OpenAI, Broadcom Unveil Jalapeño AI Chip
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI and Broadcom unveiled Jalapeño, an LLM-optimized inference chip that moved from design to manufacturing tape-out in nine months, with early testing showing performance per watt "substantially better than current state-of-the-art" per OpenAI
- OpenAI's own AI models were used to aid the chip's design and development cycle, a detail OpenAI highlights as accelerating the unusually short nine-month turnaround
- Tech Times reports Jalapeño targets 50% cheaper inference costs, framing the chip as a cost-down play for running large language models at scale
- Multiple outlets frame the unveiling as a direct challenge to Nvidia's AI chip dominance — The Next Web headline says "The target is Nvidia," while Blockonomi bills it as a bid to "challenge Nvidia's dominance" and Axios pits it against the incumbent
- TechRadar characterizes the broader strategy as an "Apple-like move to build the full stack," situating the chip inside OpenAI's reported effort to control the full AI infrastructure pipeline
- OpenAI President Greg Brockman appeared on CNBC calling the chip "a real performance improvement," the only on-the-record executive quote surfaced across the coverage
Why it matters: Jalapeño is OpenAI's first custom silicon, produced in a nine-month cycle that several outlets describe as remarkably fast, and multiple headlines explicitly position it against Nvidia. If the chip delivers the "substantially better" performance per watt OpenAI claims and the 50% cheaper inference Tech Times reports, OpenAI gains a credible path to reduce dependence on Nvidia while pursuing the "full stack" vertical integration TechRadar identifies as the strategic goal.
