According to a report, OpenAI has entered the final stages of developing its long-rumored AI chip with the goal of decreasing the company’s dependence on Nvidia. Reuters released a report on Monday. The ChatGPT creator intends to send its chip design to Taiwan Semiconductor Manufacturing Co. for fabrication within the coming months, but it has not been formally announced. The OpenAI chip is still not known to have its full capabilities. Its technical details and timeline are also unknown. However, the company intends to improve the design over time and use it as leverage in negotiations with suppliers. This could give the company the ability to control the chip design for the future. Other tech companies such as Microsoft
AmazonGoogleand Metahave created their own AI accelerator chips in the past for a variety of reasons, including cost reduction or to relieve shortages of AI processors supplied by Nvidia. Nvidia enjoys a near monopoly on high powered GPUs (such the Blackwell series for data center use).
We covered a report in October 2023 about OpenAI’s intention to create their own AI accelerator chip for similar reasons. OpenAI’s custom chip project, therefore, has been in the works since some time. OpenAI CEO Sam Altman began traveling extensively around the globe in early 2024 to raise up to $7 trillion for increasing world chip fabrication capacity.
