TLDR:
OpenAI CEO Sam Altman is reportedly seeking billions of dollars to build a global network of fabricators to develop AI chips. The goal is to address the increasing demand for high-powered chips to run complex AI systems. Currently, there is a limited number of fabs capable of producing these chips, driving up costs and leading to the need for investors. Other companies, including Microsoft and Amazon, have also ventured into producing their own AI chips.
Article:
A new report from Bloomberg reveals that OpenAI CEO Sam Altman is in talks to raise billions of dollars for a global network of fabricators to develop AI chips. The aim is to address the rising demand for high-powered chips that are required to run complex AI systems.
One of the major challenges in running AI models is having enough chips to handle computations. The bots, like ChatGPT and DALL-E, that answer prompts and generate images require substantial computational power. Nvidia, with its popular H100 GPUs, has a virtual monopoly in this space and saw its value rise above $1 trillion last year. As a result, the race to manufacture more high-powered chips has intensified.
However, there are a limited number of fabrication plants (fabs) that can produce these high-end chips. This scarcity is driving Altman and others to bid for capacity years in advance, even before they need it. It means that in order to produce new chips, OpenAI and others need deep-pocketed investors who can front the costs, which the nonprofit organization still cannot afford. Bloomberg reports that SoftBank Group and Abu Dhabi-based G42 have been in talks about raising money for Altman’s project.
Other companies, including Microsoft and Amazon, have also ventured into producing their own AI chips. Microsoft recently announced its first custom AI chip for training models, while Amazon unveiled a new version of its Trainium chip. Google’s chip design team is using its DeepMind AI to design AI processors like its Tensor Processing Units (TPU). Nvidia has even announced its next-generation GH200 Grace Hopper chips to extend its dominance in the space. Competitors like AMD, Qualcomm, and Intel have also launched AI-focused processors designed for laptops, phones, and other devices.
The demand for AI chips is expected to continue to grow as more advanced AI models are developed and deployed. These models require significant computational power, and the current limited supply of high-powered chips is driving the need for increased fabrication capacity. Altman’s efforts to raise billions of dollars for a global network of fabricators highlight the urgency and importance of developing a robust chip manufacturing ecosystem to support the AI industry.