IT Home, October 16: NVIDIA CEO Jensen Huang stated that customer orders for AI chips continue to exceed the company’s production capacity; despite recent sharp sell-offs in tech stocks, which have raised questions about how buyers are financing the industry’s massive expansion, demand remains robust.
In a recent BG2 podcast interview aired on Thursday, Jensen Huang revealed that new orders are “continuing to pour in almost every day, every week.”
We have been continuously increasing our supply, but the volume of undelivered orders continues to rise.
Major cloud service providers, emerging AI cloud computing operators, and hardware manufacturers have all placed massive orders. Jensen Huang stated that customer demand forecasts are generally accelerated by at least one year, with the largest clients planning as far as two years ahead.
However, the actual allocation NVIDIA currently provides to these customers remains far below their actual demand. Huang noted that power supply constraints, unfinished data centers, and shortages of manufacturing components are also key bottlenecks limiting the company’s ability to rapidly expand supply.
IT Home noticed that, in response to concerns that companies like OpenAI and Anthropic must generate enormous revenues to cover trillions of dollars in AI infrastructure investments, Huang Renxun countered: “People have a misunderstanding of the current state of the industry.”
According to Jensen Huang, cutting-edge AI labs account for only about 25% to 30% of NVIDIA’s overall computing demand, with the remainder widely coming from upgrades to traditional computing architectures, startup teams, and companies developing proprietary commercial applications.
Huang emphasized: "They don't need to reach a revenue scale of trillions of dollars to prove their success."