AMD challenges Nvidia's DGX Spark with $3,999 Ryzen AI Halo with Windows 11 support — Strix Halo desktop undercuts Nvidia by $700, packs 128GB of unified memory

Dr. Lisa Su, President, and Chief Executive Officer of AMD announcing the Ryzen AI Halo developer platform at CES 2026. — (Image credit: AMD)

AMD's Ryzen AI Halo Developer Platform is finally here, with preorders for the powerful $3,999 mini PC now live in the U.S. through Micro Center. The company’s new developer platform is effectively an answer to Nvidia's DGX Spark, as both systems are designed as compact AI workstations capable of running large language models locally. First unveiled at CES 2026, the developer kit is available in two variants that feature identical hardware but ship with different operating systems- Linux or Windows 11 Pro.

Notably, the DGX Spark was originally introduced at the same $3,999 price point last year. However, Nvidia recently increased the price to $4,699 due to the ongoing global supply constraints affecting LPDDR5X memory and NAND flash. Apart from pricing, the Ryzen AI Halo Developer Platform offers native Windows support, whereas the DGX Spark is limited to Linux-based environments. This gives users greater flexibility, including the option to dual-boot and use both operating systems.

Features of the AMD Ryzen AI Halo developer kit — (Image credit: AMD)

The Ryzen AI Halo dev kit is powered by the flagship Ryzen AI Max+ 395 processor, featuring 16 cores and 32 threads, with a base clock speed of 3 GHz and a max boost clock speed of 5.1 GHz. It also features 128GB of unified LPDDR5X-8000 memory, 16MB of L2 cache, and 64MB of L3 cache, onboard RDNA 3.5 Radeon 8060S graphics with 40 compute units, and a dedicated NPU capable of 50 TOPS.

Latest Videos From

Watch full video here:

The mini-PC measures 149 x 149 x 43.18 mm, has an aluminum chassis, and includes a 2TB M.2 SSD, while wireless connectivity options include Wi-Fi 7 and Bluetooth 5.4. As for I/O, it is strikingly similar to what Nvidia offers, including a 10 GbE LAN, four USB-C ports (one for power input), and an HDMI 2.1b port. Unfortunately, there is no NVLink, which, in Nvidia’s case, allows two DGK Spark machines to pair. To cool the Strix Halo chip inside AMD, it uses a unique cooling solution that includes a baseplate, direct-touch flat heatpipes, an aluminum channel heatsink, and two lateral airflow blower-style fans.

If you are a developer or simply interested in experimenting with local LLMs, Micro Center is currently accepting preorders for the Ryzen AI Halo Developer Platform, with local pickup available on July 10. Those looking for alternatives can also consider the Corsair AI Workstation 300, a slightly larger AI-focused system based on the same Ryzen AI Max+ 395 processor. It is also currently listed at a lower starting price of $2,699 for the 1TB storage model and $3,399 for the 4TB model.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

TOPICS

Kunal Khullar is a contributing writer at Tom’s Hardware. He is a long time technology journalist and reviewer specializing in PC components and peripherals, and welcomes any and every question around building a PC.

4 Comments Comment from the forums

usertests

Corsair links are out of stock, but I think there are other 128 GB Strix Halo models out there that are coming in well under $4k, so it's a tough sell.

RDNA3.5 isn't as AI-ready as RDNA4/5. I think RDNA4 is already 2-8x faster per CU depending on the data type. So waiting (until 2028?) for an RDNA5-based Medusa Halo could be the right move if you're a dabbler and not a pro(sumer).
Reply
daviburg1979

The software support is an essential differentiator here. The latest llm models just work and reliably so on the DGX, while amd's stack is unreliable. All this hardware is pointless if you can't run your app on it.
Reply
dva852

RDNA3.5 isn't as AI-ready as RDNA4/5. I think RDNA4 is already 2-8x faster per CU depending on the data type.
For LLM use, 8060S' compute capability doesn't matter for decoding (token generation). The bottleneck is the memory bandwidth, which determines token/sec, and Strix is roughly equivalent to Spark. Spark is 2-3X faster than 8060S for context processing. But for normal context size of, say, 8-32K, time-to-first-token (TTFT) is a few seconds, and is trivial compared to decoding time. It's only with larger (64K-256K) context size that TTFT becomes more noticeable.

Reviews of DGX Spark at release all knocked the memory bandwidth bottleneck. But 256-bit LPDDR5X has been decided as "good enough" for current-gen local AI, likely for lack of anything faster that's cost-effective.

That'll change once LPDDR6 hits next year, with ~50% higher bandwidth. Once the local AI train gets moving, there'll be iterative increases down the line. We're still at the starting line.

Getting back to Spark & Strix Halo, the low memory bandwidth means running large models will be slow. A variety of factors come into play, such as sparsity (dense/MoE), backend (CUDA/Vulkan/ROCm), quant type (Q4_K_M/MXFP4/NVFP4), context length, etc. The upshot is that speed ranges from very slow to usable, but never blazingly fast.

Hence why Gorgon Halo's 192GB is of marginal benefit, as the bandwidth bottleneck is still the same. You can load larger model, but it'll just run slower. Larger context would be where 8060S' lack of compute will matter. IMO, Gorgon Halo is AMD treading water until Medusa Halo arrives.

Comparatively speaking, AMD is still ahead of Intel, which will likely join the local AI APU fray with Nova Lake AX in '27, and Qualcomm would also probably get involved. Fun times ahead.

The latest llm models just work and reliably so on the DGX, while amd's stack is unreliable.
ROCm 7.1/7.2 is now stable, and faster than Vulkan, but Vulkan is more robust. CUDA still wins on polish and ecosystem support, but AMD has come a long way. For the hobbyist/enthusiast, AMD is a viable option.

That said, AMD's $4K box isn't attractive against Nvidia $4.7K DGX Spark. The target audience (devs/biz) isn't price sensitive at these price points, and will opt for Spark over Strix.
Reply
usertests

dva852 said:
That'll change once LPDDR6 hits next year, with ~50% higher bandwidth. Once the local AI train gets moving, there'll be iterative increases down the line. We're still at the starting line.
I think it could be around ~100%, if you'll get a 384-bit bus instead of 256-bit (LPDDR6 uses 12-bit subchannels).

Very conservative would be (10667 MT/s / 8533 MT/s) * 1.5 = ~1.875. The entry-level LPDDR6 speed is 10667 MT/s according to JEDEC.

dva852 said:
Comparatively speaking, AMD is still ahead of Intel, which will likely join the local AI APU fray with Nova Lake AX in '27, and Qualcomm would also probably get involved. Fun times ahead.
Sweet.
Reply