ESTIMATEDDate confidence: ESTIMATEDA target, not a fixed dateHow firmly the source words the date: confirmed, scheduled, estimated, delayed or unverified.Show all ESTIMATED pins65% CONFIDENCEOverall confidence: 65%How well the pin's source and references back up its dates.Weighted average of how firmly 5 references, the source included, support the pin's start and end times; a reference counts half as much for every 180 days older than the newestShow all pins from 50% to 74% confidence
CNBC[1] (24 June 2026) reports the companies "are aiming for initial deployment of the Jalapeño chips by the end of 2026", and the release says "designed for initial deployment by the end of 2026", so the pin takes that period's last day; Tan expects "small prototype development" in late 2026, and no deployment had been reported by 23 September 2026.- Start
- Dec 31, 202650% CONFIDENCEsource
OpenAI's Broadcom-Built Jalapeño Chips Reach First Deployment
The first deployment of Jalapeño, the inference accelerator OpenAI designed and Broadcom builds under their 10 GW deal, due by the end of 2026 in data centers with Microsoft and other partners, ahead of a planned 1.3 GW in 2027.
- OpenAI and Broadcom are aiming for initial deployment of the Jalapeño chips by the end of 2026, "expanding in the years ahead"; the chips are made by Broadcom and used by OpenAI for inference[1]
- Hock Tan told CNBC there would be "small prototype development" in late 2026, then "We will start seeing it really ramp up in '27 and really going full tilt in first half '28"[1]
- The release calls Jalapeño the first step in a multi-generation compute platform designed for initial deployment by the end of 2026; Tan said it enables gigawatt-scale data centers with Microsoft and other partners beginning in 2026, with Celestica supplying boards, racks and system integration[4]
- Tom's Hardware reported the chip slated for deployment beginning in late 2026[5]
- Broadcom said it shipped Jalapeño in its fiscal third quarter (ended 2 August 2026), that shipments to OpenAI continue in the fourth, and that Jalapeño is on track for a planned 1.3 GW deployment in 2027[2]
- The deployable system scales to 128 accelerators per rack and 2,048 in a 16-rack pod on Broadcom Tomahawk 6 Ethernet, each chip with 216 GB of HBM4 and up to 13.4 MXFP4 PFLOPS at 700W[3]