Meta’s AI Gamble

How much financial stress can tech companies handle in their efforts to win the artificial intelligence race? OpenAI CEO Sam Altman said today on CNBC he thinks the ChatGPT creator should continue “running at a loss” so it can continue to invest in computing capacity to keep advancing its AI models. In one way, he has to take that stance: He’s competing with rich companies, like Google and Meta Platforms, that have lucrative businesses to fund their development. But as AI-related capital expenditures rise, there’s surely a limit to what even those deep-pocketed companies can spend. Take Meta, whose cash balance dropped $30 billion—or 40%—in the first half of this year, as its spending on AI skyrocketed.

Meta is, arguably, making the biggest gamble of any of the big tech companies. Apart from the fact that it’s a smaller business than either Google or Microsoft—measured by revenue—it doesn’t have a cloud operation that sells AI-powered services to companies. Both Google and Microsoft are building data centers to serve their AI cloud customers. Meta is investing nearly as much on capex, including for new data centers for AI—but for what purpose? Zuckerberg has talked about delivering “personal superintelligence to everyone,” whatever that means. Sure, AI will underwrite improvements in Meta’s advertising business, but it’s highly debatable whether the ad growth that results will generate enough of a return to justify the massive investment.