At CES 2026, semiconductor giant AMD unveiled a broad lineup of new processors, including AI accelerators for data centers and refreshed mobile chips targeting laptops and handheld gaming devices.
AMD CEO Lisa Su took the stage at the 2026 Las Vegas Consumer Electronics Show to introduce a suite of next-generation artificial intelligence solutions. Central to the presentation was the "Helios" rack-scale platform, which AMD describes as a foundational blueprint for future AI computing infrastructure.
OpenAI President Greg Brockman joined the keynote, underscoring how advancements in chip technology are critical to meeting OpenAI’s massive computational demands. The October partnership between AMD and OpenAI is projected to generate billions in annual revenue for AMD. According to the company, the first shipments of its MI400 series GPUs will begin this year.
MI400 Lineup Complete, MI500 Series on the Horizon
AMD officially revealed the full Instinct MI400 product stack. The flagship MI455X powers the Helios platform, delivering high-performance computing capabilities. The company also introduced the MI440X, designed specifically for enterprise on-premises deployment. This compact eight-GPU solution supports training, fine-tuning, and inference workloads while integrating seamlessly into existing data center environments.
The previously announced MI430X will drive major supercomputing initiatives, including Oak Ridge National Laboratory's Discovery system and Alice Recoque, France’s first exascale machine.
Looking ahead to 2027, Su previewed the upcoming MI500 series, built on CDNA 6 architecture using a 2nm process node and equipped with HBM4E memory. AMD claims the MI500 series will deliver up to 1,000 times the AI performance of the MI300X released in 2023.
While AMD remains NVIDIA’s most formidable competitor in the AI accelerator space, it has yet to match the scale of NVIDIA’s success. NVIDIA reported hundreds of billions in quarterly revenue from AI chips and simultaneously launched its Vera Rubin platform on the same day.
Gorgon Point Brings Higher Clock Speeds to Laptops
Beyond data center innovations, AMD announced the Ryzen AI 400 series for notebooks, codenamed “Gorgon Point.” Seven new processors under this family are expected to ship starting in Q1 2026.
The Gorgon Point chips continue to use TSMC’s N4X process and combine standard Zen 5 cores with more compact Zen 5c variants. Key improvements over the previous generation include higher clock frequencies and faster memory support.
AMD also introduced its first desktop processor certified under Microsoft’s Copilot+ PC program. Specific specifications and launch timelines were not disclosed. This new desktop chip belongs to the Gorgon Point family and features Zen 5 CPU cores, RDNA 3.5 GPU architecture, and an XDNA 2-powered NPU.
The company expanded its embedded computing portfolio with the new Ryzen AI Embedded processors—targeting edge AI applications. The P100 and X100 series are aimed at automotive digital cockpits, smart healthcare systems, and autonomous platforms such as humanoid robots.
New Gaming Chips Could Halve Handheld Device Prices
Within the Ryzen AI Max series—also known as Strix Halo—AMD added two new SKUs: the Ryzen AI Max+ 392 (12-core) and Max+ 388 (8-core). Both retain full graphics configurations with 40 RDNA 3.5 compute units, achieving 60 teraflops of peak performance.
These chips were developed based on specific customer requirements for gaming-focused devices. Current systems featuring similar hardware, such as the GPD Win 5 or Framework Desktop, retail around $2,000. However, AMD executive Rahul Tikoo indicated that next-gen systems could see prices drop to between $1,000 and $1,500, though ongoing global memory supply constraints may impact final pricing.
