Anthropic seeks a $2 trillion Claude IPO after its alignment lead warns AI could kill all humans this decade
Published by reddit.com
A source is testimony, not proof. Citing one is a provenance fact — it records who said what, where — and never a verification that the claim is true.
Who is named
The characters and things this document talks about — 8 in alls. Each links to everything else the record knows about them.
People 3
Organisations 1
Claims made
Every claim in this document. Firsthand = asserted on its own basis; secondhand = recounted from elsewhere, carrying no independent weight.
-
Anthropic filed a confidential S‑1 on June 1, 2026 targeting a fall Claude listing near $2 trillion.
Anthropic filed a confidential S-1 on June 1, 2026, targeting a fall Claude listing near $2 trillion.
-
Anthropic staff earnestly believe AI could kill all humans with above 10 percent chance this decade.
Alignment Science lead Evan Hubinger wrote on September 8, 2026 that staff “earnestly believe AI could kill all humans,” above 10 percent this decade.
-
Claude runs workplace chat, code, and agents.
Claude already runs workplace chat, code, and agents.
-
Anthropic’s September 2026 threat report says it disrupted Claude use for cyber operations, weapons software, surveillance, and biological research.
-
People familiar with Anthropic estimated year‑end annualized revenue above $100 billion.
People familiar with the company put year-end annualized revenue above $100 billion.
-
On September 3, 2026, Sen. Bernie Sanders and Rep. Greg Casar announced a pause on advanced AI and a ban on systems that can subvert shutdown.
-
Item 105 of Regulation S‑K requires material‑risk discussion.
Item 105 of Regulation S-K requires material-risk discussion.
-
No statute covers civilization‑scale model harm; securities and tort law would.
No statute covers civilization-scale model harm; securities and tort law would.
-
Company policy page (updated August 14, 2026) describing voluntary catastrophic-risk thresholds, risk reports, and the option to pause scaling.
-
Anthropic threat‑intelligence report on Claude misuse it says it disrupted from December 2025 through August 2026 across cyber, weapons, surveillance, biological, fraud, and distillation cases.
-
Dario Amodei estimated a 25 percent chance that things go really, really badly.
Amodei’s 25% chance that things go really, really badly estimate at the Axios AI+ DC Summit on September 17, 2025.
No claims match this filter.
What it cost
Reading this document into the graph cost <$0.01.
11 claims in all. Content fingerprint 186ab5f8eea6… —
the document's text itself is not republished here; follow the link above to read it at
its source.