Nvidia faces internal strain over GPU allocation

Nvidia is facing an internal shortage of GPUs for computing, according to Hsin-Chu Wu, head of the company’s automotive business. Wu said on the Decoder podcast that Nvidia teams need the chips to train and test AI models, but internal supply is limited, forcing groups to compete for access to compute resources.

Jensen Huang often steps in to mediate disputes over chip allocation, with decisions balancing near-term revenue against longer-term strategic priorities. A major focus is autonomous driving, which Nvidia sees as part of a future where everything that moves could become automated.

Huang has described these emerging opportunities as “zero trillion dollar businesses,” markets that do not yet exist but could eventually become much larger. Nvidia is directing chips, software, wafers, and simulation resources toward automotive and robotics work despite the data center business being far larger today.

The constraints extend beyond software teams to factory capacity, making chip production itself a source of internal pressure. Nvidia’s difficulty meeting its own demand underscores how tight GPU supply remains across the AI market.