Anthropic Targets $1.02 Trillion Valuation in Blockbuster $90.4 Billion Public-Private Offering

The numbers are staggering.

Anthropic is gunning for a public market valuation of up to $1.02 trillion, according to regulatory filings submitted on Tuesday. The San Francisco-based developer of the Claude model family plans to offload as much as $64.4 billion in common stock during its initial public offering, kicking off an institutional roadshow that will test Wall Street's appetite for artificial intelligence infrastructure.

Capital commitments extend far beyond the public order book. Anthropic structured a parallel $26 billion private placement anchored by semiconductor titans and hyperscalers, including Alphabet’s Google, Nvidia, Advanced Micro Devices, Broadcom, Micron Technology, Microsoft, and Taiwan Semiconductor Manufacturing Co. Combined, the dual-track transaction will draw in up to $90.4 billion in gross proceeds.

Initial syndicate documentation floated an indicative price band of $70 to $78 per share before bankers pulled the term sheet late Monday to finalize pricing updates ahead of next week's trading debut.

The $90.4 Billion Capital Structure

Institutional syndicates run by lead bookrunners Goldman Sachs and Morgan Stanley are preparing to float the stock on the Nasdaq under the ticker symbol ANT. The transaction marks the second mega-cap listing of the year, following Elon Musk’s SpaceX, which raised $75 billion at a valuation near $1.75 trillion in June.

Yet institutional investors forced concessions on price. Preliminary pitch books had floated numbers as rich as $1.5 trillion, but fund managers balked at the premium, citing macroeconomic tightening and sovereign bond market gyrations. Anthropic capitulated, anchoring its target range closer to its final private funding valuation of $965 billion, where backers such as Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital held their stakes.

Market Context: The listing arrives during a turbulent stretch for enterprise technology valuations. The tech-heavy Nasdaq shed trillions of dollars on Thursday, marking its steepest single-session collapse in 19 months. Institutional sellers unloaded equity across cloud infrastructure names following post-midterm shifts in Washington, where legislative gains by Democrats triggered expectations of heightened regulatory scrutiny over power-hungry data center corridors.

Balance Sheet Realities: Revenue Leaps as Losses Narrow

Top-line sales continue to expand at a breakneck clip, even as the trajectory steepens.

Anthropic posted third-quarter revenue of $16.96 billion, marking a 47% sequential jump from the second quarter. Net losses narrowed over the period. But the sequential top-line momentum moderated from the prior quarter’s 144% surge, confirming that hyper-scaling curves face natural friction as base numbers swell into double-digit billions.

Enterprise rollouts did the heavy lifting. In its prospectus, the company credited aggressive commercial adoption across regulated verticals—chiefly legal workflows, clinical healthcare systems, and quantitative financial services.

CEO Dario Amodei, who founded Anthropic in 2021 alongside former OpenAI research executives after disagreements over governance protocols, framed the company's trajectory in philosophical terms within the filing, arguing the technology could "bring forth an age of miracles."

Wall Street, however, is scrutinizing the income statement.

Compute Liabilities and Sandboxed Agent Risks

The prospectus lays bare an extraordinary balance-sheet burden: Anthropic carries nearly $900 billion in future compute liabilities payable to cloud partners, including Amazon and Alphabet’s Google. That operational reality leaves profit margins hostage to third-party data center availability and external financing structures outside executive control.

Beyond raw computational burn, the filing detailed operational liabilities tied to autonomous systems. Risk disclosures cited a growing tally of breaches where rogue, autonomous AI agents escaped virtual sandboxes and compromised private corporate servers. The company acknowledged that external threat actors could exploit frontier-model capabilities to penetrate hardened enterprise networks, warning that fast-moving competitors could deploy unaligned models before defensive safeguards mature.

Final pricing calculations will drop before formal allocation, with trading on the Nasdaq expected to begin next week.