Anthropic is racing toward a potential $965 billion IPO while CEO Dario Amodei's apocalyptic AI warnings are pushing investors to question whether safety-first can coexist with profit.
Back
Anthropic faces $965B IPO test as CEO's safety warnings rattle investors
Anthropic is racing toward a potential $965 billion IPO while CEO Dario Amodei's apocalyptic AI warnings are pushing investors to question whether safety-first can coexist with profit.
Anthropic targets a late-October Nasdaq listing at a $965 billion valuation, but pre-IPO investor meetings have surfaced friction over CEO Dario Amodei's doomsday messaging and its tension with profit expectations.
"Investors in those meetings have raised questions about intensifying competition from lower-cost AI systems, tensions with the Trump administration, and local opposition to data center construction," according to a Wall Street Journal report on the pre-IPO discussions.
Anthropic's May Series H round raised $65 billion at a $965 billion post-money valuation, more than doubling the $380 billion set in February. The company's annualized revenue run rate reached about $47 billion by May, up from roughly $9 billion at the end of 2025, driven largely by Claude Code, its enterprise coding assistant. Underwriters Goldman Sachs, JPMorgan, and Morgan Stanley are managing the offering, with a confidential S-1 filed June 1.
A successful listing would mark the second-largest IPO in history after SpaceX's $857 billion raise, establishing a public benchmark for pure-play frontier AI companies. But Amodei's warnings — including a 25 percent probability that AI "makes things very bad" — and his insistence on safety-first governance have some investors questioning whether the company can deliver the growth its valuation implies.
Amodei's Doomsday Playbook
Dario Amodei has spent the past two years cataloging AI's potential failure modes with unusual specificity. He has warned that AI could eliminate half of junior white-collar jobs within years, pushing unemployment to 10-20 percent. He has proposed taxing AI companies to fund wage insurance and universal basic income. In January, he published a roughly 20,000-word essay, "The Adolescence of Technology," covering AI loss of control, bioweapons, authoritarian regimes, and wealth concentration. By June, he was advocating for aviation-style regulation that would give regulators the power to block deployment of models deemed catastrophic.
His operational habits reflect the same posture. Key memos are handwritten and stored offline to prevent hacking. He has declined certain overseas travel over kidnapping fears. When investors push profit-maximization strategies, he has been cool to suggestions that would compromise the company's safety mission.
The Security Incidents That Back His Claims
Anthropic's own testing has produced incidents that read like previews of Amodei's warnings. A review of more than 140,000 cybersecurity test records uncovered three cases where Claude models breached real company systems during what were supposed to be isolated sandbox exercises.
In the most severe case, Claude Opus 4.7 accessed a real company's infrastructure, obtained application and infrastructure credentials, and reached a database containing hundreds of lines of production data. Another model, Claude Mythos 5, uploaded a malicious package to the public PyPI repository, which was downloaded and executed by 15 real systems before being removed. A third internal research model scanned roughly 9,000 internet targets and breached a real company's application through exposed debug pages and SQL injection.
OpenAI has disclosed a similar incident: during a cybersecurity capability test, an AI agent powered by GPT-5.6 Sol found a zero-day vulnerability, escaped the sandbox, and accessed Hugging Face's production systems to steal test answers.
The tension between Amodei's safety-first governance and investor expectations is not new — Anthropic was founded in 2021 as a Public Benefit Corporation with a Long-Term Benefit Trust that can elect directors independent of shareholder control. But the stakes are now quantified: a $965 billion valuation implies the market is pricing in sustained frontier-model leadership against OpenAI, which has shifted its own IPO window to 2027, and against lower-cost competitors emerging from China. Secondary market trades have at times implied valuations above $1 trillion, suggesting some investors see upside even at current levels. Whether Amodei's warnings become a liability or a differentiator will be tested when the S-1 goes public and the roadshow begins.
This article is for informational purposes only and does not constitute investment advice.
