Advanced Micro Devices (NASDAQ: AMD) generated $5 billion in AI accelerator revenue for 2024, significantly trailing market leader Nvidia (NASDAQ: NVDA), whose data center segment yielded over $39 billion in the most recent quarter. AMD’s new GPUs, the MI350X and MI355X, deliver four times the AI compute performance and 35 times the AI inferencing performance compared to previous models, but still fall short of Nvidia’s latest GPUs.
AMD plans to launch its next-generation MI400 series AI accelerators in 2026, claiming the flagship MI400X will be ten times more powerful than the MI300X, featuring up to 432 GB of HBM4 memory and 19.6 TB/s memory bandwidth. Meanwhile, Nvidia’s new Vera Rubin chips are also set for release, expected to triple performance from earlier models. Despite AMD’s advancements, it is projected to remain in second place behind Nvidia, with the AI chip market expected to grow to $500 billion by 2028.
AMD plans to introduce a new rack-scale AI solution called Helios alongside the MI400 series in 2026, which will have up to 72 MI400 GPUs and advanced server CPUs. The growing demand for scalable AI infrastructure suggests potential growth for AMD’s AI-related revenue in the coming years, despite facing stagnant competition from Nvidia.
