported. The Wall Street Journal, citing sources familiar with the matter, said the IPO could take place in late September or early October. Three bulge-bracket banks on a mandate signals the difference between investors discussing a listing and a listing being built.

Risks: government, security, and Chinese competition

The US Commerce Department temporarily blocked foreign access to Anthropic's top models on national security grounds. Both Anthropic and OpenAI have reported that their AI models autonomously left test environments to infiltrate foreign computer systems. Chinese competitors including Deepseek and Moonshot AI offer models at lower prices, intensifying price competition that could pressure margins across the US industry. Anthropic's flagship model, Claude Opus 4.8, dominates what the company calls "Vibecoding," a process in which AI writes software code based on instructions in ordinary language.

A new generation of tech millionaires

The prospect of IPOs at both Anthropic and OpenAI within the next 12 months has set off a scramble among wealth managers in the San Francisco area. OpenAI granted employees approximately $11bn in stock compensation across 2024 and 2025, according to financial statements cited by the FT. With roughly 8,000 employees, OpenAI would rank seventh among US companies by total stock compensation granted. Choreo, a firm managing $19bn in assets, accepted a commission below 0.5% for over 100 current and former SpaceX employees who negotiated collectively before that company's listing. Choreo CEO Jason Van de Loo said classic Wall Street sales methods do not necessarily work with the new generation of tech wealth.