By hmimcu 
ChatGPT has brought a high AI craze. Recently, it has been reported that technology giant Microsoft is developing its own artificial intelligence (AI) chip to provide power for the key technology behind chat robot ChatGPT.
According to the report, both sources were directly familiar with the project. They mentioned that Microsoft began this plan to develop AI chips as early as 2019, which is codenamed “Athena” within the company.
One of them stated that the chips have been provided to a small number of Microsoft employees, as well as employees of ChatGPT developer OpenAI, who are currently testing and using these chips. The main purpose is to train large language models and support reasoning, which is a key technology behind Generative Artificial Intelligence (AIGC).
When users use ChatGPT, this software needs to recognize the instructions issued by humans and imitate them to answer the corresponding questions. This process requires processing a large amount of data, and requires extremely high chip computing power. Microsoft hopes that its self-developed chips will perform better than chips purchased from other suppliers, thereby saving it time and costs in expensive AI business.
It is reported that the “Athena” chip will be used for large-scale language model training required for generative AI. And this will compete head-on with Nvidia, which currently dominates the market for such chips. It is worth mentioning that technology giants such as Amazon, Google, and Facebook are also developing AI chips independently, but still rely on Nvidia’s chips to support training of large language models.