Venture capitalists have a professional obligation to be optimistic, but David Cahn, a partner at the storied US fund Sequoia Capital, has spotted a $US500 billion hole in the year’s biggest economic narrative: AI.
Mark Zuckerberg said the amount of computing power required to train Meta’s next AI model will “likely be almost 10 times more” than its last one., has been seeing wild swings in its share price, down more than 20 per cent in the latter half of last month and up almost 13 per cent on Wednesday. Microsoft, meanwhile,
Cahn thinks it will be, at least for consumers, but that investors who believe the hype will get burnt. He’s thinking of, but the same points apply to listed firms. A lack of pricing power as AI becomes commoditised, the difficulty of picking winners and depreciation of AI chips will all hurt, Cahn warns. Investors, he says, should not get caught in a delusion.
Meta’s chief financial officer, Susan Li, said the company had “strong returns” from AI but did not provide numbers to back up that claim. And costs will grow, as it spends on data centres and processing grunt. It raised projections for capital expenditure over the full year to a range of $US37 billion to $US40 billion, which was $US2 billion higher at the lower end.