Key Highlights
- AMD unveiled its Helios rack-scale AI platform Thursday at a San Francisco event, aiming squarely at Nvidia’s data center market leadership.
- The Helios system integrates Instinct MI455X GPUs, EPYC Venice CPUs, Pensando networking technology, and ROCm software into a unified solution.
- The company secured a partnership with Anthropic for deploying up to 2 gigawatts of Instinct MI450 GPUs, accompanied by AMD’s $5 billion investment in Anthropic.
- Microsoft plans to integrate Helios into Azure for advanced model inference tasks, with deliveries expected to commence in late 2026.
- While AMD commands approximately 4.5% of the data center GPU sector compared to Nvidia’s commanding 95%-plus share, its data center division saw 57% year-over-year growth, reaching $5.78 billion in Q1 2026.
Advanced Micro Devices is launching its most ambitious offensive yet in the AI infrastructure space, unveiling cutting-edge hardware alongside securing high-profile customer agreements — though Nvidia’s market position remains formidable.
Shares of AMD gained approximately 1.45% Thursday as the semiconductor company commenced its AI hardware showcase at Moscone West in San Francisco.
Advanced Micro Devices, Inc., AMD
At the heart of the announcement is Helios, AMD’s inaugural rack-scale AI infrastructure solution. The system integrates Instinct MI455X GPUs, EPYC Venice CPUs, Pensando networking components, and ROCm software into a comprehensive platform engineered for AI training and inference applications.
Helios represents AMD’s strategic response to Nvidia’s integrated rack-scale offerings. Nvidia currently commands more than 95% of the data center GPU marketplace. AMD holds roughly 4.5%, based on Futurum Group data reported by CNBC.
AMD is simultaneously introducing the Venice CPU for data center applications during the event — a processor that has been under development for an extended period.
The product debut comes on the heels of several partnership revelations earlier in the week that provided AMD with substantial commercial momentum entering Thursday’s presentation.
Anthropic Collaboration Details
Wednesday brought news of AMD’s collaboration with Anthropic to implement up to 2 gigawatts of Instinct MI450 Series GPUs within Helios infrastructure. The initial gigawatt deployment is scheduled to launch during the first half of 2027.
As part of this arrangement, AMD is committing a strategic equity investment reaching up to $5 billion in Anthropic.
The companies will engage in an extended engineering partnership spanning multiple years. The objective centers on leveraging Anthropic’s Claude AI to enhance workload performance on AMD Instinct GPUs and accelerate refinement of AMD’s ROCm software platform.
Additionally, AMD will implement Claude throughout its internal engineering and product development operations.
Tom Brown, co-founder and chief compute officer at Anthropic, stated: “By partnering with AMD across the stack, we are securing the capacity we need and optimizing it for training and serving Claude.”
Microsoft Collaboration
Earlier in the week, AMD broadened its collaboration with Microsoft. The agreement calls for Microsoft to implement Helios on Azure to power frontier model inference operations.
AMD anticipates beginning Helios shipments to clients — including Microsoft — during the latter half of 2026.
In October 2025, AMD also revealed a multiyear agreement with OpenAI potentially generating tens of billions in annual revenue. That arrangement provides OpenAI with the opportunity to acquire approximately 10% of AMD.
AMD’s data center business generated $5.78 billion in revenue during Q1 2026, representing a 57% increase compared to the equivalent period in the previous year. The expansion was fueled by robust demand for EPYC server processors and Instinct GPU deliveries.
CEO Lisa Su commented that the Anthropic partnership “brings together Anthropic’s leadership in frontier AI with the full strength of AMD high-performance computing.”
AMD has now secured confirmed customers in Anthropic, Microsoft, and OpenAI for its next-generation AI technology — a customer roster that represents a significant escalation in its competitive stance against Nvidia.



