From Jalapeño to SpaceX: Why the AI elite is abandoning Nvidia
openai is developing 'jalapeño', a custom inference chip with broadcom, joining spacex and apple in reducing reliance on nvidia. this shift signals a major industry pivot to vertical integration to secure compute supply and lower inference costs.
openai is working with broadcom on 'jalapeño', a custom inference chip designed to reduce its reliance on nvidia's gpus. this move places openai alongside google, apple, and spacex in a growing trend of vertical integration where companies build hardware to fit their specific software needs. by developing proprietary chips, these firms aim to insulate themselves from nvidia's supply constraints and high operational expenses. this shift pressures nvidia to innovate faster even as its customer base turns into a pool of competitors.