Microsoft subsequent month plans to unveil the corporate’s first chip designed for synthetic intelligence at its annual builders’ convention, in line with an individual with direct data. The transfer, a fruits of years of labor, may assist Microsoft reduce its reliance on Nvidia-designed AI chips, which have been briefly provide as demand for them has boomed.
The Microsoft chip, just like Nvidia GPUs, is designed for information middle servers that practice and run giant language fashions, the software program behind conversational AI options similar to OpenAI’s ChatGPT. Microsoft’s information middle servers at the moment use Nvidia GPUs to energy cutting-edge LLMs for cloud prospects, together with OpenAI and Intuit, in addition to for AI options in Microsoft’s productiveness apps.
