According to what it was produced, Openai is producing its artificial intelligence chips, starting next year

Openai is preparing to start the mass production of its artificial intelligence chips next year to be able to provide the huge computing power that its users need and reduce its dependence on NVIDIA Financial times. According to what was reported, the company designed the AI ​​chip with the American semiconductor maker, who recently announced the CEO that it has had a new customer who was placed at $ 10 billion in requests. The customer did not mention, however TimesThe sources confirmed that it was Openai, who appears to have no plans to sell chips and will only use them internally.

Reuters I mentioned In 2023, Openai was already exploring the possibility of making his artificial intelligence chips after SAM Altman blamed the GPU deficiency at the company’s API speed and reliability of the company. The News Corporation previously reported that Openai was working with Broadcom and Taiwan Semiductor Manufacturing Co. (TSMC) to create its own product. the Times He didn’t say whether Openai still has a continuous partnership with TSMC.

after GPT-5 He went out, Declare Openai’s changes in order to keep pace with “increased demand”. In addition to giving priority to paid Chatgpt users, he said that Openai will double its fleet “during the next five months.” The making of its own chips will address any possible deficiency in the graphics processing unit that the company may face in doubling its fleet, and it can also provide the company’s funds. the Times The designated artificial intelligence chips called “XPUS” such as One Openai, which will eventually take a large share of the artificial intelligence market. Currently, NVIDIA is still the main name in this industry. Recently open Its revenues for the second quarter ended on July 27 increased 56 percent compared to the same period last year, and it did not even ship any H20 chips to China.