OpenAI and Broadcom Release AI Chip Jalapeño, Cutting Inference Costs by Up to 50%

CoinFeed June 24 news, OpenAI and Broadcom jointly released the first AI chip Jalapeño optimized for large language models (LLMs). Broadcom CEO stated that the chip can reduce model operation costs by about 50% and is expected to achieve large-scale deployment faster than market expectations. Meanwhile, Broadcom anticipates starting gigawatt-scale data center construction in 2026 with Microsoft and partners to support next-generation AI computing demands. After the announcement, Broadcom's pre-market stock price rose about 3%.