AMD continues to redefine high-performance computing (HPC) and artificial intelligence (AI), building on its legacy of powering the world’s most advanced supercomputers. From Frontier—the first exascale-class system at Oak Ridge National Laboratory delivering 1.3 exaflops—to El Capitan at Lawrence Livermore, now the fastest supercomputer globally at 1.8 exaflops, AMD has set the benchmark for scientific and enterprise innovation.
The newly launched AMD Instinct™ MI430X GPU takes this leadership to the next level. Built on the next-generation AMD CDNA™ architecture, the MI430X offers 432GB of HBM4 memory and an unprecedented 19.6TB/s memory bandwidth. These capabilities enable researchers, engineers, and AI innovators to train massive language models, run complex simulations, and accelerate breakthroughs in energy, healthcare, and materials science—all on a single platform.
“The AMD Instinct MI430X is purpose-built for the convergence of AI and HPC workloads.”
Key deployments underscore its transformative potential. Discovery at Oak Ridge National Laboratory will serve as one of the U.S.’s first AI Factory supercomputers, leveraging MI430X GPUs and AMD EPYC “Venice” CPUs to advance sovereign AI and scientific computing. In Europe, Alice Recoque integrates MI430X GPUs on Eviden’s BullSequana XH3500 platform, delivering exceptional performance for double-precision HPC and AI workloads while meeting stringent energy efficiency goals.
“The AMD Instinct MI430X is purpose-built for the convergence of AI and HPC workloads,” says an AMD spokesperson. “Its massive memory bandwidth and precision support ensure balanced performance for both scientific computing and large-scale AI training—without compromise.”
Combined with AMD ROCm™ software and full-stack integration with frameworks like PyTorch and TensorFlow, MI430X delivers scalability across thousands of GPUs, ensuring optimized performance for training and inference.
From Discovery in the U.S. to Alice Recoque in Europe, AMD Instinct MI430X is driving a new generation of AI factories and scientific platforms built on open standards and energy-efficient design. As HPC and AI converge, AMD reinforces its mission: enabling global innovation and powering the next breakthroughs in science and technology.
