Joey Brookhart told Nilay Patel the filing was strong, with more than $70 billion in annualised revenue and healthy cash flow. An $80 billion loss in the first half of 2026 is a non-cash accounting charge tied to warrants and notes, he says, and the company has burned only about $10 billion of cash. Two years ago Anthropic lost money on serving its models; now it keeps a large margin. Brookhart estimates only about 30% of its computing power serves paying customers, with the rest going to research and training. Its multi-year computing commitments run to hundreds of billions of dollars.
Why it matters
The filing turns an AI lab’s economics from guesswork into public numbers, and investors will now judge Anthropic on them. Its strong margins pay for the next round of training, which the company wants to grow. The open question is whether demand spreads beyond today’s narrow set of heavy business users fast enough to justify a trillion-dollar price.
Original source
Anthropic is asking Wall Street to believe in AGI
The Verge
theverge.com
Summary by Second Order AI. Full reporting is at the original source.