AMD has announced that its most advanced GPU for artificial intelligence (AI), the MI300X, will begin shipping to select customers later this year. This move poses a significant challenge to Nvidia, which currently dominates the AI chip market with over 80% market share.
Why it matters: This development represents a significant competitive challenge to Nvidia’s dominance in the AI chip market. Technology leaders need to stay informed about advancements in AI chip technologies and their potential impact on the market landscape.
- AMD’s AI chips, or “accelerators,” have the potential to tap into a substantial untapped market, providing an alternative to Nvidia’s products. AMD CEO Lisa Su highlighted AI as the company’s largest long-term growth opportunity, with the data center AI accelerator market projected to reach over $150 billion by 2027.
- While Nvidia has traditionally been favored by AI developers due to its software package CUDA, AMD has made progress with its own software stack called ROCm, which works with an open ecosystem of models, libraries, frameworks, and tools.
- Evaluating alternative options and considering the performance, features, and cost of AI chips can help inform strategic decisions around infrastructure, development, and deployment of AI applications.