Microsoft signs on to rack AI system that rivals Nvidia

After a decade-long turnaround, chip giant Advanced Micro Devices is preparing to ship its first rack-scale AI system, called Helios, to a growing list of customers that now includes Microsoft.

It’s the first rival to Nvidia’s hugely popular Grace Blackwell and Vera Rubin systems, and aims to give the world’s most valuable chipmaker its first real competition in years.

Microsoft announced on Monday that it will use the Helios system in data centers. Meta, OpenAIOracle and others are racing to get as much computing as possible.

AMD will begin shipping to customers, including Microsoft, later this year.

Details regarding financial terms or the amount of computing capacity were not disclosed.

“We are expanding the Azure infrastructure portfolio with AMD Helios to give customers the performance, scale, and choice they need to build and run next-generation AI applications,” Microsoft CEO Satya Nadella wrote in a press release.

The new Helios system will power edge model inference and support Azure AI services for Microsoft and its AI customers. Microsoft will also add two new computing instances running on AMD’s latest “Venice” central processing units, or CPUs; one for artificial intelligence and data pipelines, and the other for semiconductor design.

This continues a long-standing partnership with AMD chips, which have been powering Microsoft’s Surface PCs and Xbox gaming consoles for many years. In 2023, Microsoft became the first company to adopt AMD’s MI300X graphics processing unit, or GPU, which rivals Nvidia’s AI chips. Microsoft also uses its own Maia chips in its data centers.

Like its peers, Microsoft needs as much computing as possible, especially as it accelerates its own model development and devotes more computing capacity to research and development. In June, it announced seven models produced in-house. Microsoft’s AI efforts have had mixed results so far, from the 365 Copilot AI assistant to the GitHub Copilot coding agent. It has been the worst-performing “Magnificent Seven” stock so far this year.

Microsoft is part of a growing number of large companies turning to AMD for AI acceleration. AMD says eight of the top 10 AI companies are running workloads on Instinct GPUs, including OpenAI, Cohere, and Elon Musk’s SpaceXAI, which is part of SpaceX.

Meta in February announced that it will use up to 6 gigawatt AMD GPUs over time; 1 gigawatt will be deployed in Helios racks later this year. OpenAI and Oracle have also made major commitments to deploy Helios with India this year. largest IT company Tata Consultancy ServicesI promise to use it too.

CNBC got the world’s first detailed look inside a Helios system from the Texas data center laboratory where it was developed and tested.

‘Lowest cost per token’

Named after the ancient Greek god who pulled the sun across the sky with the help of four horses, Helios brings together four things AMD does in-house: GPUs, CPUs, networking and software.

“We’re very focused on providing the best total cost of ownership, the lowest cost per token,” data center president Forrest Norrod told CNBC about AMD’s first-generation system. “And our customers tell us we’ve succeeded.”

In May, AMD CEO Lisa Su told CNBC’s Jim Cramer that Helios has “significant advantages” over Nvidia’s rack-scale systems “when we’re talking about inference and talking about memory bandwidth and memory capabilities.”

While AMD did not comment on the cost, Futurum Group estimates that Helios will cost between $5 and $5.5 million. That compares with Futurum’s estimates of $3.5 million to $4 million for Nvidia’s second-generation rack-scale system, Vera Rubin.

Weighing up to 7,000 pounds, the Helios is larger and heavier than Nvidia’s Vera Rubin.

According to Futurum Group, Nvidia controls more than 95% of the data center GPU market. AMD only holds 4.5% of the market, but Helios could change that.

“I think there’s a serious situation where AMD is doing excellent and can get to 20% and 25%. That’s hundreds of billions of dollars in revenue, by the way,” said Daniel Newman, analyst and CEO of Futurum Group.

Data centers accounted for the majority of AMD’s revenue in the first quarter of 2026, up 57% from the previous year. AMD told CNBC it plans to generate tens of billions of dollars in data center AI revenue starting in 2027, mostly from Helios.

In data center CPU market share, Intel It remains the clear leader, but AMD is steadily gaining ground. This CPU lead differentiates AMD from Nvidia, which launched its first server CPU in 2021 and shifted strategies to renew focus on chips this year.

‘A very different AMD’

Norrod called Helios “our baby” while showing CNBC the system’s core chips. Each of the 18 compute trays features four Instinct GPUs powered by a single EPYC central processing unit.

It was these EPYC data center CPUs that helped AMD regain a decade-long lead in the data center market.

In 2003, AMD had a groundbreaking data center CPU. quickly capture almost a quarter of the marketbut that slice faded away After a series of delays and missteps that led to major consequences Layoffs and reduced income When water takes the helm.

“It has been a very different AMD under Lisa’s leadership for the last 12 years,” Norrod said.

Things turned around after the company brought its first EPYC server CPU onto the scene in 2017.

“One of the things we did was lay out our roadmap in detail for three generations, which is very unusual,” he said. “And we delivered exactly what we said.”

Part of AMD’s roadmap included plans to launch Helios with the existing MI400 series GPU.

Each Helios tray also includes up to 12 networking chips built with technology AMD acquired when it acquired Pensando in 2022.

This was one of several acquisitions that helped develop Helios over the past few years.

AMD’s biggest acquisition to date is programmable chip company Xilinx for around $50 billion in 2022. AMD also acquired server maker ZT systems for about $5 billion in 2025 and a number of software companies that helped Nvidia develop ROCm, its open-source alternative to the widely adopted CUDA software ecosystem.

Counterpoint Research analyst Neil Shah said AMD’s Helios chips are “on par” with Nvidia GPUs and CPUs, but “the secret sauce is in the software and optimization.”

“With CUDA, I think Nvidia has a bigger ecosystem and is well ahead of AMD,” he said.

With Helios, AMD has the opportunity to make significant progress based on the success of early deployments.

“The question will be: Does AMD win because it is technologically superior? Or does it win because there is such a constraint on capacity that if they can make it, someone will buy it?” Newman said.

WRISTWATCH: First look at Vera Rubin, Nvidia’s AI system 10 times more efficient than Grace Blackwell

— CNBC’s Jordan Novet contributed to this report.