Nvidia defended the economics of its artificial intelligence spending boom after “Big Short” investor Michael Burley questioned the lifecycle of the company’s advanced chips and whether buyers would see a profit. On an earnings call late Wednesday, Chief Financial Officer Colette Kress said NVIDIA’s hardware remains productive much longer than critics have claimed, thanks to efficiencies from the company’s CUDA software system. “The long lifespan of NVIDIA’s CUDA GPUs is a significant total cost of ownership advantage over accelerators,” Kress said, according to a transcript of the call from FactSet. “CUDA’s compatibility across our large installed base extends the lifespan of NVIDIA systems far beyond their original estimated useful life. Thanks to CUDA, the A100 GPUs we shipped six years ago are now running at full capacity with a vastly improved software stack,” said Kress. He said the platform preserves the economic lifespan of graphics processing units, meaning customers can get more long-term value even as new generations of chips bring significant efficiency gains. Nvidia’s argument addresses a growing concern on Wall Street that the rapidly advancing power and efficiency of Nvidia’s chips could erase the value of previous generations of chips before enterprise buyers can monetize their AI investments. “While software updates have been proven to extend the lifespan of older chips, NVIDIA has done a good job of signaling how accurate its depreciation schedules are for large customers,” Ben Reitzes, an analyst at Melius Research, said in a note to clients. Burley’s thesis Still, Burley and other critics are contradicted. Nvidia says its latest chips are superior in performance, efficiency and features, while also promising that older chips will continue to be economically valuable. One of those defenses has to give. Burley agreed with the idea in an email to CNBC. This widely followed investor recently made headlines by revealing significant bearish positions in Nvidia and Palantir. He brought up X again in the wake of Nvidia’s blockbuster quarterly report, reiterating his theory that newer GPUs use much less power, making older hardware uncompetitive. Therefore, companies may feel they need to invest in AI hardware to catch up, not because the investment is still profitable. “Just because something is used doesn’t mean it’s profitable,” Barry wrote of X. “If that’s the direction you’re going, you might have to do it, and it’s not fun.”
