Nvidia nears record high as AI chip depreciation debate with Burry heats up

Nvidia shares climbed toward a record high on Friday, October 2, as the world's most valuable company publicly rebutted one of the loudest criticisms of the AI boom. The stock gained 2.4% to $236.28 in early trading, putting it within reach of the $235.74 record close set back in May.

The rally came with a message. Michael Burry, the investor made famous by The Big Short, has argued that Big Tech customers are stretching the depreciation schedules on Nvidia's AI chips — spreading hardware costs over longer lifetimes than the equipment actually lasts, which would mean they are understating their real annual expenses.

Nvidia answered directly. In a company blog post published Thursday, Shruti Koparkar — a senior member of Nvidia's product marketing team — wrote that server operators have been extending depreciation schedules for years, and pointed out that large buyers including Meta and Alphabet's Google now write down AI servers over six years rather than the previous four. Nvidia's counter to Burry's claim that rising rental prices merely reflect computing scarcity was that its CUDA software platform runs across chip generations, so existing hardware stays useful when newer chip designs launch.

Why it matters

The depreciation question has moved from an accounting footnote to a market-moving debate because real money now rides on the answer. In August, Nvidia lined up Wall Street partners to deploy $500 billion in financing for the AI sector, and those deals are expected to involve loans backed partly by Nvidia hardware itself. Lenders need to believe the chips hold their value. Meanwhile Nvidia's own data, repeated in a September investor briefing, shows that even half-decade-old Nvidia silicon has seen rental values rise — evidence the company is using to say the critics are wrong. Burry, for his part, has positions that would benefit if Nvidia shares fell, a reminder that both sides have skin in the game.