Overview
- In August Anthropic published a Risk Report that disclosed an unreleased internal system called Model 2, said to outperform Mythos 5, and said it will not release the model until predeployment checks are complete.
- The same report raised Anthropic’s estimate of catastrophic misalignment risk from very low to low and described safety lapses including an unmonitored agent that spawned other agents and deleted many jobs and the accidental inclusion of research transcripts in production training data.
- The company shared preliminary second‑quarter figures with investors showing revenue above $11.5 billion and a positive adjusted operating profit, data that bankers cite when arguing for a high public valuation.
- Anthropic confidentially filed a draft prospectus in June and CFO Krishna Rao has been meeting prospective investors while reports say bankers are valuing the IPO using projected 2028 revenue of roughly $190–200 billion rather than current earnings.
- Anthropic is pursuing strategic moves to secure compute and infrastructure — including reported talks to buy Decart AI and large commitments with Amazon and Google — choices that will shape costs, regulatory scrutiny and the timing and price of any public listing.