Broadcom Unveils Jalapeño Chip in Record 9 Months as Funds Shift to AMD, Micron
AVGO•Broadcom co-designed and manufactured OpenAI's custom Jalapeño AI chip in a record nine months, marking the fastest tape-out cycle in advanced semiconductors and showcasing performance-per-watt gains over current solutions. Despite this milestone and planned gigawatt-scale deployments with Microsoft and others in late 2026, institutional flow data shows Broadcom underperforming peers, with Chaikin Money Flow at -0.006 versus +0.076 for Micron and +0.169 for AMD.
1. Record-Setting Chip Development
Broadcom and OpenAI completed the Jalapeño AI inference chip from initial design to manufacturing tape-out in nine months, the fastest cycle yet reported in advanced semiconductor development. OpenAI's large-language model tools accelerated the design phase.
2. ASIC Business Model Validation
Broadcom's approach focuses on custom ASIC co-design and manufacturing fees rather than selling off-the-shelf GPUs, with CEO Hock Tan stating the model will drive frontier AI developers to commission tailored silicon. The Jalapeño launch underscores confidence in fee-based chip design services.
3. Performance and Deployment Plans
Early testing indicates the Jalapeño chip delivers substantially better performance per watt than existing solutions, positioning Broadcom to support gigawatt-scale AI deployments with Microsoft and other partners starting in late 2026. This scale could reinforce Broadcom's role in large AI infrastructure projects.
4. Institutional Flow Analysis
Despite the technical breakthrough, Broadcom's relative strength versus the SOXX benchmark stands at 53.6 and Chaikin Money Flow reads -0.006, signaling net distribution. In contrast, AMD shows +0.169 and Micron +0.076, indicating institutional funds are rotating toward memory and GPU-focused peers.
