Nvidia CEO Jensen Huang stated that demand for AI chips consistently outpaces the company's supply capacity. Despite market concerns regarding the financing of large-scale industry expansion, new orders continue to arrive daily.
Major cloud service providers, emerging AI computing operators, and hardware manufacturers have placed significant orders. Huang noted that client demand forecasts are typically planned at least one year in advance, with the largest customers projecting needs up to two years ahead.
Nvidia currently allocates fewer chips to customers than they request. Huang identified several bottlenecks limiting supply growth, including power shortages, incomplete data center infrastructure, and a scarcity of manufacturing components.
Addressing concerns about the financial viability of AI infrastructure, Huang clarified that frontier AI labs account for only 25% to 30% of Nvidia's total compute demand. The remaining demand comes from traditional computing upgrades, startups, and proprietary business applications.
