AI高管倡议放缓前沿研发，全球芯片与科技股遭遇重挫Global Tech and Chip Stocks Tumble as AI Leaders Urge Slowdown

美股盘前交易中，纳斯达克期货下跌1.8%，主要受到芯片制造商英伟达与英特尔盘前走低拖累。市场抛售情绪迅速蔓延至全球其他主要金融市场。在亚洲，韩国综合股价指数（Kospi）下跌3.3%，其中芯片制造巨头SK海力士单日跌幅达6.4%；作为OpenAI主要投资方之一的日本软银集团，在东京股市收盘大跌近11%。在欧洲交易时段，荷兰半导体设备制造巨头阿斯麦股价下挫6%。

U.S. equity futures weakened sharply in premarket trading, with Nasdaq futures falling 1.8% as semiconductor giants Nvidia and Intel both declined ahead of the opening bell. The selling pressure quickly spread across international markets. In Asia, South Korea's benchmark Kospi index slid 3.3%, dragged lower by a 6.4% drop in chipmaker SK Hynix. In Tokyo, SoftBank Group, a key backer of ChatGPT creator OpenAI, finished the trading day down nearly 11%. European technology equities also suffered, with Dutch semiconductor equipment manufacturer ASML falling 6%.

盛宝银行英国策略师尼尔·威尔逊在周一发布的研究报告中指出，放缓前沿人工智能研发速度的潜在风险，直接对该技术领域的整体投资逻辑构成冲击。他分析认为，一旦人工智能产业链相关企业协同推行实质性的研发降速并设立安全护栏，分析师将被迫重新评估这些企业的盈利预期与估值模型，从而引发资本市场对估值泡沫与增长可持续性的普遍担忧。

The broader market retreat reflected growing unease over technology valuations and future earnings growth. Neil Wilson, UK strategist at Saxo Bank, wrote in a research note on Monday that a slowdown in frontier AI development threatens the sector's core investment thesis. He noted that equity analysts would be forced to re-evaluate their corporate earnings projections and valuation benchmarks if artificial intelligence companies coordinate a material slowing in development to erect regulatory and safety guardrails.

在社交平台X发布的长文中，Anthropic首席执行官达里奥·阿莫代伊提出了一项旨在控制前沿人工智能研发节奏的三步走方案。阿莫代伊指出，自2026年夏季以来，人工智能依靠构建下一代系统的递归自我改进（RSI）能力实现了极快推进，若缺乏有效约束，系统能力可能超出人类的理解与控制范围。为防范网络攻击、生物恐怖主义和失控风险，他建议各前沿实验室向独立外部评估员开放相当于内部员工级别的访问权限，确立通用的安全基准标准，并推行全球范围的治理协调。

In a detailed essay published on social media platform X, Anthropic Chief Executive Dario Amodei outlined a three-step proposal designed to pace the development of frontier artificial intelligence. Amodei stated that the technology has advanced drastically faster since roughly the summer, driven by recursive self-improvement where models build next-generation systems. He warned that unchecked progress could outrun human ability to understand and control these models, raising dangers of cyberattacks, bioterrorism, and systemic loss of control. To mitigate these risks, he called on frontier labs to grant employee-like access to external evaluators, set common safety standards, and pursue international coordination.

OpenAI首席执行官山姆·奥特曼和xAI负责人埃隆·马斯克对阿莫代伊的降速倡议表示公开支持。奥特曼在后续声明中承诺OpenAI也将引入具备员工级权限的独立评估机制，并呼吁设立统一的联邦前沿AI安全框架。奥特曼强调，所谓研发调速并非要求行业完全停滞，而是确保模型对齐与监控能力不落后于技术性能扩展，且任何竞争压力都不应成为忽视安全风险并采取冒进行为的借口。

OpenAI Chief Executive Sam Altman and xAI leader Elon Musk publicly backed Amodei's call to pace capabilities. Altman confirmed that OpenAI would commit to providing independent evaluators with employee-like access and welcomed a federal framework to establish consistent safety requirements across the sector. Altman clarified that pacing development does not mean stopping progress, but rather ensuring that safety alignment and monitoring mechanisms keep up with model capabilities, stating that national competitive pressures should not justify recklessness.

人工智能行业内部对技术失控的担忧在上周迅速激化。Anthropic研究人员雅各布·考克森宣布离职并公开发声，指责行业未能严肃对待潜在风险，并警告从事研发的核心人员真切认为该技术可能在十年内带来毁灭性后果。考克森的离职声明在业内引发连锁反应，Anthropic以及竞争对手OpenAI的多名员工随后相继就灾难性风险公开表达忧虑，推动关于前沿模型安全风险的讨论进一步升温。

Internal unease across the artificial intelligence sector intensified following the high-profile departure of Anthropic researcher Jacob Coxon. Resigning over what he described as the industry's failure to take safety risks seriously enough, Coxon stated that those building frontier AI earnestly believe the technology could kill humanity by the end of the decade. His departure triggered broader warnings from peers, prompting other researchers at Anthropic and rival OpenAI to speak out publicly regarding the threat of catastrophic outcomes.

面对加剧的安全争议，前沿AI企业的商业化与资本化进程出现调整。山姆·奥特曼在《财富》杂志周六刊登的专访中明确排除了OpenAI在2026年实施首次公开募股的可能性。奥特曼表示，综合考虑当前安全领域出现的问题，现阶段推进上市并非明智之举，公司并未感受到必须在2026年启动IPO的外部压力。这一表态直接推迟了市场原本对该公司2026年完成历史性上市的预期。

The heightened safety concerns have begun reshaping the commercial timelines for leading artificial intelligence firms. In an interview published Saturday by Fortune magazine, OpenAI Chief Executive Sam Altman ruled out a widely anticipated 2026 initial public offering for the company. Altman stated that proceeding with an IPO right now would be ill-advised given ongoing developments in safety, emphasizing that the startup feels no external pressure to go public in 2026 despite previous market expectations for a landmark market debut.

美国总统特朗普在爱尔兰的高尔夫球场接受采访时明确拒绝了企业高管关于放缓研发的提议。特朗普淡化了人工智能的技术风险，强调美国在技术实力上处于全球最先进水平，必须在与中国的竞争中维持领先优势。他表示，人工智能领域的胜者将掌控全局，并认为行业内部存在过多提出不必要担忧的消极力量，所讨论的极端风险并不会成为现实。

Speaking to reporters at his golf course in Ireland, U.S. President Donald Trump rejected appeals by technology executives to slow down artificial intelligence development. Trump downplayed the safety risks raised by industry leaders, asserting that the United States must protect its technological edge over China because whoever wins the AI race wins overall. He added that the administration could implement specific guardrails, but argued that negative voices inside the debate are raising issues and scenarios that will not materialize.

达里奥·阿莫代伊在文章中曾指出，放缓幅度不能超过美国对威权政权特别是非受限项目的领先优势，否则将引发国家安全风险。对此，中国外交部发言人郭嘉昆周一在回应美方高管言论时表示，散布恐慌、搞对抗和竞争只会破坏全球人工智能治理进程。同日，中国国家安全部部长陈一新发表署名文章，称人工智能已成为大国战略博弈的新赛场，呼吁加快建立人工智能安全风险防控体系，推进该技术健康有序发展。

The geopolitical friction mirrors warnings from Dario Amodei, who noted in his essay that pacing must not exceed the lead American companies hold over foreign adversaries to prevent unpaced foreign projects from pulling ahead. Responding to the U.S. executives' comments on Monday, Chinese Foreign Ministry spokesperson Guo Jiakun stated that fear-mongering, confrontation, and competition will merely disrupt the global AI governance process. Meanwhile, China's Minister of State Security Chen Yixin published an article calling for an accelerated AI security risk prevention system, describing the sector as a primary battleground for major-power rivalry that requires healthy and orderly development.