OpenAI partnered with Broadcom to release a proprietary chip aimed at dramatically cutting the cost of running large language models.
Tag: custom
OpenAI Unveils Jalapeño Custom AI Inference Chip Developed in Partnership With Broadcom
The chip delivers significantly better performance per watt than Nvidia hardware for inference tasks, joining Google and Amazon in developing custom AI silicon.
OpenAI Unveils Jalapeño Custom AI Inference Chip Developed in Collaboration With Broadcom
The chip is designed to reduce OpenAI’s reliance on Nvidia hardware for inference tasks, delivering significantly better performance per watt in early testing.