AMD's Helios AI Servers Enter Full Production

Detailed close-up image of NVIDIA RTX 2080 graphics card showcasing hardware components.

Advanced Micro Devices' second-generation AI servers are in full production and will begin shipping in the coming months, CEO Lisa Su said during the AMD Advancing AI 2026 conference in San Francisco on July 23, 2026. The Helios systems combine AMD's MI455X AI accelerator with its new Venice central processor, both manufactured by Taiwan Semiconductor Manufacturing Co. Su said shipments would start at the end of the third quarter and characterized customer demand as extremely strong. She framed the product as a direct effort to capture market share from Nvidia in the fast-growing data center chip sector, particularly for inference workloads used by chatbots such as OpenAI's ChatGPT.

OpenAI and Anthropic Positioned as Anchor Customers

During Su's keynote, OpenAI's vice president of compute strategy Sachin Katti said the ChatGPT maker plans to deploy Helios at massive scale starting toward the end of 2026, with deployment accelerating through 2027. OpenAI also intends to use AMD's next-generation MI500 AI chips. The comments follow an October multiyear agreement under which OpenAI could bring in tens of billions of dollars in annual revenue for AMD while securing the option to acquire up to roughly 10% of the chipmaker. Separately, AMD announced plans to sell up to two gigawatts of its Instinct MI450 chips to AI lab Anthropic beginning in the first half of 2027, a deal that includes an investment of as much as $5 billion in the Claude maker. Anthropic co-founder Tom Brown also joined Su on stage during the event.

Market Outlook and Strategic Partnerships

Su projected the total computing market would reach $2 trillion by 2030, with $1.4 trillion from AI-accelerator chips and $220 billion from central processing units, an area where AMD has long competed against Intel and that Nvidia has entered. AMD estimated the total computing market was worth approximately $365 billion in 2025. Su also announced a partnership with Cerebras Systems under which the two companies will offer a combination of their AI servers, initially deployed in Cerebras data centers. Following the announcements, Cerebras shares rose about 5% and AMD shares fell roughly 2% by the close, with one analyst attributing AMD's decline to broader weakness in U.S. equities and a semiconductor-sector selloff.

NVIDIA Reportedly Raises RTX Memory Kit Prices

Separately, reporting on July 23, 2026 indicated that NVIDIA has raised prices for GDDR6 and GDDR7 memory kits used in its RTX GPUs. According to TechPowerUp and Wccftech, the increases affect memory components supplied across NVIDIA's RTX product line. The reports come as AMD prepared to challenge Nvidia's data center position with its Helios platform.

Industrial GPU Adapted for Desktop Use

In a separate development covered on July 23, 2026, Hackaday reported on efforts to adapt an industrial GPU for desktop use. The piece detailed modifications intended to bring workstation-class or specialized GPU capabilities to consumer-grade systems.

Share this article

FacebookX

9 sources

Sources