Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia)
TL;DR
Nvidia still sets the pace in AI chips, but major customers no longer want total dependence on one supplier, one roadmap, and one pricing model. OpenAI is working with Broadcom on Jalapeño, a custom inference chip. Google, Apple, and SpaceX are part of the same pattern: purpose-built silicon for specific workloads instead of generic GPUs for everything. This is not a quick break from Nvidia. It is a hedge: more control, better long-term costs, and hardware tuned more tightly to models, data centers, or devices.
Nauti's Take
Nvidia’s moat will not disappear overnight. The software ecosystem, manufacturing access, and developer trust are extremely hard to copy.
But the new logic is obvious: companies spending billions on AI infrastructure do not want to stay mere customers forever. Custom chips are less a hardware hobby than a leverage machine.
Even if Jalapeño and similar chips are only moderately successful, they can still pressure pricing, roadmap priorities, and dependency risk.
Briefingshow
AI infrastructure is becoming a strategic control point, not just a data center purchasing decision. When OpenAI and others build their own chips, power shifts from standard hardware toward vertically integrated stacks. Nvidia is not losing the market overnight, but its biggest customers gain leverage.
Sources