OpenAI Launches First Custom AI Chip

OpenAI has unveiled its first custom inference processor, called Jalapeño, developed in partnership with Broadcom. The chip was designed specifically for OpenAI's inference workloads, with the company's own AI models assisting in its development. Early tests show significantly better performance-per-watt than current alternatives. The move follows similar efforts by Google and Amazon to reduce reliance on Nvidia GPUs. Jalapeño targets low-cost real-time inference, particularly for coding models, while more intensive tasks like pre-training will likely still depend on Nvidia hardware. OpenAI says the chip is part of a broader strategy to optimize its full technology stack.