AMD has introduced its next-generation AMD Instinct MI400 Series GPUs, expanding its AI infrastructure portfolio with accelerators designed for frontier AI, sovereign AI, and high-performance computing (HPC) environments.
Announced at Advancing AI 2026, the new lineup includes the AMD Instinct MI455X GPU for large-scale AI factories and frontier model training, and the AMD Instinct MI430X GPU for sovereign AI deployments and scientific computing workloads.
The launch comes as organizations increasingly deploy AI across diverse environments, from hyperscale data centers running large language models to national AI infrastructure and research institutions requiring high levels of security, numerical accuracy, and operational control.
According to AMD, the MI400 Series combines advanced compute performance, high-bandwidth HBM4 memory, scalable networking, and enhanced security features to support demanding AI and HPC applications.
“The next generation of AI will span frontier AI, sovereign AI and scientific computing, and each requires infrastructure optimized for its unique demands,” said Vamsi Boppana, senior vice president for AI at AMD. He added that the MI400 Series is designed to provide purpose-built solutions across a broad range of AI deployments while maintaining flexibility through an open software ecosystem.
A key component of the portfolio is the AMD Helios rackscale solution, powered by MI455X GPUs and engineered for large-scale AI training and inference workloads. Meanwhile, the MI430X GPU targets government, research, and scientific institutions, delivering up to 288 TFLOPS of FP64 performance for computationally intensive simulations and scientific applications.
The GPUs are powered by AMD ROCm, the company’s open software platform that includes compilers, libraries, runtimes, and deployment tools for AI and HPC workloads. AMD said its commitment to open standards helps organizations avoid vendor lock-in while maintaining interoperability across hardware and software environments.
Security is also a major focus of the new platform, with features including secure boot, encrypted GPU-to-GPU communication, and hardware-based protections designed to safeguard sensitive AI and scientific workloads.
With the MI400 Series, AMD aims to address growing demand for scalable and specialized AI infrastructure as enterprises, governments, and research organizations accelerate investments in next-generation AI systems.
