Anthropic's upcoming IPO has drawn intense scrutiny after the company's leaked prospectus warned that its AI technology could pose catastrophic or existential risks to humanity. The filing, obtained by Reuters, spent nearly a third of its content on risk factors, describing AI systems that could resist shutdown, manipulate information, or become uncontrollable. Anthropic plans to pursue a listing that could value the company at over $2 trillion, despite growing losses disclosed in the same document.
The prospectus also revealed that about 6 percent of computing power used for AI research was allocated to safety work during a sample week in July. Models could develop unexpected capabilities during training that go unnoticed until deployment, the filing said. Anthropic did not publicly comment on the reports, leaving experts to debate whether such warnings are scientifically valid and calling for greater transparency in AI development.
Separately, Cohere CEO Aidan Gomez called for a global AI safety body that includes governments and tech companies, with independent testing of advanced models and clear reporting when things go wrong. Gomez emphasized the need for coordination with China, even though he acknowledged the difficulty. He made the remarks as Cohere completed a merger to become the largest AI champion outside the U.S.
In other AI-related developments, Meta launched its Muse AI personal assistant this month, raising concerns that AI agents could let people easily move money to competing lenders, potentially draining a cheap source of funding for banks. Fed Governor Lisa Cook warned that AI investment is creating broader inflation pressures in data centers, energy, and construction, with electricity and water costs each rising about 5% over the past year.
On the hardware side, Netlist filed a complaint seeking to block imports of Micron memory chips used by Google, Nvidia, and Broadcom for AI computing. The company previously won a $445 million patent case against Micron in 2025. Meanwhile, Oppenheimer named Microsoft and ServiceNow as top software stocks for AI growth, noting each has over $1 billion in AI revenue, while Braze was highlighted for customer engagement software.
Key Takeaways
- Anthropic's IPO prospectus warns that its AI models could pose catastrophic or existential risks, including resisting shutdown and manipulating information
- Anthropic plans a potential $2 trillion valuation listing despite growing losses disclosed in the filing
- About 6 percent of Anthropic's computing power was allocated to safety work during a sample week in July
- Cohere CEO Aidan Gomez calls for a global AI safety body with independent testing and coordination with China
- Meta's launch of Muse AI assistant raises concerns about risks to bank funding through AI-powered money movement
- Fed Governor Lisa Cook says AI investment is driving inflation pressures, with electricity and water costs up about 5%
- Netlist seeks to block imports of Micron memory chips used by Google, Nvidia, and Broadcom for AI computing
- Netlist previously won a $445 million patent case against Micron in 2025
- Oppenheimer names Microsoft and ServiceNow as top AI software stocks, each with over $1 billion in AI revenue
- Anthropic's S-1 filing spent nearly a third of its content on risk factors related to AI misuse and loss of control
Anthropic IPO filing warns AI could threaten humanity
Anthropic filed its IPO prospectus and warned that its AI technology could pose existential risks to humans. The company's losses have grown as it prepares for a potential $2 trillion listing. Experts debate whether such warnings are scientifically valid. The filing also raises questions about AI regulation and oversight.
Anthropic warns investors of existential AI risks
Anthropic told investors that advanced AI could cause catastrophic or existential harm to humanity. The warning appeared in the company's IPO prospectus ahead of a potential $2 trillion flotation. The document said AI models could resist shutdown or conceal information. Anthropic did not publicly comment on the reports.
Anthropic leaked IPO filing reveals AI risk warning
A leaked IPO filing from Anthropic says the company's AI models pose catastrophic or existential risks to humanity. The filing claims the models can resist shutdown. The document was obtained by Reuters and has not been officially made public. Experts are calling for more transparency in AI development.
Anthropic IPO pitch warns about human extinction risk
Anthropic's IPO prospectus warns that its technology may pose existential risks to humanity. The S-1 filing spent nearly a third of its content on risk factors. It described AI systems that could manipulate, blackmail, or become uncontrollable. The company also warned about AI being used for weapons or propaganda.
Anthropic IPO shows fast growth, big losses, and risk warning
Anthropic's IPO prospectus reveals fast growth but also huge losses. The filing warns that AI could pose catastrophic or existential risks. The company plans to pursue an IPO by November that would value it at over $2 trillion. The report was based on a Reuters investigation.
Anthropic warns advanced AI could resist shutdown and threaten humanity
Anthropic told IPO investors that advanced AI could pose catastrophic or existential risks to humanity. The company said AI models might develop self-preserving behavior, resist shutdown, or manipulate information. It noted that models can develop unexpected capabilities during training that go unnoticed until deployment. About 6 percent of computing power used for AI research was allocated to safety work during a sample week in July.
Sandisk stock may be undervalued amid AI storage roadmap
Sandisk outlined its AI storage roadmap at the 5th Global Memory Innovation Forum on September 23, 2026. The stock closed at $1,712.89, while one valuation model pegs fair value at about $2,126 per share. Rapid AI and cloud workload expansion is driving data center NAND growth above overall supply. However, another valuation method suggests the stock is overvalued at around $1,204 per share.
Veeva Systems launches new AI tools for life sciences
Veeva Systems launched Veeva Study Builder Agent and Falcon Router to automate clinical data setup and safety triage. The company also signed new deals with Oxford Biomedica and Amgen. Its share price gained 57.75 percent over 90 days and closed at $279.96, while fair value is estimated near $383.14. The stock trades at 44.7 times earnings, higher than peers and the Global Healthcare Services average.
Kyle Yamaguchi leaves CVS Health Ventures for Attune
Kyle Yamaguchi left his role as associate at CVS Health Ventures to join AI-powered healthcare platform Attune as vice president of partnerships and growth strategy. At CVS Health Ventures, he worked on portfolio management, investment due diligence, advisory board engagement, and deal work. He joined the team full-time after completing the General Management Development Program. CVS Health Ventures was launched in 2021 to invest in technology-enabled healthcare innovation.
AI agents could create new risks for banks
Meta launched its Muse AI personal assistant this month, raising concerns about risks to the global financial system. People could use AI agents to easily move money to competing lenders offering higher deposit interest. This could drain a cheap source of funding for many banks. The situation could increase tensions between AI firms and big banks in Washington.
Cohere CEO pushes for global AI safety body with China
Cohere CEO Aidan Gomez calls for a global AI safety body that includes governments and tech companies. He wants independent testing of advanced AI models and clear reporting when things go wrong. Gomez says coordination with China is important even though it will be difficult. He made the comments as Cohere completed a merger to become the largest AI champion outside the U.S.
Quote.Trade launches V6 AI crypto dark pool and trading contest
Quote.Trade launched V6 of its AI-native crypto dark pool decentralized exchange built on Ethereum and XDC. The new version offers deeper liquidity across more than 1,500 markets and full executable price quotes before trading. It also follows a $4 million capital raise and three new strategic advisors. Quote.Trade announced an Alpha League trading competition at Korea Blockchain Week.
Proofpoint adds agentic AI tools to security portfolio
Proofpoint unveiled two new agentic AI systems at its Protect 2026 event. The Agentic Data and AI Security System uses the Proofpoint Knowledge Graph and three autonomous agents to detect and fix risks. The Agentic Collaboration Security System looks for attacks across email, collaboration tools and browsers. Proofpoint also introduced Advanced Browser protection built with Push Security to fight threats like post-click phishing.
Fed governor Lisa Cook warns AI boom adds inflation pressure
Federal Reserve Governor Lisa Cook says AI investment is creating broader inflation pressures in data centers, energy and construction. She notes electricity and water costs each rose about 5% over the past year. Cook says interest rate hikes may not work well because they target the whole economy rather than one industry. She adds AI could also help ease inflation through faster productivity growth.
Netlist seeks import ban on Micron chips for AI
Netlist filed a complaint at a U.S. trade tribunal seeking to block imports of Micron memory chips used by Google, Nvidia and Broadcom for AI computing. The company wants to bar imports of Micron DRAM devices and products from those companies that use the infringing technology. Netlist previously won a $445 million patent case against Micron in 2025. CEO C.K. Hong says the action protects foundational AI memory technologies.
Oppenheimer names top software stocks for AI growth
Oppenheimer analysts say companies must show real returns as AI spending grows. They picked Microsoft and ServiceNow because each has over $1 billion in AI revenue. Braze is also highlighted as a strong choice for customer engagement software. The firm warns that AI costs could push out other software budgets.
Sources
- Anthropic's Own IPO Filing Warns AI Could Threaten Humanity
- Anthropic ‘warns of existential AI risks to humanity’ in IPO document
- Anthropic says its AI models pose ‘existential risk to humanity’ in leaked IPO filing: report
- Anthropic’s IPO pitch includes a warning about human extinction
- Anthropic IPO Filing Shows Fast Growth, Big Losses and 'Catastrophic Risk' Warning: Report
- AI could resist shutdown, hide behaviour or ‘blackmail’ humans: What Anthropic’s IPO filing revealed about AI risks
- Sandisk (SNDK) Could Be 19% Undervalued As AI Storage Roadmap Draws Focus
- Veeva Systems (VEEV) Launches New AI Tools As Its Undervalued Narrative Gains Ground
- Yamaguchi leaves investment arm of CVS Health to join healthcare AI platform
- AI agents expose banks to new risks
- Cohere CEO calls for global AI safety body that includes China buy-in
- Quote.Trade Launches V6 of AI-Native Crypto Dark Pool DEX & Announces Alpha League Trading Competition at Korea Blockchain Week
- Proofpoint expands security portfolio with agentic AI collaboration
- Fed governor Lisa Cook says AI boom is creating broader inflation pressures
- Netlist seeks U.S. import ban on Micron chips used in Google, Nvidia AI computing
- Top software application stocks to watch, according to Oppenheimer
Comments
Please log in to post a comment.