OpenAI announced its first custom inference processor, dubbed Jalapeño, developed in partnership with Broadcom. The move places the research lab alongside Google, Apple and SpaceX, all of which are creating proprietary chips to lessen dependence on Nvidia’s AI silicon.
Custom silicon gives companies tighter control over hardware, allowing designs that match the exact demands of their models and promising performance gains similar to Apple’s boost after moving away from Intel. For OpenAI, the chip is less about a clean break and more about diversifying supply risk while extracting efficiency from its workloads.
Industry analysts see the trend as a potential head‑wind for Nvidia, whose market share could erode as more AI players internalize chip design. TechCrunch’s Equity podcast explores the broader impact, including upcoming deals and how the shift might reshape the AI hardware landscape.
