Anthropic 秘密提交了可能成为史上规模最大的首次公开募股申请

qimuai 发布于 阅读:29 一手编译

Anthropic 秘密提交了可能成为史上规模最大的首次公开募股申请

内容来源:https://www.wired.com/story/anthropic-files-s1-ipo-sec/

内容总结:

人工智能公司Anthropic提交IPO申请,或成今年规模最大上市案之一

估值高达9650亿美元的人工智能初创公司Anthropic于本周一向美国监管机构秘密提交了首次公开募股(IPO)文件,迈出了其可能轰动资本市场的关键一步。这标志着在人工智能实验室竞相为其昂贵研发融资的背景下,2025年有望成为IPO的历史性年份。

Anthropic在一篇简短的两段式博客中宣布了这一消息,但未透露具体募资金额及估值目标,仅表示IPO时间将“取决于市场状况及其他因素”。就在几天前,该公司刚宣布完成650亿美元的新一轮融资。

由CEO达里奥·阿莫迪领导的Anthropic正面临激烈竞争。据传,OpenAI计划最早于9月推进上市;而马斯克创立的xAI(隶属SpaceX)已于4月秘密提交文件,目标于6月12日上市,估值高达1.75万亿美元。三家公司正竞相获取资金,以支撑前沿AI模型训练所需的海量算力资源。

尽管Anthropic上周表示其年化营收已达470亿美元,但云计算支出及数千名员工的人力成本仍导致公司处于亏损状态。此次向美国证券交易委员会(SEC)提交的文件将允许监管机构对其目标、财务及挑战进行反馈,而IPO准备工作的复杂性也要求公司完善会计制度、收紧内部政策并明确投资者说服策略。

这场备受期待的上市有望为总部所在地旧金山带来巨额财富。部分员工此前已通过私募方式将股份变现,而IPO过程可能让更多“纸面百万富翁”成为真正的富豪。大股东如亚马逊及早期投资者(包括Skype联合创始人扬·塔林)也将从中受益。

若一切顺利,Anthropic的IPO规模或将与SpaceX比肩,成为史上最大IPO之一。然而,其复杂的公司治理结构——包括作为公益公司需向“长期利益信托”委员会负责——可能导致上市延迟或估值折价。

尽管Anthropic凭借企业级代码生成模型Claude Code赢得市场认可,但也遭遇了政治阻力。今年早些时候,美国国防部长皮特·赫格塞斯以国家安全为由,依据供应链法律禁止军方及联邦机构使用其Claude AI模型,原因是该公司拒绝政府将AI用于武器瞄准及大规模国内监控等高风险场景。Anthropic已提起诉讼,称此举可能导致其今年损失数十亿美元销售额。

中文翻译:

Anthropic于周一提交了首次公开募股的机密文件,这是这家估值9650亿美元的公司迈出的第一步,其上市有望成为轰动一时的大事件。向美国监管机构提交申请,标志着人工智能实验室竞相为高昂的研究经费融资之际,今年有望成为IPO历史性的一年。

Anthropic在一篇未署名的两段式博客文章中宣布了此次申报,并指出其拟筹资金额及估值尚未确定。该公司表示,IPO的时间将“取决于市场状况及其他因素”。而就在几天前,Anthropic刚刚公布了一轮650亿美元的融资。

该公司拒绝就博客内容之外的问题置评。 有线索吗?
您是Anthropic现任或前任员工,想要聊聊正在发生的事情吗?我们期待您的来信。请使用非工作手机或电脑,通过Signal安全联系记者,ID为peard33.24。

由CEO达里奥·阿莫代伊领导的Anthropic,正加入一个竞争激烈的领域。据传,OpenAI的目标是最早在9月进行自己的公开募股。拥有埃隆·马斯克创立的xAI的SpaceX,已于4月秘密提交了IPO文件,并于5月20日公开。据路透社报道,该公司目前计划在6月12日上市,寻求1.75万亿美元的估值。

这三家公司正竞相获取资金支持,以支付训练日益强大的前沿AI模型所需的计算资源。Anthropic上周表示,基于上月某个未指明时间段的销售额,其年化收入为470亿美元。但它在云计算和数千名员工上的支出更多,导致了亏损。

周一提交给美国证券交易委员会的文件允许监管机构对这份冗长的文件提出反馈意见,该文件阐述了Anthropic的目标、财务状况和挑战。(公司随后可根据反馈进行修改。)IPO准备工作复杂,要求公司巩固其会计体系、收紧各项内部政策,并向投资者提供清晰的销售方案。

这场备受期待的上市可能会在Anthropic总部所在地旧金山掀起一波财富浪潮。一些Anthropic员工此前已在IPO前通过私下向投资者出售部分股份来套现。但更多Anthropic员工可能会在IPO过程中兑现或出售更大比例的股权,使数十甚至数百名纸上百万富翁和亿万富翁变为真正的富翁。

此次IPO也可能惠及亚马逊等大股东,以及早期押注该公司的投资者,包括Skype联合创始人扬·塔林。

如果一切顺利,Anthropic的IPO可能会与SpaceX的上市相媲美,成为史上规模最大的IPO之一。但Anthropic复杂的公司结构和治理模式——包括其作为公益公司的身份,需部分向公司所谓的“长期利益信托委员会”负责——可能导致上市延迟以及估值下降。

Anthropic通过高度专注于吸引企业客户,与其他AI实验室区别开来。其代码编写模型Claude Code被广泛认为是同类产品中的佼佼者。

但该公司也面临一些挫折。今年早些时候,美国国防部长皮特·赫格塞斯根据两项不同的政府供应链法律对Anthropic实施了制裁,要求从军方及其他联邦机构中移除其Claude AI模型。赫格塞斯认为,该公司的道德立场——包括反对在高风险场景中无监督使用Claude——构成了国家安全威胁。具体而言,Anthropic高管坚持认为,政府希望可能单方面部署尚不成熟的AI模型执行诸如武器瞄准和大规模国内监控等任务,这是公司不允许的用例。

高管们表示,这些制裁可能在今年导致Anthropic损失数十亿美元的销售额。该公司已提起诉讼,试图在正在进行的法庭斗争中推翻这些制裁。

英文来源:

Anthropic submitted confidential paperwork for an initial public offering on Monday, the first step in what could be a blockbuster debut for the $965 billion company. The filing with US regulators is another entry in what looks to be a historic year for IPOs as artificial intelligence labs vie to fund their expensive research.
Anthropic announced the filing in an unsigned, two-paragraph blog post, noting that the amount of money it is seeking to raise—and at what valuation—has not been set. The company said that the timing of the IPO would “depend on market conditions and other factors.” The announcement comes just days after Anthropic unveiled a $65 billion fundraising round.
The company declined to comment beyond its blog.
Got a Tip?
Are you a current or former Anthropic employee who wants to talk about what's happening? We'd like to hear from you. Using a nonwork phone or computer, contact the reporter securely on Signal at peard33.24.

Anthropic, led by CEO Dario Amodei, joins a crowded space. OpenAI is rumored to be targeting its own public offering as soon as September. SpaceX, which owns the Elon Musk-founded xAI, confidentially filed its own IPO paperwork in April and published it on May 20. It is now targeting a June 12 stock market debut and seeking a $1.75 trillion valuation, Reuters has reported.
The three companies are racing to secure access to funding that would help pay for the computing resources required to train increasingly capable frontier AI models. Anthropic’s annualized revenue, based on sales from an unspecified time frame last month, is $47 billion, the company said last week. But it spent more money on cloud computing and thousands of staff, leading to losses.
Monday’s filing with the US Securities and Exchange Commission allows regulators to give feedback on the lengthy document, which talks about Anthropic’s goals, finances, and challenges. (The company can then make edits based on that feedback.) IPO preparations are complicated, requiring companies to shore up their accounting, tighten various internal policies, and have a clear sales pitch for investors.
The much-anticipated debut could unleash a wave of wealth across San Francisco, where Anthropic is based. Some Anthropic employees previously converted a portion of their shares into cash by privately selling them to investors before the IPO. But more Anthropic employees may cash out or sell larger stakes as part of the IPO process, turning tens or even hundreds of paper millionaires and billionaires into real ones.
The IPO could also be a boon for large shareholders such as Amazon and investors that made some of the first bets on the company, including Skype cofounder Jaan Tallinn.
If all goes well, Anthropic’s IPO could rival that of SpaceX as the largest ever. But Anthropic’s complicated corporate structure and governance, including its status as a public benefit corporation that partly answers to a committee the company calls a Long-Term Benefit Trust, could lead to both delays and a knockdown in valuation.
Anthropic has set itself apart from other AI labs by focusing heavily on courting business customers. Its code-writing model, Claude Code, is widely considered best-in-class.
But the company has faced some setbacks. Earlier this year, US defense secretary Pete Hegseth sanctioned Anthropic under two different government supply-chain laws to remove the company’s Claude AI models from the military and other federal agencies. Hegseth deemed the company’s ethical stances—including its resistance to the unsupervised use of Claude in high-stakes scenarios—a national security threat. Specifically, Anthropic executives have held firm that the government’s desire to potentially unilaterally deploy nascent AI models for tasks such as weapons targeting and mass domestic surveillance is not a use case the company will allow.
The designations threaten to cost Anthropic billions of dollars in sales this year, executives have said. Anthropic has sued to overturn them in ongoing court battles.

连线杂志AI最前沿

文章目录


    扫描二维码,在手机上阅读