According to reports,OpenAIWorking withBroadcom(Broadcom) and other chip design companies to discuss jointly developing a new type of artificial intelligence (AI)Server chipsThe move, led by OpenAI CEO Sam Altman, is aimed at boosting the company's computing power in AI development. The project is seen as a response to the current shortage of graphics processing units (GPUs) needed for AI model development, which are currently provided mainly by Nvidia.
Source Note: The image is generated by AI, and the image is authorized by Midjourney
In addition, OpenAI also plans to recruit former employees who participated in the development of the Tensor Processing Unit (TPU) at Google to assist in chip design. Broadcom has worked with Google to develop TPU, which gives it a certain advantage in experience in customizing AI chips.
OpenAI neither fully confirmed nor denied the news. A spokesperson said, "OpenAI is in ongoing dialogue with industry and government stakeholders to increase the accessibility of AI infrastructure and ensure that the benefits of AI can benefit society broadly."
OpenAI previously revealed its intention to develop custom AI chips in September 2023, and in January 2024, Sam Altman was reported to be seeking to raise funds from institutions such as SoftBank Group and Abu Dhabi G42 to support new chip manufacturing projects. Although the amount of funds raised mentioned in the report is as high as 7 trillion US dollars, this figure far exceeds the total economic output of Australia and Japan, which is obviously too large, and may prompt OpenAI to choose to cooperate with existing companies such as Broadcom to reduce costs and achieve its goals.
This move by OpenAI also shows its ambition in the field of AI chips. Currently, Nvidia's share of the AI chip market is about 70% to 95%. OpenAI's development strategy is obviously aimed at reducing its dependence on Nvidia and thus enhancing its market competitiveness.