Ask Runable forDesign-Driven General AI AgentTry Runable For Free
Runable
Back to Blog
Technology7 min read

AMD's rival to Nvidia's GB10 AI workstation is set to go on preorder in days, but is it too little too late? | TechRadar

Preorders for AMD's attempt to compete with Nvidia's DGX Spark open in June 2026 Discover insights about amd's rival to nvidia's gb10 ai workstation is set to g

TechnologyInnovationBest PracticesGuideTutorial
AMD's rival to Nvidia's GB10 AI workstation is set to go on preorder in days, but is it too little too late? | TechRadar
Listen to Article
0:00
0:00
0:00

AMD's rival to Nvidia's GB10 AI workstation is set to go on preorder in days, but is it too little too late? | Tech Radar

Overview

News, deals, reviews, guides and more on the newest computing gadgets

Start exploring exclusive deals, expert advice and more

Details

Unlock and manage exclusive Techradar member rewards.

Unlock instant access to exclusive member features.

Get full access to premium articles, exclusive features and a growing list of member rewards.

AMD's rival to Nvidia's GB10 AI workstation is set to go on preorder in days, but is it too little too late?

Preorders for AMD's attempt to compete with Nvidia's DGX Spark open in June 2026

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

The AMD Ryzen AI Halo is powered by AMD's 16-Core 'Zen 5'-based Ryzen AI Max+ 395

It also offers 128GB of unified memory, matching that of the Nvidia DGX Spark

With an advertised MSRP of $3999, it aims to challenge both Nvidia's offerings and solutions that currently rely on a higher-end Apple Mac Mini for localized AI

AMD is finally set to release its Ryzen AI Halo, a compact AI development full-stack offering that aims to compete directly with Nvidia's DGX platform and Apple's Mac Mini.

The Platform offers a memory configuration similar to the former and allows developers to install Windows or Linux as their operating system of choice.

But AMD's entry comes almost two years after the entries of it's rivals, prompting the question: is it too little too late?

The AMD Ryzen AI Halo Developer Platform was announced by CEO Dr. Lisa Su over 4 months ago, during her CES keynote. It finally received a tentative release date, with preorders (exclusive to Microcenter in the US) beginning this June.

AMD's answer to Nvidia's dominance of the current AI developer market, however, will arrive nearly 5 months after its announcement, even as other similarly configured enterprise-class options already exist to fill the void.

AMD's new Ryzen AI 400 CPUs prioritize AI skills over GPU power

Tiny company steals AMD's thunder and challenges Nvidia with old-tech PCIe AI accelerator

Beelink's SER10 MAX mini PC, boastingthe insanely powerful AMD Ryzen AI 9 HX 470 and 32GB of DDR5, is $300 off - so act now

The specifications for the AMD Ryzen AI Halo Developer Platform (via AMD's Press Deck) (Image credit: AMD)

AMD has not directly competed with Nvidia's DGX Spark since its release in mid-October 2025, even as it has leveraged 3rd-party industrial partners such as HP (Z2 Mini G1a) and Minisforum (MS-S1 Max) to fill the void.

Despite its late launch, the Ryzen AI Halo aims to level the playing field for Team Red against Nvidia's deep-rooted and formidable offering.

While AMD's own slides show a slight lead when it comes to certain AI models, with leads ranging from 4% to 14% in terms of token generation, while touting support for Windows, a lower power cost per token generated for leading LLMs, and an NPU under the hood, things are considerably closer than one might like, all things considered.

Press slides from AMD comparing its Ryzen AI Halo to Nvidia's DGX Spark (Image credit: AMD)

AMD's onboard AMD Ryzen AI Max+ 395 has a TDP of 120W vs the Nvidia GB10 onboard the Spark (140W), even as both units come with a 240W PSU to cover the overhead of their motherboards, SSDs, and cooling.

Both have the same amount of RAM available to users (128GB), and while AMD's offering comes with Wi-Fi 7 and Bluetooth 5.4 (like the DGX Spark), it only offers a 10 Gb E ethernet port versus Nvidia's Connect X NIC that offers up to 200 Gbps speeds, making it ideal for users to hook up 2 DGX Sparks together to work with twice the parameters locally versus AMD's offering.

The DGX Spark also holds its own against AMD's offering in terms of raw compute, delivering up to 1 peta FLOP of FP4 compute, compared to AMD's advertised 60 TFLOPs of FP16. It also offers a 4TB configuration for an on-site price of $4700, doubling the storage that AMD currently offers.

Physically, the Ryzen AI Halo is slightly smaller than the DGX Spark, even as both units externalize their power supply requirements to maintain their respective form factors.

Minisforum’s MS-S1 Max for AI workloads and demanding professional tasks is $760 off

MSI (re)launches $85,000 Nvidia DGX Station workstation

Nvidia's new MFG is coming, but Intel's equivalent has a hardware advantage

One could argue that the AMD Ryzen AI Max+ 395 that the Ryzen AI Halo currently offers is a bit underwhelming for a release this late in the game.

HP has already been delivering the Z2 Mini G1a to most of its enterprise-class customers, offering a similar configuration: 128GB of unified memory, the same APU, and 2TB of storage, albeit in a much larger form factor than what AMD is currently offering.

Despite this, AMD's offering aims to broaden the options available to users who were previously looking for a smaller form factor, access to a validated tech stack, or simply the ability to purchase a solution with broader support than similar offerings off the shelf.

One thing is clear, however, for users wanting to run CUDA applications or larger models ( > 200B parameters): Nvidia's DGX Spark and its superior networking still pack a formidable punch that AMD has yet to provide an alternate for, at least in this form factor.

Follow Tech Radar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

Having built hundreds of gaming PCs and being an avid gamer in his spare time, Rahim tends to have stronger opinions about hardware than most. This is particularly on display when he gets his way with powerful, but minimalistic RGB builds even as Small Form Factor (SFF) PCs come a close second.

You must confirm your public display name before commenting

1 Could Microsoft kill the web browser at Build? Everything developers need to know about the NLWeb Protocol

2'Just imagine what could get done' — How this US startup is building a 'cheap' fab-in-a-box to do for microchips what IBM did for PCs

3'Our ability to achieve orbital AI at scale depends on our ability to access a sufficient number of AI chips, significantly more than are currently available to us': Space X admits that getting data centers in space may fall short due to a lack of chips

4 Want great coffee, fast? I'm a trained barista, and these are my top 4 capsule coffee makers from Nespresso, Keurig, and more

5'Built around volcanic materials' — Turkish startup wants to make billion-dollar radar systems near obsolete with cheap spray-in-a-can tech that claims to bring stealth capabilities to low-cost drones

Tech Radar is part of Future US Inc, an international media group and leading digital publisher. Visit our corporate site.

© Future US, Inc. Full 7th Floor, 130 West 42nd Street, New York, NY 10036.

Key Takeaways

  • News, deals, reviews, guides and more on the newest computing gadgets
  • Start exploring exclusive deals, expert advice and more
  • Unlock and manage exclusive Techradar member rewards
  • Unlock instant access to exclusive member features
  • Get full access to premium articles, exclusive features and a growing list of member rewards

Cut Costs with Runable

Cost savings are based on average monthly price per user for each app.

Which apps do you use?

Apps to replace

ChatGPTChatGPT
$20 / month
LovableLovable
$25 / month
Gamma AIGamma AI
$25 / month
HiggsFieldHiggsField
$49 / month
Leonardo AILeonardo AI
$12 / month
TOTAL$131 / month

Runable price = $9 / month

Saves $122 / month

Runable can save upto $1464 per year compared to the non-enterprise price of your apps.