Advanced Micro Devices (AMD) scored a major victory in the artificial intelligence arms race on Monday, announcing that Microsoft (MSFT) will deploy its new Helios rack-scale AI system across the Azure cloud platform. The expanded partnership sent AMD shares up nearly 5%, with investors betting the deal could finally crack Nvidia’s (NVDA) iron grip on the data center GPU market.

Microsoft plans to use Helios to power frontier AI model inference, Azure AI services, and customer workloads. AMD said it will begin shipping the system to Microsoft and other customers in the second half of 2026. The two companies did not disclose the financial terms or the size of the compute order.

“Through our collaboration with AMD, we are expanding the Azure infrastructure portfolio with AMD Helios to give customers the performance, scale and choice they need to build and run the next generation of AI applications,” said Microsoft Chairman and CEO Satya Nadella.

The announcement marks the first time Microsoft has committed to AMD’s complete rack-scale AI platform, putting Helios in direct competition with Nvidia’s Grace Blackwell and Vera Rubin systems. AMD CEO Lisa Su called the deployment “an important milestone” and said the companies are extending a long-running collaboration to scale the next wave of AI infrastructure.

What’s Inside Helios

Helios is AMD’s first integrated rack-scale AI system, a departure from the company’s traditional approach of selling GPUs and CPUs separately. The platform combines four key components into a single, turnkey solution designed for large-scale AI training and inference.

The system packs AMD’s Instinct MI455X GPUs alongside sixth-generation EPYC “Venice” central processors. Each compute tray houses four Instinct GPUs driven by a single EPYC CPU, plus up to 12 networking chips based on technology AMD acquired through its 2022 purchase of Pensando. The entire stack runs on AMD’s open-source ROCm software platform.

Forrest Norrod, who heads AMD’s data center business, said the core value proposition is straightforward. “We developed Helios to deliver the best total cost of ownership and the lowest cost per token,” Norrod said. “Customer feedback tells us we’ve hit that target.”

Lisa Su has previously claimed Helios carries “significant advantages” over Nvidia’s rack-scale offerings in inference performance, memory bandwidth, and memory capabilities.

Research firm Futurum Group estimates Helios will cost between $5 million and $5.5 million per rack. That’s notably higher than the $3.5 million to $4 million price tag Futurum projects for Nvidia’s second-generation Vera Rubin system. AMD has not publicly discussed pricing.

Beyond the GPU: A Broader Azure Footprint

The partnership extends well beyond Helios. Microsoft will introduce two new virtual machine series powered by sixth-generation AMD EPYC Venice processors. One line targets agentic AI and data pipeline workloads, while the other is designed for semiconductor design applications.

Microsoft is also expanding its use of AMD’s Pensando Data Processing Units across Azure’s network backbone. The company plans to integrate AMD technology with its own Azure Boost acceleration platform to improve networking performance and efficiency throughout its global cloud infrastructure.

The deal deepens a partnership that stretches back years. AMD chips have long powered Microsoft’s Surface laptops and Xbox gaming consoles. In 2023, Microsoft became one of the first major cloud providers to adopt AMD’s MI300X GPUs for AI workloads. Microsoft also develops its own custom Maia AI chips for data center deployment.

AMD’s Growing Customer Roster

Microsoft joins an increasingly impressive list of Helios adopters. Meta (META) has committed to deploying up to six gigawatts of AMD GPU capacity over time, with the first gigawatt running on Helios racks. OpenAI, Oracle (ORCL), and India’s Tata Consultancy Services have also made significant deployment commitments.

AMD says eight of the world’s top 10 AI companies now run workloads on its Instinct GPUs. The company reported first-quarter 2026 data center revenue of $5.78 billion, up 57% year-over-year, with total revenue reaching $10.25 billion. AMD expects second-quarter revenue of approximately $11.2 billion, representing roughly 46% annual growth.

Wall Street has taken notice. KeyBanc maintained an Overweight rating and $725 price target on AMD shares, citing strong demand for server processors and AI. William Blair initiated coverage with a Market Perform rating, pointing to the company’s expanding potential in AI infrastructure.

AMD stock closed Monday around $512, up from a previous close of $495.76. The shares have rallied roughly 139% year-to-date. The average analyst price target sits at $541.31, according to TipRanks data, with the consensus rating at Strong Buy based on 28 Buy and eight Hold ratings over the past three months.

The Nvidia Question

Despite the momentum, AMD faces a steep climb. Nvidia still controls more than 95% of the data center GPU market, leaving AMD with approximately 4.5%. Futurum Group analyst Daniel Newman believes Helios could change that dramatically.

“I believe there is a significant case where AMD performs exceptionally well and can reach 20% to 25% market share,” Newman said. “This is hundreds of billions of dollars in revenue, by the way.”

Counterpoint Research analyst Neil Shah cautioned that hardware parity isn’t enough. “AMD’s Helios chips are comparable to Nvidia’s GPUs and CPUs in performance,” Shah said. “But the key lies in software and optimization.” Nvidia’s CUDA platform remains the dominant software ecosystem for AI development, a moat AMD’s ROCm platform has yet to breach.

Newman raised another fundamental question: “Is AMD winning because of technology leadership, or simply because compute supply is so tight that anyone who can build chips finds buyers?” The answer, he said, will depend heavily on how Helios performs in early real-world deployments.

AMD’s Advancing AI event, scheduled for July 22 and 23, could provide additional catalysts. Jefferies analysts have suggested that Anthropic might be the next major customer to announce an AMD partnership. Lisa Su has said customer interest in the upcoming MI450 chips and Helios platform is running stronger than the company originally expected.

The company plans to generate tens of billions of dollars in data center AI revenue starting in 2027, with the bulk expected to come from Helios. Whether AMD can translate early customer wins into sustained market share gains remains the defining question for investors — and for Nvidia.