AMD has introduced the Ryzen AI Embedded P100 and X100 x86 processors designed to power AI-driven applications at the edge, such as automotive digital cockpits, smart healthcare equipment, and humanoid robotics.
AMD had previously launched several Ryzen Embedded families, starting with the Ryzen Embedded V1000 in 2018, and more recently, the Ryzen Embedded 8000 family with a 16 TOPS NPU, yet without “AI” in the name. The new AMD Ryzen AI Embedded family is based on high-performance “Zen 5” core architecture, an RDNA 3.5 GPU for real-time visualization and graphics, and an XDNA 2 NPU for low-latency, low-power AI acceleration.
The P100 Series processors target in-vehicle infotainment and industrial automation, while the X100 Series processors feature higher CPU core counts and AI TOPS performance for more demanding physical AI and autonomous systems.
Six models are currently available in the Ryzen AI Embedded P100 family: AMD Ryzen AI P121, P123, P121i, P132i, P122a, and P132a for commercial, industrial (i), and automotive (a) applications.
Commercial Temp | Industrial Temp | Automotive Grade |
|||||
|---|---|---|---|---|---|---|---|
| Model # | P121 | P132 | P121i | P132i | P122a | P132a |
|
| CPU | “Zen 5” CPU Cores | 4 | 6 | 4 | 6 | 4 | 6 |
| Max Frequency | Up to 4.4 GHz | Up to 4.5 GHz | Up to 4.4 GHz | Up to 4.5 GHz | Up to 3.65 GHz | Up to 3.65 GHz |
|
| L3 Shared Cache | 8 MB |
||||||
| GPU | Work Group Processors | 1 | 2 | 1 | 2 | 2 | 2 |
| 4K120/8Kp120 Displays | 4 / 2 |
||||||
| GPU Max Frequency | 2.7 GHz | 2.8 GHz | 2.7 GHz | 2.8 GHz | 2.0 GHz | 2.4 GHz |
|
| NPU | TOPS | 30 | 50 | 30 | 50 | 30 | 50 |
| Memory | DDR5 (ECC) | 5600 MT/s | N/A |
||||
| LPDDR5X (ECC) MT/s | 7500 MT/s | 8000 MT/s | 7500 MT/s | 8000 MT/s | 7500 MT/s w/RAS | 7500 MT/s w/RAS |
|
| I/O | 10GE Ports w/TSN | 2 |
|||||
| USB 4.0 | 2x USB4 | N/A |
|||||
| Other USB | 1x USB 3.2 | 1x USB3.1 | 3x USB2 | 1x USB2 (Secure BIOS) |
||||||
| Power & Thermal | Nominal TDP | 28 W | 28 W | 28 W | 28 W | 28 W | 45 W |
| TDP Range | 15-54 W | 15-54 W | 15-54 W | 15-54 W | 15-30 W | 25-45 W |
|
| Junction Temperature | 0 to 105°C | -40 to 105°C |
|||||
| Package, Reliability | Package | 25 mm x 40 mm |
|||||
| Longevity | 2.5 Years (Standard) | Up to 10 Years (Extended) | AEC-Q100 |
|||||
We’re not shown a similar table for the X100 family, as they look to be formally launched a little later. The AMD Ryzen AI Embedded P100 Series is said to deliver up to a 2.2x multi-thread and single-thread performance boost over the “previous generation”. More specifically, the AMD Ryzen AI Embedded P132a processor delivers up to an estimated 84% higher single-thread performance and 125% higher multi-thread performance on SPECrate®2017_int_base compared to the AMD Ryzen Embedded V2A46.
The RDNA 3.5 GPU delivers an estimated 35% faster 3D graphics rendering on GFXBench 5.0.0 Vulkan 5 normal offscreen, and can power up to four 4K (or two 8K) displays simultaneously at 120 frames per second. Finally, the AMD XDNA 2 NPU delivers up to 50 TOPS, or up to 3x higher AI inference performance using AI models such as vision transformers, compact LLMs, and CNNs.
The new Ryzen AI Embedded processors are supported by the Ryzen AI Software. The software stack is built on the open-source, Xen hypervisor-based virtualization framework to enable multiple operating systems to run simultaneously in isolation. For instance, Yocto Linux or Ubuntu can power the HMI, while FreeRTOS manages real-time control, and Android or Windows runs richer applications, all running safely in parallel.
The 4-6-core AMD Ryzen AI Embedded P100 processors mentioned above are already sampling to early access customers. Tools and documentation are available, and mass production is expected to start in Q2 2026. The more powerful 8-12-core P100 Series processors for industrial automation applications should start sampling in Q1 2026, while the up to 16-core X100 Series processors will start sampling in H1 2026, which I suppose really means Q2 2026. More details can be found on the product page and the press release.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
I’m done salivating over these ECC-capable Ryzen-embedded processors. Seems like almost no vendors use them in any mainstream designs, and the exceptions are almost always either prohibitively expensive, or really hard-to-find, or both…
That’s because ECC is not usually necessary in consumer devices. So it’s used in lower volume, higher margins embedded/industrial applications. The most affordable option is probably systems using SoC with IBECC (In-Band ECC).
Indeed. Intel although with mediocre products does have still an edge on ECC support in price-optimized solutions at least.