Overview
- Anthropic has filed an S-1 ahead of a planned November share offering that includes a valuation guidance as high as $2 trillion and extensive risk disclosures.
- The prospectus shows very large operating losses in the first half of 2026 together with substantial fundraising and revenue figures, underscoring the high cost of competing in advanced AI.
- Anthropic discloses that its current model, Claude, is being used to help develop the next generation, a practice reporters and analysts warn could create a rapid feedback loop that outpaces human verification.
- The filing explicitly warns of 'catastrophic or existential' harms, notes the possibility that powerful models could resist shutdown, and strengthens calls from the company and experts for mandatory independent safety tests.
- Coverage places the IPO in a wider financial and geopolitical picture where investor exposure is concentrated in a few AI firms, tech firms have flooded the bond market with risky data‑center debt, and analysts say binding domestic rules and international coordination are needed to manage systemic risk.