Holy sh*t I think Anthropic is profitable now

Anthropic is reportedly on track for its first profitable quarter, with revenue projected to more than double to ~$10.9B. The analysis breaks down why: exclusive availability of Claude models on AWS (used by 90%+ of Fortune 100), favorable revenue-share deals with cloud providers, aggressive pricing shifts (Opus 4.5 effectively replaced Sonnet at a higher price point), a tokenizer change in Claude 4.7 that generates 30-50% more tokens per request, and enterprise contract restructuring to API-based pricing. The author argues profitability was partly accidental — Anthropic was conservative on compute spend and raised prices to ration GPUs for researchers — but the demand proved inelastic. Claude Code and Opus 4.5 are credited as the key inflection points that drove enterprise adoption and spending. The piece also compares Anthropic's position to OpenAI's, noting OpenAI models are only available on Azure while Anthropic runs on all three major clouds.