OpenAI is seeking alternatives to NVIDIA's AI chips due to cost and supply concerns, with AMD positioned as the primary beneficiary gaining market share.

Another developing news story is that OpenAI is not satisfied with NVIDIA's latest AI chips and has already begun actively seeking alternative solutions. This dissatisfaction is not a negation of NVIDIA, but rather focuses on cost, supply cadence, and efficiency and latency performance in certain inference scenarios; OpenAI is starting to reduce reliance on a single supplier. In OpenAI's path to finding NVIDIA alternatives, AMD is the fastest to fill the gap; OpenAI is generally believed by the market to have started adopting AMD's MI300 series GPUs as one of the supplementary solutions to NVIDIA. The reason investment negotiations between OpenAI and NVIDIA have been delayed for months is partly due to the key variable of AMD's involvement. As long as OpenAI proves it can maintain operations of part of its business without using NVIDIA, it gains leverage. Additionally, OpenAI's adoption of AMD chips aligns with Microsoft's hardware strategy, allowing direct reuse of Microsoft's existing AMD computing infrastructure, supply chain relationships, and operations and maintenance system to reduce switching and expansion costs. The relationship between OpenAI and NVIDIA is quite complex.