Anthropic Selects Banks to Lead IPO Process

Anthropic has selected Morgan Stanley and Goldman Sachs to lead its initial public offering, and JPMorgan Chase is also on the deal, Bloomberg reported on June 3, 2026, citing people familiar with the matter. Anthropic said on June 1 that it filed confidentially with the SEC by submitting a draft registration statement, and Bloomberg reports a listing could come as soon as October. Anthropic and the banks declined to comment, according to Bloomberg.

The filing follows Anthropic's $65 billion Series H, which the company says valued it at $965 billion post-money, up from a $380 billion valuation at its $30 billion Series G in February 2026. At that level Anthropic is the most valuable AI startup and is racing OpenAI to be first among the frontier labs to list publicly.

SpaceX's own S-1 disclosed a related compute arrangement: SpaceX supplies Anthropic with roughly 325,000 Nvidia chips for about $1.25 billion per month, part of a commitment reported at about $45 billion over three years tied to SpaceX's Colossus GPU clusters. The disclosure highlights the scale of compute spending behind frontier models and the concentration of that spending among a small set of suppliers.

For practitioners and investors, an Anthropic listing would create a rare public comparable for a pure-play frontier AI developer, with disclosures on revenue, margins, and compute costs that are currently private. Large, fixed compute commitments also illustrate vendor-concentration and capital-intensity risks common across the sector.

Confidential filings followed by bank mandates typically signal an IPO timetable, but pricing, size, and timing remain subject to market conditions. Key questions include the eventual valuation range, the disclosed economics of multi-year compute contracts, and whether OpenAI moves on a similar schedule.