Anthropic CEO Dario Amodei 发出 3800 字警告，AI 股票遭重挫

一度近乎“刀枪不入”的 AI 交易如今正式受到冲击。

周一盘前交易中，AI 股票遭遇重挫。Anthropic（ANTH.PVT）CEO Dario Amodei 发布了一篇悲观的 3,800 字文章，动摇了整个半导体与 AI 基础设施领域的投资者信心。

动量型 AI 股票首当其冲：Marvell（MRVL）、SK Hynix（SKHY）、Micron（MU）和 Super Micro（SMCI）均下跌 6%；CoreWeave（CRWV）、SanDisk（SNDK）、Intel（INTC）和 AMD（AMD）的跌幅也都超过 4%。

即便是 AI 领域最重要的两家公司——Nvidia（NVDA）和 Broadcom（AVGO）——在开盘后也下跌了 3%。

仅仅两天之内，Anthropic（ANTH.PVT）CEO Dario Amodei 和 OpenAI（OPAI.PVT）CEO Sam Altman，就凭借他们创造、并推动由顶尖程序员组成的团队加速发展的强大技术，成功让全人类感到不安。

Amodei 在周六发布的一篇文章中表示：“在过去几个月里，我逐渐确信，要充分应对这些风险，需要更加审慎——不仅要投资于风险防范，还要控制能力提升的速度，以便风险防范有时间跟上。”

“我们必须放缓提升 A.I. 模型能力的步伐。进展看起来仍会很快，而我们必须明智地利用所获得的时间。”

Altman 在该文发布后不久表示，他同意这位竞争对手的看法。

随后，他在周一凌晨发布了一则语气阴郁的 X 帖文。

Altman 在美东时间凌晨约 12:57 发布的X 帖文中表示：“AI 的发展有两种可能变得极其糟糕的方式，而我们必须避免它们。第一种是，我们可能将对未来的控制权交给 AI。这是不可接受的；我们毫不掩饰地站在人类一边，AI 必须始终服务于人类。为确保这一点，我们需要找到方法，确保对齐与安全技术始终领先于模型能力的进步。”

他继续说：“第二种是，我们可能最终生活在一个权力过度集中的世界。如果某个拥有非凡能力的 AI 被一个人或一家公司用来将其世界观强加给所有人，结果可能会极其反乌托邦。要避免这两种威胁，需要在一条狭窄的中间道路上前行；例如，一个国家可能获得过多权力。另一个例子是，某一家实验室最终拥有过多权力。”

Bernstein 分析师 Madison Rezaei 在一份报告中解释称，这一框架的核心实际上是尽可能降低安全风险，包括引入独立审查者、协调安全标准以及开展国际合作。

Rezaei 补充道：“目前，这并不是要求降低资本开支，或停止模型训练。然而，许多投资者已经开始质疑：如果训练放缓，会发生什么？”

在 Rezaei 覆盖的公司中，对任何 AI 放缓最为敏感的公司是 CoreWeave。

Rezaei 表示：“我们估计，其 [CoreWeave] 现有已启用美国电力容量中，有 25% 位于 Tier 3 和 Tier 4 市场；其已签约电力容量中，另有约 74% 位于这些市场。积压订单主要由不论是否使用均需付费的 take-or-pay 合同构成，因此我们预计这部分不会受到威胁；但如果训练开发放缓，已签约但尚未售出的农村地区电力容量需求可能会出现回落。”

Brian Sozzi 是 Yahoo Finance 执行编辑、《Power Players With Brian Sozzi》播客主持人，也是 Yahoo Finance 编辑领导团队成员。可在 X 平台关注 Sozzi：@BrianSozzi、Instagram 和 LinkedIn。如有新闻线索，请发送邮件至 brian.sozzi@yahoofinance.com。

The often-teflon AI trade has officially been rattled.

AI stocks are getting absolutely crushed in premarket trading on Monday after Anthropic (ANTH.PVT) CEO Dario Amodei dropped a gloomy 3,800-word essay that rattled investor confidence across the entire semiconductor and AI infrastructure complex.

Momentum AI names took it on the chin, with Marvell (MRVL), SK Hynix (SKHY), Micron (MU), and Super Micro (SMCI) each down 6%, while CoreWeave (CRWV), SanDisk (SNDK), Intel (INTC), and AMD (AMD) were all off by more than 4%.

Even the two most important names in AI — Nvidia (NVDA) and Broadcom (AVGO) — are sliding 3% after the opening bell.

Within the course of two days, AI titans at Anthropic (ANTH.PVT), CEO Dario Amodei, and OpenAI (OPAI.PVT), CEO Sam Altman, have managed to scare the hell out of the human race with the powerful technology they created and pushed their teams full of pioneering super-coders to supercharge.

"Over the last few months, I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up," Amodei stated in an essay on Saturday. "We must slow the pace at which we improve the capabilities of A.I. models. Progress will still seem fast, and we must make wise use of the time we gain."

Altman said right after the post that he agreed with his rival.

He then followed that up on Monday with a dark X post early on today.

"There are two ways AI progress could go very badly, and that we must avoid," Altman said in an X post around 12:57 am ET. "First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve people. To ensure that, we need ways to ensure that alignment and safety techniques stay ahead of progress in model capabilities."

He continued, "Second, we could end up in a world with too much concentration of power. If an extraordinarily powerful AI is used by one person or company to impress their worldview onto everyone else, the results could be extremely dystopian. Avoiding these two threats requires walking a narrow middle path; for example, one country could gain too much power. Another example is one lab ending up with too much power."

The framework is really about minimizing security risks — independent reviewers, coordination on safety standards, and international cooperation, Bernstein analyst Madison Rezaei explained in a note.

"At this point, it's not a call for a lowering of capex or stopping model training. However, many investors have started to question what happens if training slows," Rezaei added.

The most exposed name in Rezaei's coverage universe to any AI slowdown is CoreWeave.

"We estimate that 25% of their [CoreWeave] existing active U.S. power, with an additional ~74% of their contracted power, is situated in Tier 3 and Tier 4 markets. The backlog is primarily comprised of take-or-pay contracts, so we do not anticipate a threat there, but could see a pullback in demand for the contracted, not-yet-sold rural power if training development slows," said Rezaei.

Brian Sozzi is Yahoo Finance's Executive Editor, host of the 'Power Players With Brian Sozzi'podcast and a member of Yahoo Finance's editorial leadership team. Follow Sozzi on X @BrianSozzi ,Instagram , and LinkedIn . Tips on stories? Email brian.sozzi@yahoofinance.com.

Click here for in-depth analysis of the latest stock market news and events moving stock prices

Read the latest financial and business news from Yahoo Finance