人工智能初创公司 Anthropic 于当地时间 11 月 19 日正式在纳斯达克挂牌上市,以约 813 亿美元的募资规模打破由 SpaceX 保持的 750 亿美元纪录,成为历史上规模最大的首次公开募股 146。发行价定为每股 83 美元,完全稀释后的市值预计达到 1.5 万亿美元 13。尽管部分来源报道首日收盘价为 97.17 美元(较发行价上涨约 17%),对应市值约 1.27 万亿美元,但另有数据显示开盘后股价曾飙升至 111.5 美元,随后在次日盘前继续下跌约 9%,显示出市场对其估值的剧烈波动与分歧 3456。此次 IPO 不仅使 Anthropic 成为首家上市的前沿 AI 研发机构,也被华尔街视为观察全球 AI 投资热潮的重要风向标 1。
财务数据显示,Anthropic 营收呈现爆发式增长,截至 10 月底其年化收入运行率已突破 810 亿美元,2026 年上半年营收达 162 亿美元 134。然而,公司同期账面净亏损高达约 800 亿至 807 亿美元,每股收益为负值,巨额亏损主要源于算力基础设施投入及公允价值变动等一次性费用 146。招股书披露,Anthropic 承诺到 2045 年在计算基础设施等方面投入约 9000 亿美元,其中一半需在 2030 年前完成,并与亚马逊、谷歌等云厂商签署了接近这一规模的长期算力协议 36。尽管股票认购超额倍数约为 6 倍,且仅约 15% 的股份分配给散户投资者,但分析师警告此举可能引发市场波动,且有观点认为此次 IPO 未能提振整体 AI 板块信心,甚至导致亚洲股市出现自 2020 年以来最糟糕的一周表现 35。
在商业成功之外,Anthropic 的“AI 安全”叙事及其技术风险引发了广泛争议。CEO 达里奥·阿莫代伊承认技术存在被滥用、经济破坏及失控等真实风险,并承诺不会为了商业利益或政治压力而牺牲 AI 安全,同时呼吁放缓最先进模型的研发速度 136。然而,就在上市前夕,多位图灵奖和诺贝尔奖得主签署《反 Anthropic 宣言》,批评该公司言行不一,指责其删除了“若安全措施不到位则暂停训练”的承诺 6。此外,竞争对手 OpenAI 已将原定今年的上市计划推迟至 2027 年,理由同样涉及对技术安全的担忧,这使得 Anthropic 的上市表现成为行业内外审视 AI 治理与商业化平衡的关键案例 3。
Artificial intelligence developer Anthropic completed the largest initial public offering in history on November 19, raising $81.34 billion by listing its shares on the Nasdaq 146. The company’s debut surpassed the previous record held by SpaceX, which had raised $75 billion in a prior offering 134. Shares were priced at $83 per share, valuing the company at approximately $1.27 trillion based on the closing price, though fully diluted estimates reached $1.5 trillion 1346. On the first day of trading, the stock opened at roughly $111 to $111.50 before settling at $97.17, representing a 17% increase from the issue price 346. Despite this surge, market volatility persisted, with pre-market shares dropping approximately 9% on the following day, November 20 6.
The financial landscape of Anthropic presents a stark contrast between explosive revenue growth and significant losses driven by heavy infrastructure spending. As of late October, the company reported an annualized revenue run rate exceeding $81 billion 34. In the first half of 2026, revenues reached $16.2 billion, while total revenues for the first nine months stood at $33.24 billion 16. However, growth slowed significantly, with third-quarter revenue rising 47% compared to 144% in the second quarter 3. The firm recorded a net loss of approximately $80 billion to $80.7 billion during the first half of the year or nine-month period, largely attributed to one-time expenses and fair value changes related to massive computing commitments 16. To support its AI models, Anthropic has signed long-term agreements with cloud providers like Amazon and Google, pledging nearly $900 billion for computational infrastructure through 2045, with half of that expenditure scheduled before 2030 36. The IPO was heavily oversubscribed, with only about 15% of shares allocated to retail investors, while analysts noted that nearly all employees became millionaires due to the listing 235.
Anthropic’s entry into public markets has sparked broader debates regarding AI safety, governance, and market confidence. CEO Dario Amodei publicly acknowledged real risks associated with AI technology, including potential misuse and economic disruption, stating that he would not sacrifice AI safety for commercial interests or political pressure 36. This stance contrasts with reports that the company removed a prior commitment to pause training if safety measures were inadequate, leading Turing Award and Nobel laureates to sign an "Anti-Anthropic Manifesto" criticizing the firm’s inconsistency just days before the IPO 6. While some commentators viewed the successful listing as a vote of confidence in the sector, others pointed to simultaneous declines in Asian stock markets and weak performance in other AI-related equities as signs of underlying market fragility 35. Additionally, rival OpenAI postponed its own planned IPO to 2027, citing concerns over technological safety, suggesting that Anthropic’s path serves as a complex benchmark for the industry 3.
评论
还没有评论,欢迎留下第一条。