OPENAI BUILDS ITS OWN CHIP: JALAPEÑO GOES FROM ZERO TO TAPE-OUT IN NINE MONTHS OpenAI has been paying Nvidia billions for years. That arrangement just got a lot less permanent. On June 24, OpenAI and Broadcom unveiled Jalapeño — OpenAI’s first custom AI inference chip and the first tangible product of a partnership the two companies announced in October 2025. The name sounds casual but the ambition behind it is not. The chip was designed from concept to manufacturing tape-out in nine months, which OpenAI and Broadcom claim is the fastest development cycle ever achieved in high-performance advanced semiconductors. Jalapeño is built for inference, the part of the AI pipeline where trained models actually answer your questions. It is not a training chip. That distinction matters because inference is where the money flows at scale. Every time you send a message to ChatGPT, a chip like this has to process it. If OpenAI can run those workloads on its own hardware instead of Nvidia GPUs, the cost savings are enormous and the strategic leverage shifts dramatically. OpenAI used its own AI models to accelerate parts of the design process. The chip targets initial deployment by end of 2026. Early testing shows substantially better performance per watt than current alternatives. This is not a research project. It is a plan to own the stack. Keywords: OpenAI Jalapeño chip, Broadcom AI chip, custom inference processor, OpenAI Nvidia alternative