Elon Musk signals SpaceX and Tesla may build their own AI chips, challenging NVIDIA's dominance in chip supply

Musk publicly stated that a major AI semiconductor company is not expanding production fast enough to meet demand — widely understood as criticism of NVIDIA, which supplies chips for AI training and inference (the computational step where an AI produces answers). This suggests SpaceX and Tesla may pursue an in-house chip design strategy to reduce dependency on external suppliers.

If SpaceX and Tesla develop their own chips, they could optimize silicon specifically for their AI workloads — Tesla's autonomous driving and energy systems, SpaceX's rocket optimization and launch scheduling — rather than buying generic chips designed for many use cases. This typically reduces cost-per-task and speeds up the time between when an AI system is trained and when it produces answers.

For business professionals: NVIDIA's ability to control chip availability and pricing has become a bottleneck for AI companies. If major players like Tesla and SpaceX move to custom silicon, it signals a shift away from buying off-the-shelf chips, which could reshape competition in semiconductor supply and make NVIDIA's dominance less automatic in future years. For Tesla investors specifically, this could lower long-term operating costs for autonomous vehicles and energy products.