OpenAI is in discussions with chip designers including Broadcom to develop new AI chips to reduce reliance on scarce and expensive graphics processing units (GPUs) used in developing AI models such as ChatGPT, GPT-4 and DALL-E3, sources familiar with the matter told The Information.
To reduce this reliance, OpenAI is considering developing its own AI chips. The Microsoft-backed company has hired ex-Googlers who developed the online search giant's own AI chip, the Tensor Processing Unit, a sign of its commitment to developing AI server chips, the article added, citing three people familiar with the matter.
“OpenAI is in ongoing discussions with industry and government stakeholders about increasing access to the infrastructure needed to make the benefits of AI broadly available,” an OpenAI spokesperson told The Information.
Also, Bloomberg News reported earlier this year that OpenAI CEO Sam Altman was looking to raise billions of dollars to build a network of semiconductor foundries, teaming up with major chipmakers such as Intel, Taiwan Semiconductor Manufacturing Co. (TSMC) and Samsung Electronics Co. as potential partners for the highly ambitious project.
In this regard, OpenAI's move to develop AI chips will significantly reduce reliance on external GPU suppliers and increase the development capabilities of AI models. OpenAI partners with leading companies in the industry and tends to invest heavily in the semiconductor manufacturing process, ensuring a robust and self-sufficient infrastructure for the continued advancement of AI innovation. This puts its custom hardware at the forefront of artificial intelligence advancements.
