Anthropic IPO Target Signals AI’s Public Market Stress Test
Anthropic is weighing an IPO that could match SpaceX’s $75 billion debut while revising enterprise data controls for its AI models.
Jurgen Goldmeier ·

Anthropic IPO plans could seek at least $75 billion, people familiar said, matching SpaceX’s record sale and testing demand for frontier AI.
The Claude developer is running calculations for a possible public filing as soon as late August, according to people familiar with the matter, who were not authorized to discuss private plans. Recent investor briefings led by Chief Financial Officer Krishna Rao did not settle a valuation, the people said.
The discussions remain fluid, and the eventual size of the offering could change. Anthropic declined to comment on the potential listing, the people said.
SpaceX sets the target
SpaceX raised $75 billion in its initial sale, making it the largest first-time share offering on record, according to compiled IPO data. The total rose to $86.2 billion after the overallotment option, a mechanism often used when early trading supports extra share sales.
Anthropic’s target follows a private funding surge across artificial intelligence companies. The company raised $65 billion in May at a $965 billion valuation, while OpenAI raised $122 billion in March at an $852 billion valuation, according to figures cited by people familiar with those transactions.
The financial picture is expanding quickly but remains costly. Anthropic had preliminary second-quarter revenue above $11.5 billion, compared with $787 million in the same period of 2025, and its annualized revenue run rate reached $65 billion by the end of July, people familiar said.
Company documents reviewed by reporters showed positive adjusted operating income in the second quarter. The same documents showed a net loss of almost $42 billion in 2025, roughly five times the $8.3 billion loss reported for the prior year.
Compute costs shape the filing
Anthropic and its rivals are spending heavily to train and run frontier models, people familiar said. In one agreement with data center owners, Anthropic committed to buy computing resources from SpaceX in a deal that could be worth tens of billions of dollars over the next three years.
The company is also arranging financing before any public sale. A revolving credit facility is expected to exceed its roughly $10 billion target, people familiar with the matter said, giving Anthropic another source of liquidity before it tests public investors.
Morgan Stanley, Goldman Sachs Group and JPMorgan Chase are working with Anthropic on the IPO, and other banks could be added, people familiar said. The company is also considering super-voting shares that would give Chief Executive Officer Dario Amodei and co-founders greater control while Amodei holds about 2% of the company, according to a person familiar with the matter.
Customer data becomes listing risk
Anthropic is preparing a separate change for business customers using its most capable AI models. A new safety system expected later this year would still require enterprise customers to retain data for 30 days, but would let them keep it on their own cloud infrastructure, according to a person familiar with the plan.
The shift follows a June policy under which Anthropic said it would retain customer data for its Mythos and Fable models, as well as future frontier offerings, for 30 days to detect cyberattacks using its technology. The company said then that the data would not be used to train its AI systems.
Anthropic acknowledged in a recent company report that the policy could be unpopular with customers accustomed to zero retention and could create business risks if competitors did not adopt similar rules. The revised system has been developed with more than 100 customers in highly regulated industries, the person familiar said.
OpenAI is testing a rival approach called private safety processing with Databricks and Microsoft and plans a wider release in September, the company said. The competition puts data governance beside model performance as a selling point for enterprise AI customers.
Public markets face AI scale
A sale above SpaceX’s record would put 2026 within reach of the strongest year for U.S. IPO volume. Newly listed companies had raised $160.6 billion through August 19, compared with the 2021 record of $195.2 billion, according to compiled IPO data.
If Anthropic files near the target and investor demand holds, U.S. IPO volume would likely pass the prior record, Anthropic would gain capital for computing contracts, and AI peers would receive a new valuation benchmark. If losses, governance terms or data-retention concerns weigh on demand instead, the company may need to adjust size or pricing, while the sector faces closer scrutiny of cash burn and customer controls.