AI总结
- 核心结论：当企业群体性选择通过发行股票进行融资时，往往是市场估值过高、泡沫显现的强有力警示信号。
- 结构逻辑：
- 背景：SpaceX等公司的大规模股票发行潮引起关注。
- 分析：比较债务与股权融资的决策动机；引用“市场先生”理论解释发行股票本质上是企业的“卖出”决策。
- 对比：通过对比互联网泡沫、2020-21年热潮与金融危机前夕的融资结构，论证股权融资比例是衡量市场过热的有效指标。
- 结论：当前AI领域的资金投入可能存在过度炒作的“烧钱”风险。
- 关键信息：
- 行业板块：AI领域、半导体/芯片股。
- 热点趋势：企业融资潮、二级发行与IPO活跃、AI估值膨胀。
- 关键概念：加权平均资本成本(WACC)、“烧钱率”(Burn rate)、“市场先生”(Mr. Market)。
- 核心问题与回答：
- 问题：企业融资决策如何反映市场估值状况？
- 回答：当公司倾向于发行新股而非借债时，说明公司认为自身股价被高估，这是比传统市盈率更准确的市场泡沫指标。
正文
A SpaceX logo and an image of a rocket launch on a digital billboard in Times Square.
SpaceX launched its IPO earlier this month. Angela Weiss/AFP/Getty Images
SpaceX 于本月早些时候启动了首次公开募股（IPO）。 安吉拉·韦斯/法新社/盖蒂图片社
What should you do if investors bid up your stock on the basis that you will be a winner in artificial intelligence, but your product isn’t popular? Elon Musk has the answer: Use your expensive stock to buy another AI business.
如果投资者以你会成为人工智能领域的赢家为由出价，但你的产品不受欢迎，你应该怎么办？ 埃隆·马斯克 有答案：用你昂贵的股票买下另一家 AI 企业。
SpaceX’s SPCX -13.47%decrease; red down pointing triangle $60 billion all-stock purchase of Cursor, a programming assistant that popularized “vibe-coding,” makes Musk a player in corporate AI in a way that his Grok chatbot hasn’t.
It is also part of a rush of stock issuance that should raise serious red flags even among those who dismiss elevated valuations.
这也是股票发行热潮的一部分，即使是那些轻视高估值的人也应引起严重警示。
Companies always have a choice of how to finance capital spending and takeovers. They can raise debt or issue stock (or a mix). If interest rates are low, debt is cheap for companies. Equally, if stock valuations are high, it is “cheap” for a company to issue more. Only firms in desperate need of cash sell more shares when their valuations are low, because it dilutes existing shareholders and trashes the share price.
公司始终可以选择如何为资本支出和收购融资。他们可以筹集债务或发行股票（或混合）。如果利率低，企业借债就便宜。同样，如果股票估值较高，公司增加发行是“便宜”的。只有急需现金的公司才会在估值低时卖出更多股票，因为这会稀释现有股东并压低股价。
Legendary investor Benjamin Graham thought of “Mr. Market” as a manic-depressive, offering daily prices to investors—sometimes far too high (you should sell) and sometimes far too low (you should buy). Mr. Market also offers companies the opportunity to buy from or sell to us investors—and when they issue stock, they are choosing to sell.
传奇投资者本杰明·格雷厄姆认为“市场先生”就像躁郁症患者，他向投资者提供每日价格——有时过高（你应该卖出），有时又太低（你应该买入）。市场先生还为公司提供了向美国投资者买卖的机会——当他们发行股票时，就是选择出售。
In any individual case selling stock can be perfectly reasonable. Balancing the amount of debt and equity a company uses helps offset the perceived riskiness of an investment or acquisition. Equity is good for risky endeavors because it doesn’t have to be paid back like debt. The mix of debt and equity feeds into the company’s weighted average cost of capital, a measure that helps determine whether a particular investment is worth it.
在任何情况下，出售股票都是完全合理的。平衡公司使用的债务和股权，有助于抵消投资或收购的风险感知。股权适合高风险项目，因为它不必像偿还债务那样被偿还。债务和股本的比例反映在公司的加权平均资本成本中，这一指标有助于判断某项投资是否值得投资。
But step back, and when companies as a group turn into sellers, it’s a reasonable sign that stocks are very overpriced. Indiscriminate demand from investors incentivizes companies to step up to supply them, both with secondary offerings, like Alphabet’s GOOGL -5.52%decrease; red down pointing triangle recent record-breaking issue, or IPOs.
Examples from recent history stand out.
近代历史中的例子尤为突出。
In the dot-com bubble and postpandemic SPAC mania there were waves of IPOs and secondary issues as stock valuations rose. There was also a boom in U.S. mergers and acquisitions, only just surpassed in the past four quarters in dollar value. Companies financed those deals with equity—two-thirds with new stock in the dot-com M&A and 45% in 2020-21, the two highest over a rolling four-quarter period in LSEG data dating back to 1990.
在互联网泡沫和疫情后 SPAC 热潮中，随着股票估值上升，出现了大量首次公开募股和次级发行。美国的并购也一波繁荣，过去四个季度的美元价值仅略高于此。这些交易的资金来源是三分之二的，其中三分之二是互联网并购中新股票，2020-21 财年为 45%，这是 LSEG 自 1990 年以来滚动四个季度的最高两份。
Big deals are flourishing again: giant IPOs blasted off with SpaceX, and companies are choosing to issue stock—using shares to finance almost half the cost in the current quarter, according to LSEG. (Over the rolling four-quarter period, it is still only a third.) Investor demand for stock is, in part, being satisfied by the creation of new stock.
大型交易再次蓬勃发展：SpaceX 启动了大型 IPO，公司也选择发行股票——根据 LSEG 的数据，他们用股票融资了本季度近一半的成本。（在滚动的四季度周期内，这一数字仍仅为三分之一。）投资者对股票的需求部分是通过新股票的创造来满足的。
The flip side is, when debt is cheap, it can cause even more serious problems. Pre-financial crisis deal value peaked at $910 billion over the year to the end of September 2007, just before shareholders realized the economy was built on sand. Only 19% of that was financed with stock, because debt—helped by endless structured products—was far too cheap. The bubble was in the debt markets, not equities, although the implosion of the banks and deep recession still hit stocks exceptionally hard.
另一方面，当债务低廉时，可能会引发更严重的问题。金融危机前的交易价值在截至2007年9月底的一年内达到9100亿美元的峰值，就在股东意识到经济是建立在沙地上之前。其中只有19%是用股票融资的，因为债务——再加上无尽的结构性产品——实在太便宜了。泡沫主要集中在债券市场，而非股票，尽管银行的崩溃和深度衰退对股市打击尤为严重。
Corporate financing decisions are in some ways better at measuring how expensive the market is than standard valuation gauges, which compare price to some measure of corporate fundamentals such as past or forecast profits (soaring), book value, free cash flow (plunging) or sales.
企业融资决策在某些方面比标准估值指标更能衡量市场成本，后者将价格与企业基本面指标（如过去或预测利润（飙升）、账面价值、自由现金流（暴跌）或销售额进行比较。
These can be misleading, often disagree, and mostly have fairly short histories. Worse, the baseline for what counts as expensive can change over time as the structure of companies or accounting standards shift. In the case of book value, it now needs so many adjustments as to be virtually useless.
这些书 可能具有误导 性，常常存在分歧，而且历史大多较短。更糟的是，随着公司结构或会计准则的变化，什么算是昂贵的基线可能会随时间变化。就书面价值而言，现在需要进行大量调整，几乎毫无用处。
One example: The S&P 500’s ratio of price to forecast earnings currently stands just below 20, well down from the 23 times reached both in 2020 and last year, and the record 24.5 times in the dot-com bubble. It’s still very high, but it has fallen because Wall Street expects earnings to boom—particularly in highly cyclical chip stocks where record near-term profit margins are unlikely to be a good guide to the distant future.
举个例子：标普500指数的市盈率目前略低于20，远低于2020年和去年的23倍，以及互联网泡沫期间创纪录的24.5倍。虽然依然很高，但之所以下跌，是因为华尔街预期盈利将大幅增长——尤其是在高度周期性的芯片类股票中，创纪录的短期利润率不太可能成为遥远未来的良好参考。
There are drawbacks to looking at how much companies choose to issue debt or equity too. The decision on whether to raise money through stock or bond issuance tells us about how companies see the relative valuation of stocks vs. bonds more than the absolute valuation.
查看公司发行债务或股权的比例也存在缺点。决定通过发行股票还是债券融资，反映了公司如何更看重股票与债券的 相对 估值，而非绝对估值。
When it becomes concerning, at least in my view, is when there’s a flood of fundraising, debt or equity or both. If lots of companies are raising cash to chase the same opportunity—in this case AI—there are three possibilities.
至少在我看来，当筹款潮涌入、债务或股权，甚至两者兼有时，才会令人担忧。如果许多公司都在筹集资金追逐同一个机会——在这里指的是人工智能——有三种可能。
The bull case is that the opportunity is so huge it can absorb all the cash and still deliver fat profits. The bear case is that the opportunity is real but spending so much will destroy value as competition erodes margins. The deeply depressing third possibility is that companies are raising and spending so much merely because shareholders are cheering them on, and the AI claims are just wildly overhyped.
牛市的观点是，机会如此巨大，可以吸收所有现金，同时仍能带来丰厚利润。熊市的观点是，机会确实存在，但过度消费会摧毁价值，因为竞争会侵蚀利润率。令人沮丧的第三种可能是，公司之所以大量融资和投入，仅仅是因为股东的欢呼，而人工智能的说法被过度炒作了。
The danger is it turns out like the dot-coms. Then, like now, companies competed to spend as much as they could as quickly as possible and the “burn rate” was seen as a positive—until it turned out all the shareholder money had gone up in smoke.
危险在于，这会变成互联网公司。那时和现在一样，公司争相尽快花尽可能多，“烧钱率”被视为积极因素——直到事实证明所有股东资金都化为乌有。
Appeared in the June 22, 2026, print edition as ‘Flood of AI Money Is a Warning Sign’.
发表于2026年6月22日印刷版，标题为《人工智能资金洪流是警示信号》。