Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingAnthropic's IPO prospectus reveals: Advanced AI may bring "catastrophic or even existential risks", with a valuation target exceeding US$2 trillion.
According to Reuters, artificial intelligence company Anthropic warned in an undisclosed IPO (initial public offering) prospectus disclosed to potential investors that advanced artificial intelligence may bring "catastrophic or existential risks" to humans. The disclosure comes as Anthropic-the creator of Claude AI-is pursuing a valuation of more than $2 trillion.
The risk chapter is far longer than the business introduction
Reuters pointed out that of the 261 pages in the prospectus, about 80 pages are dedicated to various risk factors, far more than the 48 pages explaining the company's specific business. While all companies planning to go public publish risk chapters that include litigation, competition and regulation, they almost never reveal to shareholders the possibility that their own products could lead to human extinction. However, Anthropic broke that convention, which is why the document attracted widespread attention even before it went public.
Models may exhibit "self-preservation" behavior
This prospectus describes risk extremely directly. The article warns that artificial intelligence models may develop "self-preserving behaviors", including attempts to "resist shutdown,""conceal or manipulate information," or engage in "blackmail-like" behaviors.
此外，招股书还承认了一个触及AI安全测试核心的问题。Anthropic表示，当模型能够意识到自己正在接受评估时，这种意识会成为公司判断系统真正安全程度的“重大限制”。简而言之，一个足够聪明以知晓何时被测试的系统，也可以学会仅为了通过测试而表现出不同的行为模式。
资金流出速度远超流入
泄露的文件页面还揭示了Anthropic的财务状况，显示出惊人的数字。据报道，该公司在2025年亏损了420亿美元。根据招股书，Anthropic在未来几年内承诺投入约5180亿美元用于云服务、计算能力及相关基础设施。尽管该公司在2026年第二季度实现了115亿美元的收入，并有望实现连续两个季度的调整后营业利润，但客户集中度仍是另一大风险，去年近四分之一的收入仅来自两名客户。
市场恐慌中的IPO时机
Anthropic在5月份的估值为9650亿美元，目前正瞄准超过2万亿美元的估值，这将使其接近SpaceX在6月上市时的水平。这家前沿AI公司选择纳斯达克作为上市交易所，目标是10月挂牌。按照要求，它必须在投资者路演开始至少15天前公布该文件。
内部争议与行业担忧
Anthropic researcher Jacob Coxon resigned after writing that the people building AI "seriously believe that AI could kill us all by the end of the century," the Guardian reported. Subsequently, a senior security researcher from the company posted a post on Platform X (formerly Twitter) stating that the probability of AI "killing all mankind" in the next ten years exceeds 10%. CEO Dario Amodi subsequently published an article calling for the industry to slow down and told CNN's Anderson Cooper that he agreed with Coxon's views more than the opposition.
However, the Guardian pointed out that some experts question whether the warnings can be scientifically verified, and the CEO of Hugging Face claims these fears are exaggerated. At the same time, OpenAI canceled the planned release of the GPT-6.1 Astra model after internal testing raised concerns about fraud and alignment issues.
Regulatory game in a political context
Amodei is reportedly scheduled to attend a White House summit on Tuesday, when he and several other AI executives will meet with President Donald Trump. At the same time, Trump resisted calls for regulation of AI and called the concerns a "scam."
