OpenAI unveiled its first custom processor, dubbed Jalapeño, on June 24, 2026. The chip is specifically designed to run large language models, marking a significant shift in the company's hardware strategy.
The development cycle for the Jalapeño chip was completed in approximately nine months. This timeline is notably shorter than the industry standard of three years, a feat achieved by utilizing OpenAI's own AI models to assist engineers during the design process.
Jalapeño serves as the initial component of a broader infrastructure ecosystem. The project is part of a larger initiative to deploy 10 gigawatts of computing power powered by OpenAI's proprietary accelerators.
This development signals a new phase where AI contributes to its own hardware foundation. These chips will support the training and operation of future AI generations, which are expected to further optimize hardware design.
