How to invest in artificial intelligence

Private startups or public markets?

本文主要讨论应该如何投资人工智能行业：是投资初创公司，还是投资已经上市的大企业。本文认为，尽管过去历次技术革新很多是颠覆式的，很多新的公司迅速崛起壮大，但人工智能时代的革新却主要来自微软、NVIDIA等已上市的大公司。因此进行人工智能行业投资可能机会更多在二级市场，而非初创企业。

IT HAS BEEN a torrid 18 months for investors who bet on tech. SoftBank, a Japanese investment firm that epitomised the 2010s boom in venture capital for companies with rapid-growth ambitions, is still smartingfrom the shift to a world of higher interest rates and lower corporate valuations. But there is one area in which the firm, run by Son Masayoshi, its charismatic founder, wants to peekabove the parapet: investments in artificial intelligence (AI).

注：“Torrid” 本意是干燥，torrid zone热带，想象一下热带沙漠撒哈拉，torrid用来描述非常炎热和干燥的事物，引申为过度和强烈，不仅是热的，而且令人无法抗拒和承受。

smart在这里是熟词生义，smart除了表示聪明的以外，还可表示引起刺痛的。大概聪明的头脑容易说出令人刺痛的话吧。

"Charismatic"形容一个人具有无法抵御的吸引力和感染力，能够赢得人们的信任和支持，并达成自己想要实现的目标或愿望。一些政治领袖往往会有Charisma。

"Peek above the parapet"是指突出自己的意见或建议，在团队或集体中发表不同于常规的观点。这个习语来源于军事作战中设置在壕沟前面的矮墙，军人需趴在其后方探头窥视敌军情况，并使自己的身体暴露在敌军火力下。因此这种行为往往也会面临其他人的反对、批评和压力。"Parapet"是提供防护和额外支撑的构筑物，用于城堡、要塞等建筑物。

The advances of generative-AI platforms, such as ChatGPT, have left just about every investor discussing what to make of the incipient industry, and which firms it might upturn. Mr Son sees parallels with the early period of the internet. Generative AI could provide a new pipeline of initial public offerings—and the foundation for the next generation of mega-cap tech firms.

注：incipient早期的、初期的，指一个事物处于其初期的阶段，还没有完全成熟。例如"incipient technology"可以表示一项高新技术，将来可能有很大的发展前途。

Investors face two questions. The first is which frontier technologies will make market leaders a fortune. That is difficult enough. The second, establishing whether the value will accrue to upstarts backed by venture capital or existing technology giants, is at least as tricky. Nobody knows if it is better to have the best chatbot or plenty of customers—having a head start in a whizzy new tech is not the same as being able to make money from it. Indeed, lots of the value of revolutionary innovation is often captured by existing giants.

注："Whizzy"技术革新的；技术先进的，这个词最初在计算机技术和互联网领域中被广泛使用，用来形容某种程序或软件能够快速、有效地实现某个功能或者任务。

Alphabet, Amazon and Meta are three of the seven largest listed companies in America, worth a combined $3.4trn. They were founded between 1994 and 2004, emerging at a time when internet technology was new and people were spending an increasing amount of time online. Alibaba, a Chinese e-commerce giant, is another similar example (SoftBank’s early $20m stake in the company helped cement Mr Son’s reputation as an investor). Spotting tech trends, and developing the best platforms, generated a gargantuan amount of value for early and even not-so-early investors. Legacy firms struggled to jump on the bandwagon.

Will the story be the same this time around? The insights of Clayton Christensen, a management guru who pioneered a theory of innovation just as the internet giants were bursting onto the scene, can provide a useful guide. Christensen noted that smaller companies often gain traction in low-end markets and entirely new ones, which the largest incumbents eschew. The incumbents focus on deploying new technology for their existing customers and lines of business. They are not incompetent or ignorant of technological progress. Instead, they follow the seemingly correct path from a profit-maximising perspective—until it is too late and they are fatally undermined.

注：Clayton Christensen世界创新大师、哈佛商学院教授,曾两次获得“麦肯锡奖”,荣获2011年《哈佛商业评论》“当代50名最具影响力的商业思想家”，曾在 1990 年代中期提出了颠覆性创新理论。

eschew 避开。其中-sch其实就是shy害羞，害羞的人往往会避开很多东西。

Investors like Mr Son, excited about the future of startups that focus on AI, are implicitly presuming that a period of disruptive innovation is under way. But most of the recent excitement about generative-AI platforms has focused on their potential as a new technology to be deployed, not as companies which could open up brand new markets. In the case of other recent technological innovations, incumbents have won the day. Elad Gil, a venture capitalist, has noted that the value of previous advances in machine learning, the broader category of which generativeAI is a part, have accrued almost entirely to incumbents. The early internet startups have benefited, as have Microsoft and chip firms like Nvidia and Micron. The earlier stages of machine learning produced no listed firms that might be considered the Amazon or Google of their niche.

注：这是文章观点的核心支持证据：the value of previous advances in machine learning, the broader category of which generative AI is a part, have accrued almost entirely to incumbents。（机器学习领域革新几乎全部来自现有大型科技公司，而生成式人工智能也属于机器学习技术的一部分。）

Christensen’s insights make clear that revolutionary innovation does not always end up being revolutionary in business terms. Yet existing tech firms are now spending enormous sums on AI, suggesting they should be well-placed if the tech does turn out to revolutionise business. It is possible an investment in a broad index fund tracking existing listed tech firms will end up outperforming the equivalent investment in private, strictly AI-focused startups.

Theories about why innovation is sometimes disruptive and sometimes not are more often discussed by students of business and management than stockpickers. But the difference between the two possibilities is crucial in assessing whether the next generation of listed tech companies, with market capitalisations in the hundreds of billions of dollars, is to be found among private AI firms. As things stand, it looks more likely that the market value of the technology will end up as a new string to the bow of already giant tech firms.

注：As things stand是英语习语，通常用于表示政治、商业、社会的当前状态。例如：“As things stand, we are still waiting for the results of the investigation.”（就目前情况而言，我们还在等待调查结果。）

“a new string”是一种隐喻，指“新的方法或机会”。本文中是指人工智能技术可能会成为现有大型科技公司的机会和方向。

扫码进经济学人打卡答疑群