For most of the past couple of years the best argument that American stocks were not in a bubble was that things simply didn’t feel manic(狂热的，躁狂的) enough. Yes, prices had soared, and yes, valuations had risen so high that shareholders’ expected future returns were looking increasingly disappointing. Yes, the odds of unexpected returns were receding as well, with more or less everyone already convinced that corporate earnings would keep rocketing up(飞速上升) while artificial intelligence revolutionised the economy.
在过去几年的大部分时间里，认为美国股市并非泡沫的最佳论据是，事情尚未显得足够狂热。没错，股价飙升了，估值高企以至于股东预期的未来回报正变得越来越令人失望。没错，意外回报的可能性也在消退，因为几乎所有人都已经确信，在企业盈利将随着人工智能革命经济而继续飙升。
But a proper bubble needs more. It is not enough for investors merely to book tomorrow’s profits today. Share prices really come unstuck(脱离，松开) from reality only after a mania takes hold. Then comes the grim doggedness(阴沉的固执) of revellers who know the next morning’s hangover will be dreadful and are determined to put it off for as long as possible. Then, when it no longer can be, comes the crash.
但一个真正的泡沫需要的远不止这些。投资者仅仅将明天的利润计入今天的账本是不够的。股价只有在狂热占据主导之后才会真正脱离现实。随之而来的是狂欢者那种阴沉的顽固——他们知道第二天早晨的宿醉将无比难受，却决心尽可能推迟它。然后，当再也无法推迟时，崩盘降临。
Has the manic phase now arrived? Imagine telling a time-traveller from a decade ago that on June 11th SpaceX sold shares which valued it at $1.8trn, over 90 times its annual revenue. Mad? No—a fantastic deal for buyers, since the rocketry(火箭技术，火箭学) firm’s bankers reportedly told them it would increase its AI arm’s sales by a factor of 100 by 2030, to $322bn. Why the focus on revenue rather than the conventional way of valuing a company, relative to profits? Because SpaceX doesn’t make any; last year it lost $5bn. Four trading days after its listing its share price had risen by over 40%, valuing it at $2.5trn. In the meantime it agreed to pay $60bn in shares for Cursor, an AI-coding firm.
那么狂热阶段到了吗？想象一下，告诉一个来自十年前的时间旅行者，6月11日SpaceX出售的股票将其估值定为1.8万亿美元，是其年收入的90多倍。疯了吗？不——对买家来说这是一笔极好的交易，因为这家火箭公司的银行家据称告诉他们，到2030年其人工智能部门的销售额将增长100倍，达到3220亿美元。为什么关注收入，而不是传统的以利润为基础的公司估值方式？因为SpaceX根本不赚钱；去年它亏损了50亿美元。上市后四个交易日，其股价上涨超过40%，估值达到2.5万亿美元。与此同时，它同意以600亿美元的股票收购AI编程公司Cursor。