OpenAI and Broadcom jointly release LLM inference chip Jalapeño
Original title:OpenAI and Broadcom unveil LLM-optimized inference chip
AI summary
OpenAI and Broadcom released Jalapeño, their first self-developed inference accelerator, designed from scratch for current and future LLMs. Early tests show its performance-per-watt significantly outperforms existing SOTA. Engineering samples have run workloads such as GPT-5.3-Codex-Spark in the lab at target frequency and power. The chip took only 9 months from design to tape-out, with OpenAI models accelerating part of the process. OpenAI plans to deploy gigawatt-scale data centers with partners including Microsoft starting in 2026, rolling out multiple generations of computing platforms.
Key Takeaways
- OpenAI and Broadcom released Jalapeño,their first self-developed inference accelerator,designed…
- Early tests show its performance-per-watt significantly outperforms existing SOTA.
- Engineering samples have run workloads such as GPT-5.
- 3-Codex-Spark in the lab at target frequency and power.
- The chip took only 9 months from design to tape-out,with OpenAI models accelerating part of the…
Quick Facts
- Published
- Updated
- Source
- OpenAI:官网动态(RSS · 排除企业/客户案例)
- Category
- Products & Apps
- AI score
- 61
OpenAI 首次亲自设计芯片,和 Broadcom 联手推出专为 LLM 推理优化的 Jalapeño,从设计到流片仅 9 个月。虽然还只是早期测试,但性能功耗比大幅领先,一旦大规模部署,推理成本可能跳水,用 ChatGPT 的每个人都能感知到更快更便宜。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: OpenAI and Broadcom jointly release LLM inference chip Jalapeño. June 24, 2026. https://new.nat6.net/en/news/cmqs319vw0qkcslp52gzwr77q/