In the nearly three years since AI took center stage in Silicon Valley, the major players, with the exception of Nvidia, whose chips will likely still be in utilize after the collapse, have still not shown what their long-term AI business model will be. OpenAI, Anthropic, and the AI tech giants are burning through billions, inference costs haven’t come down (these companies still lose money on almost every user query), and the long-term viability of their enterprise programs is a massive question mark at best. Is a product that justifies hundreds of billions of investment a search engine replacement? A substitute for social media? Workplace automation? How will AI companies price energy and computation costs that remain sky-high? If copyright lawsuits don’t get in the way, will they have to license their training data and will they pass on these additional costs to consumers? AND latest MIT study made waves – and helped fuel the latest wave of bubble fears – with the revelation that 95 percent of companies that implemented generative AI saw no profit from the technology at all.
“Usually, uncertainty decreases over time,” Goldfarb says. People learn what works and what doesn’t. This did not happen with AI. “Over the last few months,” he says, “we’ve realized that there’s a rough edge, and some of the earliest claims about AI’s effectiveness were ambiguous or not as great as initially claimed.” Goldfarb believes that the market still underestimates the difficulties of integrating AI into organizations, and he is not alone. “If we underestimate this difficulty as a whole,” Goldfarb says, “then we are more likely to create a bubble.”
The closest historical equivalent to artificial intelligence may not be electric lighting, but radio. When RCA began broadcasting in 1919, it immediately became clear that it had powerful information technology. However, it was less clear how this would translate into business. “Would radio be loss-making marketing for department stores? A public service enabling Sunday sermons to be broadcast? An advertising-supported entertainment medium?” the authors write. “They were all possible. They were all subject to technological narratives.” As a result, radio became one of the largest bubbles in history, reaching its peak in 1929 before losing 97 percent of its value in a crash. This wasn’t a random sector; RCA, along with Ford Motor Company, was the most heavily traded stock on the market. It was as The Novel Yorker wrote. recently wrote“The Nvidia of its time.”
Fair game
Why Toyota valued at $273 billion while Tesla is worth $1.5 trillion investors – when Toyota sold more cars than Tesla last yearand brought three times more income? The answer has to do with Tesla’s status as a “pure” investment in electric (and to a lesser extent, autonomous) cars. In 2010, Elon Musk used all the invigorating uncertainty surrounding electric vehicles to tell an internal combustion engine-free story that was so enticing that investors were willing to bet enormously on a volatile startup over proven workhorses. A purely fun company is one whose fate is tied to the emergence of a specific innovation that entrepreneurs could tell more invigorating and fantastic stories about, and these are needed to inflate the bubble. They are a tool through which narratives are turned into material bets.
So far this year, according to Silicon Valley Bank58 percent of all VC investments went to AI companies. Retail investors don’t have many obvious pure-play investments – another criteria for inflating a bubble – but there are a few massive ones. Nvidia is at the top of the list because it has staked its future on building chips for artificial intelligence companies and has become… the first $4 trillion company in history during. According to Goldfarb and Kirsch, when a sector sees a lot of clear situations, it is more likely to overheat and develop a bubble. SoftBank plans to sink tens of billions of dollars into OpenAI, the purest artificial intelligence game, although it is not yet open to retail investment. (If and when this finally happens, analysts speculate OpenAI Could Become the First Trillion-Dollar IPO.) Investors also supported pure-play gaming companies such as Perplexity (currently valued at $20 billion) and CoreWeave (Market capitalization of $61 billion). In the case of artificial intelligence, these pure-play investments are particularly concerning as the largest companies become increasingly interconnected. Nvidia just announced a proposed $100 billion investment in OpenAI, which in turn relies on Nvidia chips. OpenAI relies on Microsoft’s computing power, the result of a $10 billion partnership, and Microsoft in turn needs OpenAI’s artificial intelligence models.