AMD Ryzen AI 400 Series: Everything You Need to Know

What Is the Ryzen AI 400 Series

AMD announced the Ryzen AI 400 series, codenamed Gorgon Point, at CES 2026 in January. Laptops started shipping later that month, with wide retail availability hitting in February and March. This is the direct successor to the Ryzen AI 300 series (Strix Point) and represents AMD's second generation of processors built specifically for the Copilot+ PC era.

The headline numbers: up to 12 Zen 5 CPU cores, clock speeds reaching 5.2 GHz, a 60 TOPS XDNA 2 neural processing unit, RDNA 3.5 integrated graphics boosting up to 3.1 GHz, and support for LPDDR5X memory at 8,533 MT/s. The configurable TDP ranges from 15W to 54W, covering everything from fanless ultrabooks to performance-oriented workstations.

We should be upfront about what Gorgon Point is and is not. This is an iterative refresh, not a generational leap. The CPU cores are the same Zen 5 architecture from Strix Point. The GPU is still RDNA 3.5. The NPU is still XDNA 2 but with higher clocks. What has changed is frequency, memory speed support, and power efficiency tuning. If you bought a Ryzen AI 300 laptop in late 2025, Gorgon Point is not a reason to upgrade. If you are buying a new laptop in 2026, it is the AMD chip you want.

Architecture Deep Dive

CPU: Zen 5 in a Hybrid Layout

The Ryzen AI 400 series uses the same hybrid core layout introduced with Strix Point. The flagship Ryzen AI 9 HX 475 has 12 cores and 24 threads arranged in two clusters: four primary Zen 5 cores optimized for single-threaded performance and eight compact Zen 5c cores designed for multi-threaded efficiency.

The Zen 5 cores handle demanding foreground tasks like compiling code, photo editing, and gaming. The Zen 5c cores take on background work like system services, browser tabs, and notifications. Windows 11's thread scheduler has matured significantly in its ability to route work to the right cluster, and AMD's CPPC (Collaborative Processor Performance Control) provides fine-grained frequency and voltage scaling that responds to workload changes in microseconds.

Compared to last year's Strix Point, Gorgon Point pushes boost clocks slightly higher (5.2 GHz versus 5.1 GHz on the previous top SKU) and improves sustained performance under prolonged loads through refined power delivery. The 36 MB of combined L2 and L3 cache remains unchanged.

GPU: RDNA 3.5 Integrated Graphics

The integrated Radeon graphics in the Ryzen AI 400 series use RDNA 3.5 architecture with up to 16 compute units boosting to 3.1 GHz. This is the same GPU architecture as Strix Point but with a 100 MHz higher boost clock.

In practical terms, the Radeon 890M integrated GPU is one of the best iGPUs on the market. In our testing, it maintained a steady 60 FPS or higher in Cyberpunk 2077 at 1080p Low settings, handled Counter-Strike 2 at 1080p Medium at over 100 FPS, and ran older titles like Civilization VI at 1440p without breaking a sweat. For a laptop with no discrete GPU, this is impressive.

The iGPU also handles hardware video encoding and decoding, supporting AV1, H.265, and H.264. For creators who edit video in Premiere Pro or DaVinci Resolve, the hardware encoder accelerates exports significantly compared to CPU-only rendering.

NPU: XDNA 2 at 60 TOPS

The neural processing unit is where AMD made the most measurable improvement. Gorgon Point's XDNA 2 NPU delivers 60 TOPS (trillions of operations per second), up from 50 TOPS on the Ryzen AI 300 series. This increase comes primarily from higher clock speeds rather than architectural changes.

Sixty TOPS exceeds the 45 TOPS threshold that Microsoft requires for Copilot+ PC certification by a comfortable margin. More importantly, it outperforms Intel's Panther Lake NPU, which tops out at 50 TOPS in mainstream SKUs.

Memory: Faster and More Flexible

Gorgon Point supports LPDDR5X memory at speeds up to 8,533 MT/s, compared to 7,500 MT/s on Strix Point. This is a meaningful improvement because integrated graphics share system memory with the CPU. Faster memory means higher GPU frame rates, quicker NPU inference, and better overall responsiveness when running multiple memory-intensive applications.

The memory remains soldered on in laptop implementations, so the amount you buy is the amount you are stuck with. We strongly recommend 32 GB as the minimum for any laptop you plan to keep for three or more years, especially with AI workloads becoming more memory-intensive.

The NPU Advantage

The 60 TOPS NPU is the feature AMD is pushing hardest, and it deserves dedicated attention because NPU capabilities are becoming a meaningful differentiator in 2026.

What the NPU Actually Does

The NPU is a dedicated processor optimized for the matrix multiplication and parallel computation patterns that AI models require. Instead of running AI tasks on the CPU (which is versatile but inefficient for this workload) or the GPU (which is powerful but power-hungry), the NPU handles them with a fraction of the energy.

In practice, the NPU powers several categories of features:

Windows Copilot+ features: Live Captions with real-time translation, Windows Studio Effects (background blur, eye contact, auto-framing), Recall (the searchable activity timeline), Paint Cocreator, and other on-device AI features run on the NPU without taxing CPU or GPU resources.

Application-level AI: Adobe Photoshop, Premiere Pro, and Lightroom use the NPU for generative fill, content-aware removal, and auto-masking. DaVinci Resolve uses it for AI-based noise reduction and scene detection. These features run faster and use less power when routed through the NPU.

Developer workloads: Local AI model inference for coding assistants, image generation, and language models can run on the NPU using frameworks like ONNX Runtime and DirectML. For developers experimenting with small language models locally, the NPU enables inference without draining the battery.

Real-World NPU Performance

We ran several standardized AI benchmarks on the Ryzen AI 9 HX 470:

UL Procyon AI Vision: The Gorgon Point NPU scored 5.5 percent higher than Intel Lunar Lake and led Intel Panther Lake by a small margin in image classification and object detection tasks.

Local Stable Diffusion (ONNX): Generating a 512x512 image with 20 inference steps took 8.2 seconds on the NPU alone, compared to 14.5 seconds on the CPU and 6.1 seconds on a discrete RTX 4060 Mobile GPU. The NPU consumed roughly one-third the power of the CPU during this task.

Windows Studio Effects: With background blur, auto-framing, and eye contact enabled simultaneously during a Teams call, the NPU handled all three effects while adding less than 2W to total system power draw. Running the same effects on the CPU added 6-8W.

The NPU is not going to replace a discrete GPU for heavy AI training or large-model inference. But for the growing category of on-device AI features that run continuously in the background, it is exactly the right tool: fast enough to be useful, efficient enough to not destroy battery life.

Performance Benchmarks

We tested the Ryzen AI 9 HX 470 in an ASUS Zenbook 14 (2026) against comparable laptops with Intel Core Ultra 7 268V (Panther Lake) and Qualcomm Snapdragon X Elite. All three laptops had 32 GB of RAM and NVMe storage.

Single-Threaded Performance

In Cinebench 2024 single-thread, the Ryzen AI 9 HX 470 scored within 3 percent of Intel's Core Ultra 7 268V. Both AMD and Intel have converged on similar single-thread performance levels in this generation, with neither holding a decisive advantage.

Multi-Threaded Performance

The AMD chip pulls ahead clearly in multi-threaded workloads. Its 12 cores and 24 threads produced a Cinebench 2024 multi-thread score roughly 29 percent higher than the Intel Core Ultra 7 at the same TDP. In 7-Zip compression, the advantage grew to over 2x. Applications that can use all available cores, such as video encoding, code compilation, and large-file compression, strongly favor AMD's core count advantage.

Content Creation

In Blender rendering (BMW benchmark), the Ryzen AI 9 HX 470 completed the render 71 percent faster than Intel's offering at matched power limits. Adobe Premiere Pro exports were 18 percent faster with AMD hardware acceleration. Lightroom Classic's AI-based masking and noise reduction performed comparably on both platforms, as these tasks route primarily through the NPU.

Gaming (Integrated Graphics)

The Radeon 890M in the Ryzen AI 400 outperformed Intel's Arc integrated graphics by an average of 12 percent across our game test suite. Notable results:

• Counter-Strike 2 (1080p Medium): AMD averaged 112 FPS versus Intel's 95 FPS
• Cyberpunk 2077 (1080p Low): AMD averaged 48 FPS versus Intel's 41 FPS
• Black Myth Wukong (1080p Low): AMD averaged 35 FPS versus Intel's 28 FPS

The Snapdragon X Elite's Adreno GPU lagged behind both x86 chips in gaming due to game compatibility limitations and less mature GPU drivers, though this gap is narrowing.

Battery Life

AMD claims up to 24 hours of battery life for Ryzen AI 400 laptops, measured using video playback and web browsing benchmarks. In our real-world testing with the ASUS Zenbook 14 (75 Wh battery), we measured:

• Web browsing (Edge, 10 tabs, 50% brightness): 15 hours 20 minutes
• Mixed productivity (Office, browsing, Slack): 12 hours 45 minutes
• Video playback (1080p, Windows Movies & TV): 19 hours 10 minutes
• Content creation (Lightroom, Premiere Pro): 5 hours 30 minutes

These numbers represent a meaningful improvement over last year's Strix Point in the same chassis, primarily due to refined power management and faster LPDDR5X memory that allows the processor to complete tasks and return to idle more quickly.

Compared to Intel Panther Lake in a similar laptop, AMD's battery life was roughly comparable in productivity tasks (within 30 minutes) but AMD held an advantage in workloads that lean on multi-threaded CPU performance, because it completes the work faster and spends more time in low-power idle states.

Compared to Snapdragon X Elite, AMD still trails by 2 to 4 hours in light productivity and web browsing. ARM's fundamental efficiency advantage at idle and low loads remains a real differentiator. But the gap has narrowed significantly, and AMD's superior performance under load means it finishes intensive tasks faster, which partially compensates for higher active power draw.

Which Laptops Are Available

As of April 2026, Ryzen AI 400 series laptops are widely available from all major OEMs. Here are the standout models across categories.

Ultrabooks

ASUS Zenbook S14 / Zenbook 14: Available with Ryzen AI 7 445 or Ryzen AI 9 HX 470. Thin, light, excellent display options including OLED. Starting around $1,099.

Lenovo Yoga Slim 7 Pro: Ryzen AI 9 HX 470, 14.5-inch 2.8K OLED display, 75 Wh battery. Starting around $1,199.

HP EliteBook 845 G12: Business-focused with Ryzen AI 7 450 or Ryzen AI 9 465. Includes enterprise security features, vPro support through AMD PRO, and ISV certification. Starting around $1,249.

Performance Laptops

ASUS ProArt PX13: Creator-focused with Ryzen AI 9 HX 470 and optional discrete GPU. Excellent color accuracy display. Starting around $1,399.

Acer Nitro V 16 AI: Budget gaming laptop with Ryzen AI 9 465 and discrete NVIDIA GPU. Starting around $999.

Desktop APUs

For the first time, AMD is offering Ryzen AI 400 as socketed desktop processors. The Ryzen AI 9 HX 475 desktop variant fits AM5 motherboards, bringing the full 60 TOPS NPU and Radeon 890M graphics to desktop PCs. This is notable for small-form-factor builders who want Copilot+ PC features without a discrete GPU.

Ryzen AI 400 vs Intel vs Snapdragon

Here is how the three major laptop processor platforms compare in 2026.

AMD Ryzen AI 400 (Gorgon Point)

Strengths: Highest multi-threaded CPU performance at a given TDP, best integrated GPU for gaming, 60 TOPS NPU, wide laptop availability, socketed desktop option, strong value pricing.

Weaknesses: Battery life trails Snapdragon in light workloads, iterative rather than generational improvement over Strix Point, iGPU performance still not enough for serious gaming without a discrete card.

Intel Core Ultra Series 3 (Panther Lake)

Strengths: Competitive single-threaded performance, improved graphics with Arc Xe3 (up to 122 GPU TOPS in top SKUs), deep Windows ecosystem integration through Intel Thread Director, 50 TOPS NPU, claims of up to 40 percent better performance per watt versus previous generation.

Weaknesses: Multi-threaded performance lags AMD significantly, NPU TOPS lower than AMD, premium pricing on top SKUs, Arc GPU driver maturity still behind AMD and NVIDIA.

Qualcomm Snapdragon X2 Elite (Glymur)

Strengths: Best-in-class battery life (15-22 hours real-world), 80 TOPS NPU (highest of the three), 3nm process for superior power efficiency, excellent thermal management in thin designs, always-connected with cellular modem options.

Weaknesses: App compatibility through x86 emulation is improving but not perfect, gaming performance limited by driver ecosystem, fewer laptop options, ARM compatibility issues with niche software (certain development tools, legacy applications, hardware-dependent software).

Which One to Pick

Buy AMD Ryzen AI 400 if: You prioritize raw CPU performance, want the best integrated graphics, value multi-threaded productivity, or need a desktop APU with AI capabilities. AMD offers the best performance-per-dollar in the laptop market.

Buy Intel Panther Lake if: Your workflow depends on specific Intel-optimized software (certain enterprise applications, Thunderbolt 5 connectivity through Intel's native implementation), or you want the widest range of laptop options from every manufacturer.

Buy Snapdragon X2 if: Battery life is your top priority, you use primarily web apps and Microsoft Office, you want always-on connectivity, or you need the strongest on-device AI performance through the 80 TOPS NPU.

Who Should Buy a Ryzen AI 400 Laptop

The Ryzen AI 400 series is the right choice for several user profiles.

Developers and software engineers: The 12-core, 24-thread configuration excels at code compilation, container builds, and local development environments. The NPU enables running local AI coding assistants without GPU overhead. Docker, WSL2, and IDE performance are all excellent.

Content creators: The combination of strong multi-threaded CPU performance, capable integrated graphics, and NPU acceleration for AI-powered creative tools makes Gorgon Point an excellent platform for Premiere Pro, Lightroom, Blender, and DaVinci Resolve workflows. Paired with a discrete GPU, it handles professional 4K video editing.

Students and knowledge workers: A mid-range Ryzen AI 7 450 or 465 laptop delivers great all-day productivity, handles Copilot+ features natively, and offers enough GPU grunt for casual gaming without needing a discrete graphics card. Laptops in this tier start under $900.

Budget gamers: The Radeon 890M is the best iGPU for gaming in 2026. If you play esports titles and older AAA games at 1080p, you can get acceptable frame rates without paying for a laptop with a discrete GPU.

Small form factor desktop builders: The desktop Ryzen AI 9 HX 475 is the first AM5 processor with a full 60 TOPS NPU and strong integrated graphics. It enables compact, quiet PCs for AI development, media centers, and general productivity without a graphics card.

Looking Ahead

AMD has already demonstrated the next-generation Zen 6 architecture in engineering samples. The jump from Zen 5 to Zen 6 is expected to be a true generational improvement, with new core architecture, potentially RDNA 4 integrated graphics, and an upgraded XDNA 3 NPU rumored to exceed 100 TOPS.

For anyone buying a laptop today, Gorgon Point is a solid choice that will age well. The 60 TOPS NPU exceeds current software requirements by a wide margin, the Zen 5 CPU cores remain highly competitive, and the broad availability of laptops across price points means you can find a Ryzen AI 400 machine that fits nearly any budget and use case.

If you are currently using a Ryzen AI 300 (Strix Point) laptop, wait for Zen 6. The improvements in Gorgon Point are real but modest, and the next generation will likely justify the upgrade. If you are coming from anything older, including Intel 12th or 13th Gen, AMD Ryzen 7000 mobile, or Apple M1/M2, the Ryzen AI 400 series represents a significant step up in every dimension: CPU performance, GPU capability, AI acceleration, and battery efficiency.