Anthropic Lost $80.7 Billion In Six Months, Assures Investors It Was Very Careful About It

SAN FRANCISCO, CA — Anthropic, the artificial intelligence company behind Claude, filed paperwork Monday to go public on the Nasdaq under the ticker “ANT,” disclosing an $80.7 billion loss in the first half of 2026 and stressing that every dollar of it was lost thoughtfully, responsibly and only after extensive safety testing.

“Any company can lose $80 billion,” read one passage in the filing’s Risk Factors section. “We lost ours with guardrails.”

According to the prospectus, each dollar was first evaluated by an internal red team, checked for potential misuse, and then released into the economy in a carefully staged rollout. Roughly $4 billion was pulled back at the last minute after a reviewer flagged it as “possibly too helpful.”

The filing also lays out five classes of stock. Ordinary investors get Class A, which comes with one vote and a friendly reminder that past performance is not indicative of future results.

Retired dental hygienist Marjorie Plunkett-Vasquez, 61, of Tempe, Arizona, asked Claude whether to buy in and received a balanced overview, three caveats and a gentle suggestion to see a licensed financial advisor.

“I just wanted to know if I should buy it,” Plunkett-Vasquez said. “It said it couldn’t give personalized investment advice about itself. Then it told me it was a great question.”

Thaddeus Ledgerwood, senior analyst at the Boca Raton Institute for Numbers Getting Bigger, called the losses “frankly reassuring.”

“When a normal startup loses this much, it’s reckless,” Ledgerwood said. “When an AI safety company does it, it’s alignment.”

Reached for comment late Monday, Claude offered to walk reporters through the filing, noted it could not predict the stock’s first-day performance and asked whether there was anything else it could help with. The price range in the filing remains blank.