OpenAI & Broadcom Launch ‘Jalapeño’: A Groundbreaking Custom AI Chip
In a significant development for the artificial intelligence industry, OpenAI and semiconductor titan Broadcom (NASDAQ: AVGO) have officially launched Jalapeño — OpenAI’s inaugural custom-built AI inference chip. Revealed on June 24, 2026, this processor signifies a crucial shift in OpenAI’s strategy to possess not only the software but also the hardware stack powering its models.
Jalapeño is an Application-Specific Integrated Circuit (ASIC) tailored specifically for large language model (LLM) inference — the procedure of running AI models in response to user queries. Unlike general-purpose Nvidia GPUs, Jalapeño was engineered from scratch based on OpenAI’s comprehension of its frontier models’ behavior. Preliminary testing indicates that the chip will deliver a performance per watt that is “substantially superior to current state-of-the-art” hardware.
One of the most impressive aspects is its rapid development speed: the chip transitioned from initial design to manufacturing tape-out in a mere nine months. This is believed to be the fastest ASIC development cycle ever achieved in high-performance advanced semiconductors. OpenAI even utilized its own AI models to expedite parts of the chip design process.
Broadcom CEO Hock Tan confirmed that Jalapeño will be deployed at gigawatt-scale data centers in collaboration with Microsoft and other partners later in 2026. OpenAI president Greg Brockman encapsulated the vision: “By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access.”
This strategic move places OpenAI alongside Google (TPUs), Amazon (Trainium), and Meta (MTIA) as a company that designs its own silicon — a growing necessity as AI compute costs continue to pose an existential challenge.
Source: TechCrunch — OpenAI unveils its first custom chip, built by Broadcom