OpenAI has partnered with Broadcom to develop and deploy 10 gigawatts of custom AI chips and computing systems. This collaboration is a significant step towards meeting OpenAI's rapidly growing computing demands. The partnership will span over the next four years, involving billions of dollars in investment. OpenAI is intending to design its own GPUs, using its AI model expertise to inform hardware development. Broadcom will co-develop the specialized chips with OpenAI. These chips will be deployed starting in the second half of next year. Broadcom's expertise lies in custom AI chip design, catering to specific AI applications. The partnership began 18 months ago with custom chip creation. The collaboration has expanded to include server racks and networking equipment. This deal signifies the growing hardware focus of AI giants.