At Advancing AI 2026, AMD unveiled “AMD Kria AI Solutions,” a new robotics product line. Combining the Kria AI System-on-Module (SOM), a robotics developer platform, and an open software ecosystem, AMD says it aims to accelerate development for physical AI — AI that operates in the real world, such as robots. At its core is FPGA technology inherited from AMD’s $35 billion acquisition of Xilinx in 2022, which the company says enables low-latency processing such as 23-degree-of-freedom robotic hand control. Foundation Future Industries’ humanoid robots already use AMD’s X100 series chips.
What This Article Covers
- The technical characteristics of Kria AI SOM and the role of Xilinx-derived technology
- Where AMD stands relative to NVIDIA in the robotics market
- Actual adoption cases
Overview
The market for robotics-focused AI chips has largely been led by NVIDIA. AMD is aiming to make a serious push into this market centered on FPGA (Field-Programmable Gate Array — a semiconductor whose circuit configuration can be rewritten for the application) technology from its 2022 Xilinx acquisition. FPGAs excel at low-latency, low-power processing, making them well suited to real-time applications like robot joint control and sensor processing.
Details
The following points were confirmed through a primary source — AMD’s official newsroom.
- At Advancing AI 2026, AMD unveiled “AMD Kria AI Solutions,” combining the Kria AI SOM, a robotics developer platform, and an open software ecosystem, aimed at accelerating physical AI robotics development
- The Kria AI SOM uses the AMD Ryzen AI Embedded X100 series, featuring heterogeneous computing that integrates CPU, GPU, NPU, and FPGA
- FPGA technology inherited from the 2022 Xilinx acquisition is said to enable low-latency control, such as 23-degree-of-freedom robotic hand control
- Foundation Future Industries’ Phantom humanoid robot uses the X100 series, with AMD claiming 2.5x faster inference and 3x faster training compared to NVIDIA
Industry Impact
AMD’s entry into robotics shows that competition in AI accelerators is expanding beyond data centers and edge PCs into the new frontier of physical AI (robotics). It’s also notable that, four years after the Xilinx acquisition, FPGA technology is now clearly being positioned as one of AMD’s key differentiators. That said, the claimed performance advantage over NVIDIA comes from AMD itself, and no third-party verification exists yet.
Relevance to Japan
Nothing in the announcement reviewed for this article directly mentions Japanese companies. That said, Japan has a strong robotics industry, and the competition between AMD and NVIDIA for robotics-focused chips and platforms is a topic worth watching for its relevance to component sourcing decisions among Japanese robot makers.
What to Watch Next
- Customer and adoption cases beyond Foundation Future Industries
- Third-party verification of the claimed performance advantage over NVIDIA
- Kria AI SOM’s mass production timeline and pricing
- How the competitive dynamic with NVIDIA’s robotics product strategy evolves
Summary
AMD has made a serious entry into the robotics market with “AMD Kria AI Solutions,” built around Xilinx-derived FPGA technology to pitch low-latency processing. Foundation Future Industries has already adopted the chips, but the claimed performance edge over NVIDIA remains AMD’s own assertion. What’s next is to track additional customer cases and third-party verification.
