Michael Burry, the investor made famous in “The Big Short,” has warned about an artificial intelligence bubble. So has Dean Baker, an economist who identified the U.S. housing bubble ahead of the 2008 financial crisis. Even Jamie Dimon, the chief executive of JPMorgan Chase, has expressed caution.

因电影《大空头》(The Big Short)闻名的投资者迈克尔·伯里曾警告人工智能存在泡沫。曾指出2008年金融危机前美国地产泡沫的经济学家迪恩·贝克也发出过类似警告。甚至连摩根大通首席执行官杰米·戴蒙也表达了谨慎态度。

But in Silicon Valley, some people are less worried. In fact, they say, a bubble might be a good thing.

但在硅谷，有些人却没那么担心。事实上，他们认为泡沫可能是一件好事。

In particular, venture capital investors, the optimistic risk-takers who provide money for start-ups, view the frothiness as a necessary part of the innovation machine. The excitement created by a bubble can drive new breakthroughs, their thinking goes.

尤其是风险投资人——这些为初创企业提供资金的乐观主义冒险家——将这种泡沫化倾向视为创新机器中不可或缺的一部分。他们的逻辑是，泡沫所创造的热情能够推动新的突破。

“Bubbles are a part of technology and innovation,” said Tomasz Tunguz, an investor at the venture capital firm Theory Ventures. These frenzies are important for allowing crucial infrastructure to get built, even if they lead to some “capital destruction” along the way, he added.

“泡沫是科技和创新的一部分，”风险投资公司Theory Ventures投资人托马什·通古兹说道。他还说，即使这些狂热在此过程中会导致一些“资本毁灭”，但它们对于让关键基础设施得以建成至关重要。

It is a common refrain in the nation’s tech capital. How else can risky, cutting-edge companies attract the money, talent and excitement they need to build the future?

在美国的这个科技之都，这是个常见的观点。要不然，那些充满风险的前沿公司该如何吸引构建未来所需的资金、人才和关注度呢？

“If you were to rely on a purely organic, rational way of financing big technology platform shifts, they would probably never happen,” said Samir Kumar, an investor at Touring Capital, a venture capital firm. Once you get past the excessive exuberance, he added, “there are long-term benefits.”

“如果你要依赖一种纯粹有机、理性的方式来为大型科技平台的转变提供融资，那可能永远不会发生，”风险投资公司Touring Capital投资人萨米尔·库马尔表示。而且一旦你跨越了过度繁荣的阶段，“就会看到长期的好处，”他还说。

This counterintuitive thinking stands in contrast to rising concerns over a potential A.I. mania, with dealmaking that never seems to stop. This week alone, OpenAI and Nvidia were close to landing an elaborate $500 billion data center deal in Ohio. Amazon, Google, Meta and Microsoft said they had spent a combined $170 billion on capital expenditures in the second quarter, up 72 percent over last year.

这种反直觉的想法与人们对潜在人工智能狂热的日益担忧形成了鲜明对比，这一领域的交易活动似乎永不停歇。仅在本周，OpenAI和英伟达就即将敲定一项在俄亥俄州建设价值达5000亿美元的复杂数据中心协议。亚马逊、谷歌、Meta和微软表示，他们在第二季度的资本支出总计达到1700亿美元，同比大幅增长了72%。

Venture funding in the first half of the year also hit $413 billion, more than all of 2025, according to PitchBook, which tracks private investment. Young companies that have not released a product or earned a dollar of revenue are landing valuations that sound absurd, in one case reaching $32 billion. And there is so much money sloshing around Silicon Valley that nearby San Francisco has a mansion shortage.

追踪私募投资的机构PitchBook的数据指出，今年上半年的风险投资额达到了4130亿美元，超过了2025年全年的总和。一些尚未推出产品或赚到一分钱收入的新兴公司正斩获听起来极其荒谬的估值，其中一个甚至达到了320亿美元。硅谷的财富是如此横流满溢，以至于邻近的旧金山出现了豪宅短缺。

The bubble-is-good philosophy reveals how much the tech industry is still pressing the pedal to the metal, four years into the A.I. boom. Many investors said the funding behavior was justified because companies were using the technology to hit unfathomable levels of growth.

“泡沫有益”的哲学揭示出，人工智能狂热持续了四年之后，科技界依然在全力踩油门。许多投资者表示，这种融资行为是有道理的，因为各类公司正在利用这项技术来实现令人难以置信的增长水平。

OpenAI, valued at $730 billion, is generating $2 billion in revenue a month. Anthropic, valued at $900 billion, is producing nearly $4 billion a month. Google and Meta have boasted of nearly a billion people using their A.I. models. The demand for A.I., investors said, is seemingly unlimited.

估值为7300亿美元的OpenAI每月产生20亿美元的收入。估值为9000亿美元的Anthropic每月产生近40亿美元的收入。谷歌和Meta都宣称有近十亿人在使用他们的人工智能模型。投资者表示，对人工智能的需求似乎是无限的。

“V.C.s only really make money when there are big technology shifts to invest behind,” said Pratyush Buddiga, an investor at Susa Ventures.

“风险投资人只有在有重大技术变革可供投资时才真正能赚到钱，”Susa Ventures投资人普拉蒂尤什·布迪加说道。

(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The two companies have denied those claims.)

（《纽约时报》已向OpenAI和微软提起诉讼，指控其侵犯了人工智能系统相关新闻报道的版权。这两家公司均否认了这些指控。）

William Quinn, an associate professor at Queen’s University Belfast and a co-author of “Boom and Bust: A Global History of Financial Bubbles,” said past bubbles, and the easy money that came with them, had obvious positive elements: They drove innovation and long-term economic growth.

贝尔法斯特女王大学副教授、《繁荣与萧条：金融泡沫的全球史》(Boom and Bust: A Global History of Financial Bubbles)一书的合著者威廉·奎因表示，过去的泡沫以及随之而来的容易获取的资金具有明显的积极因素：它们推动了创新和长期经济增长。

“In Silicon Valley, there are things that everyone seems to believe that seem crazy to the rest of the world, and on a lot of those things, it seems like Silicon Valley was right and the rest of the world was wrong,” he said.

“有些事情在硅谷似乎所有人都深信不疑，而在世界其他地方看来却很疯狂；但在很多这类事情上，似乎硅谷是对的，而世界其他地方是错的，”他说。

Bubbles also tend to go on for a lot longer than people think they will, Mr. Quinn said. Pessimists who see irrational behavior and sit on the sidelines can miss years of gains before any correction or crash. By the time the comedown happens, no one is listening to their bubble warnings anymore.

奎因表示，泡沫持续的时间通常也比人们想象的要长得多。在发生任何回调或崩盘之前，看到非理性行为就选择旁观的悲观主义者可能会错失多年的收益。等到轰然倒塌真正发生时，已经再也没有人在听他们的泡沫警告了。

“There’s always the joke that they’ve predicted 10 of the last three crashes,” he said.

“大家总是开玩笑说，他们预测到了过去三次崩盘中的十次，”他说。

The Dot-Com Lesson

互联网泡沫的教训

Tech industry veterans still carry scars of the 2000 dot-com crash and its painful aftermath. For several years after, it was nearly impossible for start-ups to raise money.

科技行业的资深人士依然带着2000年互联网泡沫破裂及其痛苦后果留下的伤疤。那之后的好几年里，初创公司几乎不可能筹集到资金。

The handful of companies that powered through to the other side — Amazon, PayPal, eBay — went on to become behemoths. And over time, the fiber-optic network that exuberant investors overbuilt in that boom helped more people get online, fueling the growth of new tech companies like Facebook.

少数挺过难关的公司——亚马逊、PayPal、eBay——后来都成长为了巨头。随着时间的推移，狂热的投资者在那个繁荣时期过度建设的光纤网络帮助了更多人接入互联网，从而推动了像Facebook这样的新型科技公司的增长。

Investors who witnessed the dot-com crash remained wary of history’s repeating itself. In the early 2010s, they warned of a bubble among mobile apps with no revenue. Then came fears over “unicorn” start-ups that were valued at $1 billion or more, followed by nonsensical crypto projects and, eventually, the overlapping investment manias of 2021.

见证了互联网崩盘的投资者始终对历史重演保持警惕。在2010年代初期，他们警示过没有收入的移动应用存在泡沫。随后是对估值达到或超过10亿美元的“独角兽”初创公司的担忧，接着是荒谬不经的加密货币项目，并最终发展为2021年相互交织的投资狂热。

But none of those moments led to a major crash.

但这些节点都没有引发严重的崩溃。

Charles Hudson, an investor at Precursor Ventures, said the 2021 frenzy taught him that investors got little credit from sitting things out in bubbly times. Some investors made lots of bad, silly bets in 2021, but their backers did not punish them for the excesses. The lesson, he said, was to keep investing, even when things felt overheated.

Precursor Ventures的投资者查尔斯·哈德森表示，2021年的狂热给了他一个教训：在泡沫时期选择置身事外的投资者并不会因此获得什么赞赏。一些投资者在2021年做出了许多糟糕、愚蠢的下注，但他们的出资人并没有因为这些过激行为而惩罚他们。他说，这给人的启示是：即使在环境看起来过热时，也要继续投资。

The A.I. boom has been trickier to navigate, with obscene dollar amounts and the market moving at hyperspeed. Even so, Mr. Hudson said he knew he could not sit it out. He has focused on backing even younger companies and homed in on the “application layer,” or software start-ups that use A.I.

人工智能热潮的走向更加难以把握，涉及的金额巨大，市场发展速度也异常迅猛。即使如此，哈德森表示他知道自己不能置身事外。他将重点放在资助更为年轻的公司上，并聚焦于“应用层”，即使用人工智能的软件初创公司。

“It would be incredibly dangerous, and maybe irresponsible, to just say, ‘I’m not investing in A.I. at all because I think pockets of the business are a bubble,’” he said.

“如果仅仅因为认为这个行业的某些局部存在泡沫，就声称‘我完全不投资人工智能’，那将是极其危险的，甚至是不负责任的，”他说。

Sudheendra Chilappagari, an investor at Battery Ventures, was in school when the dot-com bubble happened. (At 33, he is older than some of the top A.I. start-up founders, who were not yet born.) He joked that he felt jealous of his older colleagues because they started investing before there was so much noise and chaos.

Battery Ventures投资人苏德辛德拉·奇拉帕加里在互联网泡沫发生时还在上学。（现年33岁的他比一些顶尖人工智能初创公司的创始人年龄还要大，他们当时甚至尚未出生。）他开玩笑说，他很羡慕资历更老的同事，因为他们在没有这么多噪音和混乱之前就已经开始投资了。

He said he was taking a lesson from their experience in the 1990s search engine wars. At the time, Battery Ventures invested in Infoseek, one of the earliest search engines — only to get crushed by Yahoo and others that then got crushed by Google.

他表示，自己正从他们90年代搜索引擎之战的经历中吸取教训。当时，Battery Ventures投资了最早的搜索引擎之一Infoseek——结果被雅虎等公司碾压，后者随后又被谷歌碾压。

“It’s not always the first mover that captures the enterprise value,” Mr. Chilappagari said. In a so-called supercycle like the A.I. boom, he expects continuous waves of possible investments.

“捕捉到企业价值的并不总是先行者，”奇拉帕加里说道。在像人工智能爆发这样的所谓“超级周期”中，他预计会出现一浪接一浪潜在的投资机会。

The investors acknowledge that many factors could set off a crash: geopolitical clashes, competitive threats from cheaper open-source A.I. models, societal backlashes against data centers or more safety problems like OpenAI’s rogue hacking of a partner company. Even disappointment over how long it takes investments to pay off could deflate the market.

投资者承认，许多因素都可能引发崩溃：地缘政治冲突、来自更便宜的开源人工智能模型的竞争威胁、社会对数据中心的反弹，或是更多类似于OpenAI对合作公司进行恶劣黑客攻击的安全问题。甚至是对投资多久能获得回报感到失望，都可能让市场泄气。

“You don’t want to get delusional,” Mr. Kumar of Touring Capital said.

“你不能让自己陷入妄想，”Touring Capital的库马尔说道。

Darian Shirazi, an investor at the venture capital firm Gradient Ventures, said his profession had become more nerve-racking because the typical strategy of hitting one home-run deal to drive profits was so much more extreme. Venture capital firms might previously have hoped for a few of their investments to sell for $1 billion or $2 billion. In the A.I. boom, that potential outcome has increased to $50 billion.

风险投资公司Gradient Ventures的投资者达里安·席拉齐表示，他的职业变得更加令人神经紧绷，因为这种通过一次“全垒打”式的交易来带来利润的典型策略如今已经极端得多。过去，风险投资公司可能希望他们的几笔投资能以10亿或20亿美元的价格出售。而在人工智能繁荣时期，这一潜在结果已暴涨至500亿美元。

“Your winner can be wildly outsized,” he said. “If it works, it works much better than anything we’ve ever seen in the history of venture capital.”

“你赌赢的那一把可以大到难以置信，”他说。“一旦成了，那就是远远超越我们在风险投资历史上见过的任何投资的大胜。”