What would you like to know?

Answer

capital spending

A large share of Nvidia's data-centre revenue comes from a small number of very large customers building out AI infrastructure at a scale few other companies could attempt. That concentration is a double-edged feature: it has driven extraordinary growth, but it also means the company's near-term prospects are tied unusually closely to the capital budgets of a handful of firms, several of which face their own scrutiny over how quickly AI spending is translating into revenue.

• Electrical power availability for new data centres has become a tighter constraint than chip supply in several regions.
• Construction and cooling infrastructure lead times now often exceed the time it takes to manufacture the chips themselves.
• Competing chip designs, both from established rivals and from customers building their own silicon, are narrowing the performance gap in specific workloads.
• Export restrictions on advanced chips to certain markets remain a source of policy risk that could affect addressable revenue.