SAN FRANCISCO — OpenAI has unveiled its first custom-built artificial intelligence processor, developed in partnership with Broadcom, as the company accelerates investment in next-generation AI infrastructure and high-performance computing.

The new inference chip, named Jalapeño, has been specifically engineered to run AI models more efficiently by handling the complex calculations required to generate responses for applications such as conversational AI and coding assistants. The move is designed to reduce reliance on third-party graphics processors while improving performance, energy efficiency, and long-term scalability.

Industry experts say the announcement reflects a broader shift across the U.S. technology sector, where leading AI companies are investing heavily in proprietary hardware to meet surging demand for generative AI services. Custom silicon is expected to play an increasingly important role as organizations expand AI-powered products, cloud computing platforms, and enterprise applications.

The chip was designed in collaboration with Broadcom and is expected to become part of OpenAI's expanding AI infrastructure strategy. Analysts believe purpose-built processors could help lower operational costs while delivering faster inference performance for future AI systems. The development also reinforces the United States' leadership in advanced semiconductor innovation and artificial intelligence research.

As investment in AI infrastructure continues to surge across the technology industry, businesses and investors are closely watching how custom chips, cloud computing, and next-generation data centers will shape the future of America's digital economy.