Anthropic Taps Morgan Stanley, Goldman for IPO

Anthropic has officially chosen Morgan Stanley and Goldman Sachs to spearhead its initial public offering, with JPMorgan Chase also joining the underwriting team, according to a Bloomberg report on Tuesday. The developer behind the Claude AI model is considering a public debut as soon as October, following a confidential filing on Monday. Additional banks might be brought on board, and the specifics of the offering remain subject to change.

This selection formalizes the underwriting syndicate for what could rank among the largest technology IPOs in history. Anthropic’s listing is part of a wave of major offerings this autumn. SpaceX is slated for a listing as early as June 12 at a $1.8 trillion valuation, while OpenAI is also preparing its own filing with Goldman Sachs, Morgan Stanley, Citigroup, and JPMorgan.

The SpaceX computing deal

The most eye-catching detail came not from Anthropic’s own filings but from SpaceX’s IPO paperwork. It reveals that SpaceX supplies Anthropic with AI computing capacity, including approximately 325,000 Nvidia chips at a cost of $1.25 billion per month. The agreement runs through May 2029 and allows either party to terminate with 90 days’ notice after an initial three-month period.

This arrangement makes SpaceX simultaneously an Anthropic supplier, a competitor (through its Grok chatbot), and a fellow IPO candidate. At $1.25 billion monthly, the deal represents $15 billion in annualized computing costs for Anthropic. That figure puts both the scale of Anthropic’s infrastructure needs and the thin margins it must navigate into sharp context, even for a company projecting $10.9 billion in quarterly revenue. When the S-1 becomes public, it will need to disclose this as a material related-party arrangement given the competitive overlap.

The race to list

Anthropic and OpenAI are vying for investor attention, the same underwriting banks, and ultimately the same pool of public market capital. Both have held discussions with Morgan Stanley, Goldman Sachs, and JPMorgan. The company that lists first sets the valuation benchmark for the AI sector and secures priority access to institutional allocations.

Anthropic’s position has strengthened dramatically this year. Its Mythos cybersecurity model rattled global markets. Its coding agents sent legal software stocks spiraling in February. Its revenue trajectory, from $4 billion annualized in July 2025 to a projected $50 billion run rate by July 2026, has eclipsed OpenAI’s growth. The $965 billion private valuation already surpasses OpenAI’s, reversing a hierarchy that had held since both companies’ founding.

The Pentagon’s supply-chain risk designation of Anthropic remains the most significant overhang for the IPO. The company’s refusal to grant the military unrestricted model access led to a designation typically reserved for foreign adversaries, and Anthropic has said it could jeopardize billions in revenue. How the S-1 discloses and quantifies this risk will be closely scrutinized by public market investors.

Founded by OpenAI defectors

Anthropic was founded in 2021 by former OpenAI staffers including CEO Dario Amodei, positioning itself as a more responsible AI steward than its competitors. Claude and its underlying technology have gained traction with enterprise customers in finance, healthcare, and software development. The enterprise AI spending surge has benefited Anthropic disproportionately. The company is on pace for its first profitable quarter, with $559 million in operating profit on $10.9 billion in Q2 revenue.

The IPO will test whether a company that has raised $65 billion in private capital, achieved a $965 billion valuation, and is spending $1.25 billion per month on a single computing contract can sustain investor enthusiasm in the public market. Morgan Stanley and Goldman Sachs are betting it can.