OpenAI Tackles GPU Shortage with Dive into AI Chip Development

OpenAI is reportedly in discussions with chip designers, including Broadcom, to develop a new AI chip, aiming to reduce its dependence on costly and scarce GPUs currently used for AI models like ChatGPT, GPT-4, and DALL-E3, according to sources cited by The Information.

To mitigate this reliance, OpenAI is contemplating the creation of its own AI chips. The Microsoft-backed company is recruiting former Google employees who were involved in creating Google’s AI chip, the tensor processing unit, indicating a strong commitment to developing an AI server chip, as reported by three individuals familiar with the matter.

An OpenAI spokesperson informed The Information, “OpenAI is having ongoing conversations with industry and government stakeholders about increasing access to the infrastructure needed to ensure AI’s benefits are widely accessible.”

Furthermore, Bloomberg News reported earlier this year that OpenAI CEO Sam Altman plans to raise billions of dollars to establish a network of semiconductor manufacturing plants. Potential partners for this ambitious project include major chipmakers such as Intel, Taiwan Semiconductor Manufacturing Co., and Samsung Electronics.

Developing its own AI chips would significantly decrease OpenAI’s reliance on external GPU suppliers and enhance its capabilities in AI model development. OpenAI has a history of collaborating with industry giants and making substantial investments in semiconductor production, which would ensure a robust and self-sufficient infrastructure for ongoing AI innovation. This strategic move places custom hardware at the forefront of advancing artificial intelligence.