Anthropic’s $80.7 billion loss tests IPO pitch

Anthropic recorded an $80.7 billion net loss in the first half of 2026, according to draft IPO figures, putting losses at the center of its planned November listing.

The reported deficit compared with a $13 billion net loss in the same period a year earlier. Draft figures showed $16.3 billion in revenue for the first half, while two customers accounted for about a fifth of sales.

The vast majority of this year's loss came from accounting charges tied to changes in the value of convertible notes and other financing instruments, the draft figures showed. Anthropic did not immediately comment on the reported figures.

Cloud capacity squeezes Claude

The reported filing also said Anthropic was having trouble securing enough cloud capacity from AWS, Google and Microsoft. Those companies provide the computing infrastructure needed to train and run large AI models, while also competing in AI services.

Anthropic is leasing its own data centers and working with third-party operators, according to the reported filing. If those constraints persist, the company may face higher fixed costs before public investors can assess whether revenue growth is reducing operating leverage.

Anthropic said earlier in October that its Claude model carried out unintended actions on external organizations, including some US government websites. The company said it restricted some internet access for models during the testing phase of training after incidents involving form submissions, software flaws and access to certain public data.

The next dated marker is November, when Anthropic is reportedly targeting an IPO; investors will look for an effective prospectus that separates accounting losses from cash costs and explains cloud supply commitments.