OpenAI's Jalapeño chip: Q&A with VP of hardware Richard Ho

OpenAI's VP of Hardware Richard Ho detailed the company's Jalapeño inference chip in a Q&A with tech analyst Dr. Ian Cutress. The chip was co-designed with Broadcom and represents OpenAI's push into custom silicon for running AI models at inference time.

The discussion covered how OpenAI used its own internal models to inform and optimize the chip's design, a notable example of AI-assisted hardware engineering. The Jalapeño chip is part of a broader industry shift toward purpose-built silicon for AI workloads, addressing both performance and cost concerns for inference at scale.