United States firms unveil custom Jalapeño AI processor | JQJO

United States – OpenAI and Broadcom have jointly introduced Jalapeño, a custom intelligence processor designed to power next-generation, large-scale artificial intelligence models, marking OpenAI’s formal entry into semiconductor design. The companies said the chip is engineered specifically for the computational demands of both AI training and inference, aiming to deliver higher efficiency and speed than prevailing industry-standard hardware. OpenAI’s engineers worked with Broadcom to tailor the chip’s architecture to the needs of deep-learning clusters and high-demand workloads, aligning it with the company’s future AI systems and long-term infrastructure plans.

United States – The collaboration draws on Broadcom’s established chip-manufacturing and design expertise while incorporating OpenAI’s detailed technical requirements for its data center operations. Executives framed the Jalapeño launch as part of a broader effort to reduce reliance on traditional third-party chip architectures and to strengthen control over critical hardware amid ongoing global chip shortages. Industry analysts say the processor is optimized to improve performance and potentially lower operational costs in large-scale data centers running intensive AI workloads. Neither OpenAI nor Broadcom has provided a specific timeline for mass rollout or integration of Jalapeño into existing OpenAI facilities.