AMD named OpenAI as an early design partner for MI450 GPU, with CEO Sam Altman expressing strong enthusiasm for the partnership.

At the Advancing AI keynote in San Jose, California on Thursday, AMD CEO Lisa Su announced that OpenAI has served as a "very early design partner" for the company's upcoming Instinct MI450 GPU, which she said will usurp Nvidia's fastest AI chips next year. Su disclosed that ChatGPT has provided "significant feedback on the requirements for next-generation training and inference" regarding the MI450. OpenAI CEO and founder Sam Altman then took the stage to personally endorse the GPU, saying he is "extremely excited for the MI450." Altman praised its memory architecture as "great for inference" and stated, "I believe it can be an incredible option for training as well." He reflected on his initial skepticism of the project's ambitious specs, saying, "When you first started telling me what you're thinking about for the specs, I was like, there's no way. That just sounds totally crazy. It's too big. But it's really been so exciting to see you all get close to delivery on this. I think it's going to be an amazing thing."

AMD's MI400 series will pack 432 GB of HBM4 memory, offering 50 percent more memory capacity and bandwidth than Nvidia's Vera Rubin platform while delivering roughly the same compute performance. The company's Helios server rack, designed by AMD "from the grounds up as a rack-scale solution," will contain seventy-two of the MI450 GPUs. Su stated that when "Helios launches in 2026, we believe it'll set a new benchmark for AI at scale."

Altman explained that OpenAI faces substantial computing demands due to its shift toward reasoning models, which has "put pressure on model efficiency and long, complex rollouts" given the lengthy responses these models generate. He emphasized the urgency of their infrastructure needs: "We need tons of compute, tons of memory, tons of CPUs as well. And our infrastructure ramp over the last year and what we're looking over the next year has just been a crazy, crazy thing to watch."

Su described AMD's multi-year collaboration with OpenAI, noting that the partnership evolved through joint work with Microsoft Azure, which has served as an important cloud partner to both companies. She highlighted a key principle that emerged from their collaboration: "One of the things that really sticks in my mind is when we sat down with your engineers, you were like, 'Whatever you do, just give us lots and lots of flexibility because things change so much. And that framework of working together has been phenomenal.'" Beyond OpenAI, AMD also received on-stage endorsements for its Instinct GPUs from executives at Microsoft, Meta, Cohere, and Oracle Cloud Infrastructure on Thursday.