Anthropic联合创始人兼首席执行官达里奥·阿莫代伊现身白宫会议。画面前景中,是当天刚刚辞职的英伟达首席执行官黄仁勋。图片来源:Photo by Kent NISHIMURA / AFP via Getty Images
Anthropic即将启动史上最受瞩目的IPO之一。然而风云突变,就在距离即将上市之际,这个行业却痛失最具影响力的人物与头号“推销员”。
上周四,英伟达(Nvidia)联合创始人、长期担任首席执行官的黄仁勋辞去公司职务。他此前被曝出两段婚外情:一名对象是公司客户,另一名则是他在英伟达的直属高管。
这个微妙的节点给Anthropic的上市增添了新压力。市场本就将此次IPO视作检验投资者AI热情的“压力测试”。按计划,这家AI实验室将于本周三或周四挂牌,就在此前几天,英伟达将发布该公司史上首份没有黄仁勋掌舵的财报。
英伟达是Anthropic的紧密合作伙伴,也深度参与了这家AI实验室的IPO。根据Anthropic的申报文件,英伟达已同意在一笔与IPO同步进行的260亿美元私募配售中,认购约100亿美元股份。两家公司在商业层面也高度绑定。Anthropic已承诺投入数百亿美元,以保障英伟达最先进芯片的供应。
投资者和分析师对《财富》杂志表示,他们预计,英伟达的震荡不会波及Anthropic的核心业务,但可能冲击整个AI行业的叙事逻辑。
华尔街给予英伟达股票“卖出”评级的分析师寥寥无几,Seaport Research Partners分析师杰伊·戈德堡便是其中之一。他并不认为黄仁勋的离职会让这笔IPO脱轨。
他表示:“他们将顶着英伟达财报电话会议的压力上市,但我认为这不会让上市计划受阻。这笔IPO体量庞大,背后又有一众机构力挺,他们不会轻易撤回上市计划。”
戈德堡表示,更大的风险将在Anthropic上市后显现。在黄仁勋正式离职前夕、媒体爆出他可能离开英伟达之后,戈德堡曾警告称,英伟达的动荡,包括首席执行官离职,会在Anthropic发布上市后首份季报时带来压力。他表示:“至少,这会引发质疑:‘AI叙事能否维持下去?这种投入还能持续吗?’他们将面临更棘手的问题。”
风投机构M13合伙人卡尔·阿洛马尔也认为,Anthropic的基本面未发生变化。他表示:“在我看来,Anthropic现在的处境和这则消息传出之前一样良好、稳健。Anthropic的财务表现极为出色,公司也已充分证明,其现有算力储备足以支撑业务需求,而且资金消耗相当节制。”
与部分竞争对手相比,Anthropic对英伟达的依赖更低。除英伟达硬件之外,公司部分模型还运行在谷歌(Google)的TPU、亚马逊(Amazon)的Trainium芯片以及AMD的GPU上。据报道,Anthropic还一直在探索自研芯片。
曾在沃顿商学院(Wharton School)任教的会计师、调查记者弗朗辛·麦肯纳表示:“英伟达已承诺在Anthropic IPO中投资100亿美元。待IPO最终启动时,这笔投资很可能仍会落地。但如果有人原本指望英伟达在必要时出手救助CoreWeave,现在恐怕得另寻出路了。”
不过,乔治敦大学(Georgetown University)麦克多诺商学院(McDonough School of Business)副教授杰森·施洛策表示,英伟达近期的风波以及围绕其首席执行官判断力的质疑,也可能波及其合作伙伴,招致外界更严格的审视。他预计市场会“更仔细地审查Anthropic与英伟达及其关联客户的关系”。
对硅谷而言,Anthropic与竞争对手OpenAI吸纳了海量资本,几乎所有风投机构均持有其股份——这两家俨然成了私募市场的“指数基金”。这也意味着,Anthropic预计下周进行的IPO,将成为一次事关重大的流动性事件——不仅关系到Anthropic超过400家投资方,也关系到整个AI热潮。市场普遍认为,如果Anthropic成功上市,AI热潮仍有广阔的延伸空间;如果IPO失败或被无限期推迟,将释放明确的负面信号,表明AI泡沫正走向破灭。
但事情未必会演变成一场崩盘。事实上,至少有一位风投人士认为,眼下的局面反而对Anthropic构成明确利好。Anthropic首席执行官达里奥·阿莫代伊长期主张更严格管制对华芯片出口,一直与黄仁勋意见相左。在AI监管问题上,这家AI实验室也与黄仁勋存在分歧。黄仁勋尤其反对设置许可要求,认为这会加大客户采用开源AI模型的难度。
一位投资人表示:“要说影响,我反而认为Anthropic是此事的受益者。他们跟黄仁勋本就没有那么亲近……而且他们正在自研芯片,如今供应链层面的掣肘得以解除,对Anthropic而言无疑是件好事。”因事态敏感且自身在AI领域布局广泛,该投资人要求匿名。(财富中文网)
译者:刘进龙
审校:汪皓
Anthropic即将启动史上最受瞩目的IPO之一。然而风云突变,就在距离即将上市之际,这个行业却痛失最具影响力的人物与头号“推销员”。
上周四,英伟达(Nvidia)联合创始人、长期担任首席执行官的黄仁勋辞去公司职务。他此前被曝出两段婚外情:一名对象是公司客户,另一名则是他在英伟达的直属高管。
这个微妙的节点给Anthropic的上市增添了新压力。市场本就将此次IPO视作检验投资者AI热情的“压力测试”。按计划,这家AI实验室将于本周三或周四挂牌,就在此前几天,英伟达将发布该公司史上首份没有黄仁勋掌舵的财报。
英伟达是Anthropic的紧密合作伙伴,也深度参与了这家AI实验室的IPO。根据Anthropic的申报文件,英伟达已同意在一笔与IPO同步进行的260亿美元私募配售中,认购约100亿美元股份。两家公司在商业层面也高度绑定。Anthropic已承诺投入数百亿美元,以保障英伟达最先进芯片的供应。
投资者和分析师对《财富》杂志表示,他们预计,英伟达的震荡不会波及Anthropic的核心业务,但可能冲击整个AI行业的叙事逻辑。
华尔街给予英伟达股票“卖出”评级的分析师寥寥无几,Seaport Research Partners分析师杰伊·戈德堡便是其中之一。他并不认为黄仁勋的离职会让这笔IPO脱轨。
他表示:“他们将顶着英伟达财报电话会议的压力上市,但我认为这不会让上市计划受阻。这笔IPO体量庞大,背后又有一众机构力挺,他们不会轻易撤回上市计划。”
戈德堡表示,更大的风险将在Anthropic上市后显现。在黄仁勋正式离职前夕、媒体爆出他可能离开英伟达之后,戈德堡曾警告称,英伟达的动荡,包括首席执行官离职,会在Anthropic发布上市后首份季报时带来压力。他表示:“至少,这会引发质疑:‘AI叙事能否维持下去?这种投入还能持续吗?’他们将面临更棘手的问题。”
风投机构M13合伙人卡尔·阿洛马尔也认为,Anthropic的基本面未发生变化。他表示:“在我看来,Anthropic现在的处境和这则消息传出之前一样良好、稳健。Anthropic的财务表现极为出色,公司也已充分证明,其现有算力储备足以支撑业务需求,而且资金消耗相当节制。”
与部分竞争对手相比,Anthropic对英伟达的依赖更低。除英伟达硬件之外,公司部分模型还运行在谷歌(Google)的TPU、亚马逊(Amazon)的Trainium芯片以及AMD的GPU上。据报道,Anthropic还一直在探索自研芯片。
曾在沃顿商学院(Wharton School)任教的会计师、调查记者弗朗辛·麦肯纳表示:“英伟达已承诺在Anthropic IPO中投资100亿美元。待IPO最终启动时,这笔投资很可能仍会落地。但如果有人原本指望英伟达在必要时出手救助CoreWeave,现在恐怕得另寻出路了。”
不过,乔治敦大学(Georgetown University)麦克多诺商学院(McDonough School of Business)副教授杰森·施洛策表示,英伟达近期的风波以及围绕其首席执行官判断力的质疑,也可能波及其合作伙伴,招致外界更严格的审视。他预计市场会“更仔细地审查Anthropic与英伟达及其关联客户的关系”。
对硅谷而言,Anthropic与竞争对手OpenAI吸纳了海量资本,几乎所有风投机构均持有其股份——这两家俨然成了私募市场的“指数基金”。这也意味着,Anthropic预计下周进行的IPO,将成为一次事关重大的流动性事件——不仅关系到Anthropic超过400家投资方,也关系到整个AI热潮。市场普遍认为,如果Anthropic成功上市,AI热潮仍有广阔的延伸空间;如果IPO失败或被无限期推迟,将释放明确的负面信号,表明AI泡沫正走向破灭。
但事情未必会演变成一场崩盘。事实上,至少有一位风投人士认为,眼下的局面反而对Anthropic构成明确利好。Anthropic首席执行官达里奥·阿莫代伊长期主张更严格管制对华芯片出口,一直与黄仁勋意见相左。在AI监管问题上,这家AI实验室也与黄仁勋存在分歧。黄仁勋尤其反对设置许可要求,认为这会加大客户采用开源AI模型的难度。
一位投资人表示:“要说影响,我反而认为Anthropic是此事的受益者。他们跟黄仁勋本就没有那么亲近……而且他们正在自研芯片,如今供应链层面的掣肘得以解除,对Anthropic而言无疑是件好事。”因事态敏感且自身在AI领域布局广泛,该投资人要求匿名。(财富中文网)
译者:刘进龙
审校:汪皓
Anthropic is about to embark on one of the most hotly anticipated IPOs in history. After a dramatic turn of events, it will now be doing so a week after the industry lost one of its most powerful figures and best salesmen.
Nvidia’s longtime CEO and co-founder Jensen Huang resigned from the company on Thursday following reports alleging he had engaged in two extramarital affairs: one with a customer suspected by U.S. intelligence to be a Chinese intelligence asset and another with an Nvidia executive who reports to him.
The timing has thrown new pressure on Anthropic’s listing, which was already seen as a stress test for investors’ appetite for AI. The AI lab is expected to list on Wednesday or Thursday next week, days after Nvidia reports quarterly earnings—the company’s first-ever without Huang at the helm.
Nvidia is a close partner of Anthropic and intimately involved in the lab’s upcoming IPO. According to Anthropic’s filings, the chipmaker has agreed to buy about $10 billion of shares in a $26 billion private placement alongside the IPO. The two companies are also closely tied commercially, with Anthropic committing to tens of billions of dollars to secure access to Nvidia’s most advanced chips.
Investors and analysts told Fortune that while they don’t expect the Nvidia shock to affect Anthropic’s core business, it could cause problems for the broader AI story.
Jay Goldberg, an analyst at Seaport Research Partners and one of the few on Wall Street with a sell rating on Nvidia’s stock, doesn’t expect Huang’s exit to derail the deal.
“They’re going to go public right in the face of Nvidia’s earnings call, and I don’t think that derails it,” he said. “It’s such a massive IPO and there’s so many people behind it now that they don’t pull the IPO.”
The bigger risk, he said, comes once Anthropic is a public company. Speaking shortly before Huang’s departure but after reports that he was likely to be leaving the company, Goldberg warned that turmoil at Nvidia, including the loss of its CEO, would create a headwind for Anthropic when it reports its first quarterly results. “At the very least it will create this narrative of, ‘Can the AI narrative maintain? Can we keep doing this?’” he said. “They’ll get much tougher questions.”
Karl Alomar, a partner at venture firm M13, agreed that Anthropic’s fundamentals haven’t changed. “From where I stand, Anthropic is still in as good, as strong a position as it was before this news came out,” he said. “Anthropic’s numbers are fantastic, and they’ve clearly shown that they have excess capacity for what they need. They haven’t spent that much money.”
Anthropic is also less dependent on Nvidia than some of its rivals, running some models on Google’s TPUs, Amazon’s Trainium chips, and AMD’s GPUs, as well as Nvidia’s hardware. Anthropic has also reportedly been exploring building its own chips.
“Nvidia has committed to invest $10 billion in the Anthropic IPO. That will likely still happen when the IPO eventually happens,” said Francine McKenna, an accountant and investigative journalist who has taught at the Wharton School. “But anyone who was looking at Nvidia to bail out CoreWeave if needed should probably look elsewhere.”
However, the recent trouble at Nvidia and questions surrounding the CEO’s judgement could spill over into more scrutiny for its partners, according to Jason Schloetzer, an associate professor at Georgetown University’s McDonough School of Business. He said he expected there to be “much closer scrutiny of [Anthropic’s] ties to Nvidia and customers associated with Nvidia.”
For Silicon Valley, Anthropic and competitor OpenAI have needed so much capital that almost every venture capital firm in town owns a piece of them—making them the private market equivalent of an index fund. That makes Anthropic’s expected IPO next week a high-stakes liquidity event—not only for the more than 400 backers of Anthropic, but for the AI boom as a whole. There’s a broad sense that if Anthropic succeeds, the AI boom has far more runway. If it flops or pushes back indefinitely, it would be an unequivocal negative signal that the AI bubble is heading towards a pop.
It may not be a bust. Indeed, at least one venture capitalist sees a clear upside for Anthropic. Anthropic CEO Dario Amodei has long pushed for tighter controls on chip exports to China, putting him at odds with Huang. The AI lab has similarly clashed with Huang on AI regulation, with Huang particularly keen to avoid any licensing requirements that would make it more difficult for customers to adopt open-source AI models.
“If anything, I think Anthropic is a beneficiary here,” said the investor, who asked not to be named given the situation’s sensitivity and their investments across AI. “They weren’t that close to Jensen…They’re building their own chip, and so the supply chain considerations that get freed up are good for Anthropic.”