AMD to Supply 'Helios' AI Rack to Microsoft, Accelerating Its Pursuit of Nvidia
Expanding Big Tech Clientele
Challenging Nvidia's Market Share
AMD has agreed to supply rack-scale artificial intelligence (AI) systems to Microsoft (MS). By expanding its customer base to include Meta, OpenAI, Oracle, and MS, AMD is accelerating its efforts to catch up in the AI hardware market, where Nvidia has maintained a dominant lead.
On July 20 (local time), AMD announced that Microsoft will introduce its AI rack system, 'Helios,' which integrates a graphics processing unit (GPU), central processing unit (CPU), networking, and software into a single cabinet, on its Azure cloud platform. Azure is Microsoft's cloud platform.
Helios is AMD's next-generation rack-scale platform for AI training and inference. AMD first previewed Helios in June of last year and unveiled the system at the Open Compute Project (OCP) Global Summit in October of the same year. At the 'CES 2026' information technology (IT) and electronics exhibition in January of this year, AMD showcased specific system configurations and performance metrics. This product is positioned to compete with Nvidia’s Grace Blackwell and next-generation Vera Rubin-based rack systems.
Lisa Su, AMD's Chair and Chief Executive Officer (CEO), stated, "AMD and Microsoft have been building high-performance infrastructure together for years," adding, "We are expanding the scope of collaboration across the entire stack of AMD AI solutions on Azure." She continued, "Microsoft's adoption of AMD products is a significant milestone as it provides Azure customers with industry-leading computing solutions and enables the joint expansion of next-generation AI infrastructure."
Satya Nadella, Microsoft's Chair and CEO, also said, "Customers are looking for AI infrastructure optimized for a broad range of workloads—from training and inference to data preparation, search, and reinforcement learning. By introducing Helios, we will provide customers with the performance, scalability, and wide array of choices needed to develop and run next-generation AI applications."
CNBC highlighted that Microsoft has joined the ranks of large enterprises increasingly adopting AMD products for AI computing infrastructure. According to AMD, 8 out of the top 10 AI companies are carrying out AI tasks on Instinct GPUs. This includes companies such as OpenAI, Cohere, and SpaceXAI under SpaceX.
In February of this year, Meta announced plans to deploy AMD GPUs on a long-term basis, scaling up to as much as 6GW. The company plans to first build a 1GW-scale system through Helios racks in the second half of this year. Both OpenAI and Oracle have signed large-scale contracts to adopt Helios this year, and Tata Consultancy Services, the largest IT company in India, has also pledged to use the system.
Hot Picks Today
End of Selling Spree in Sight: "Leverage Unwinding Nearing Completion, KOSPI Target at 12,500," Says JPMorgan
- "$300 Gone in One Date"...U.S. Gen Z Ditches Restaurants for Parks
- "It Took Three Hours Just to Exit the Parking Lot"... Vehicles Stuck at Asan Medical Center
- "Broke the Piggy Bank... Salaries Just Aren't Enough": The Stocks Individual Investors Are Rushing to Buy Worldwide
- Zara’s “Death Pants” Too Stylish to Give Up: “I Almost Died Just 30 Minutes After Going Out”
According to market research firm Futurum Group, Nvidia holds more than 95% of the data center GPU market, while AMD's share remains around 4.5%. However, there is analysis suggesting that Helios has the potential to change the market landscape. Daniel Newman, CEO and analyst at Futurum Group, stated, "It is entirely possible for AMD to expand its market share to 20–25% with Helios leading the way," adding, "This is a market with hundreds of billions of dollars in revenue opportunities at stake."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.
