In the CES 2026 opening keynote, AMD CEO and Chair Dr. Lisa Su unveiled new products across the client, graphics, embedded, and commercial segments. From the world’s fastest gaming processor to powerful new AI PCs and next-generation AI accelerators, AMD showcased how it is advancing AI innovation from consumer devices to the data center.
Below is a high-level recap of the key AMD announcements, links to the press releases, and a list of industry analysts who can provide additional context on the news. A news overview sheet is also attached.
Client and Graphic
- New Ryzen™ AI 400 and Ryzen AI PRO 400 Series processors deliver next-generation AI experiences for both consumer and commercial users. Built on the advanced “Zen 5” architecture and featuring AMD XDNA™ 2 NPUs, both series offer up to 60 TOPS NPU of AI compute, exceeding the requirements of Microsoft Copilot+ PCs and enabling powerful AI performance.
- Additional SKUs in the AMD Ryzen™ AI Max+ Series extend the portfolio’s high-performance AI compute, integrated desktop-class graphics, and unified memory architecture to premium ultra-thin laptops, workstations, and compact mini-PCs. These new processors are optimized for demanding creative and AI workloads, and immersive gameplay without compromise.
- The new AMD Ryzen™ AI Halo mini-PC is designed to accelerate AI development. Built on the Ryzen AI Max+ platform, this compact unit offers desktop-grade AI compute, integrated graphics, and easy access to pre-installed applications and models, making it ideal for running LLMs locally with minimal friction.
- The Ryzen™ 9850X3D processor joins the Ryzen X3D family as the fastest gaming processor on the market. Utilizing powerful “Zen 5” CPUs and 2nd Gen AMD 3D V-Cache™ technology, the processor delivers impressive performance gains and maximum efficiency in today’s most demanding titles.
- The new ROCm 7.2 update supports the Ryzen AI 400 Series processors and is now available as an integrated download through ComfyUI.
- ROCm has seen a 10x year-over-year increase in downloads, driven by doubled platform support across Ryzen and Radeon in 2025 and expanded availability across Windows and Linux distributions.
Data Center
AMD unveiled the full MI400 Series lineup, built to serve every AI market. Leading the portfolio is Helios, AMD’s blueprint for yotta‑scale compute, delivering up to 3 AI exaflops in a single rack for trillion-parameter training with maximum bandwidth and efficiency, creating the blueprint for yotta-scale compute.
- The lineup also includes:
- AMD Instinct MI440X, an 8-GPU on-premise solution targeting enterprise AI for training, fine-tuning, and inference workloads that integrates seamlessly into existing infrastructure.
- AMD Instinct MI430X is a high-precision GPU solution targeting sovereign AI, HPC, and hybrid computing.
- AMD disclosed MI500 Series details (launching in 2027), which it set to deliver up to a 1,000x increase in AI performance compared to the AMD Instinct MI300X with AMD CDNA™ 6 architecture, 2nm process technology, and HBM4E memory.
Embedded
- New Ryzen™ AI Embedded P100 and X100 Series processors with “Zen 5” CPUs, RDNA™ 3.5 graphics, and XDNA™ 2 NPUs, deliver up to 50 AI TOPS in a single compact chip and feature 3x AI performance and 35% faster GPU rendering power for real-time 4K/8K visualization. The space-efficient design and unified open software stack help to accelerate and simplify embedded system development for AI-driven automotive, industrial, and autonomous systems.
- AMD will announce its collaboration with Generative Bionics to develop the GENE1.0, a next-generation humanoid robot powered by AMD, showcasing a new category of compute that merges robotics, AI, and human-centric design.
