OpenAI has unveiled its first custom artificial intelligence chip, Jalapeño, marking a major step in the tech giant’s efforts to build more of the computing infrastructure that powers its AI models and products, the company announced on Wednesday.
The chip, developed in partnership with semiconductor company Broadcom, is designed specifically for large language model (LLM) inference, which is the process through which AI systems such as ChatGPT generate responses to users.
OpenAI said Jalapeño is the first accelerator in a planned multi-generation computing platform intended to make AI services faster, more reliable and more affordable.
OpenAI said early testing indicates the chip will deliver a significant efficiency advantage over current industry-leading hardware, though final performance figures are still being measured.
- “While OpenAI is still measuring final performance, early testing shows that Jalapeño will deliver performance per watt substantially better than current state-of-the-art,” the company said, adding that a detailed technical performance report will be presented in the coming months.
Greg Brockman, President and Co-Founder of OpenAI, tied the chip to the company’s broader infrastructure ambitions.
- “The world is moving to a compute-powered economy. Jalapeño is part of our long-term full-stack infrastructure strategy to make compute more abundant, resulting in AI which is faster, more reliable, more affordable for people and businesses, and can be used to solve more important problems,” he said.
Richard Ho, who leads OpenAI’s hardware program, said the chip was built entirely around the specific demands of large language model inference.
- “We optimized the architecture around the kernels, memory movement, networking, and serving patterns that matter most for frontier AI models. Based on early testing, Jalapeño will efficiently execute our most important workloads close to the hardware’s theoretical limits,” he said.
Broadcom CEO Hock Tan described the partnership as part of a much longer-term infrastructure commitment.
- “Our collaboration with OpenAI represents a fundamental commitment to scaling the physical infrastructure required for the next decade of AI. This is just the beginning of a multi-generation roadmap,” he said, adding that the collaboration would enable gigawatt-scale data centre deployments with Microsoft and other partners beginning in 2026.
