AMD is making one of its most significant pushes into AI infrastructure, unveiling new hardware and announcing major customer partnerships as it seeks to challenge Nvidia’s dominant position in data center GPUs.

AMD stock was up roughly 1.45% on Thursday as the company began its AI hardware event at Moscone West in San Francisco. The central product launch is Helios, AMD’s first rack-scale AI system, which is designed to combine multiple components into one integrated platform for AI training and inference. The rack-scale approach reflects a broader industry shift as AI models grow larger and more complex, pushing hyperscalers and model developers to procure fully integrated systems rather than individual GPUs.

Helios brings together Instinct MI455X GPUs, EPYC Venice CPUs, Pensando networking chips, and AMD’s ROCm software stack. The system is AMD’s response to Nvidia’s rack-scale AI infrastructure offerings, which have helped Nvidia maintain a commanding position in the market. Nvidia’s CUDA software ecosystem has long been considered one of its strongest competitive moats, as developers have built years of tooling and optimization around it. AMD’s ROCm serves as its counterpart, and the company has been investing heavily to close the software gap, though it remains a key area of focus.

Nvidia currently holds more than 95% of the data center GPU market, while AMD has about 4.5%, according to Futurum Group estimates cited by CNBC. AMD is also formally launching its Venice CPU for data centers at the San Francisco event, a chip that has been in development for some time.

The launch comes after several partnership announcements earlier in the week that gave AMD commercial support for its next-generation AI hardware.

AMD Announces Anthropic Partnership

On Wednesday, AMD announced a partnership with Anthropic to deploy up to 2 gigawatts of Instinct MI450 Series GPUs inside Helios systems. The first gigawatt of deployment is scheduled to begin in the first half of 2027. The scale of the planned deployment underscores the enormous power requirements of next-generation AI training and inference workloads, a growing constraint for the data center industry.

As part of the agreement, AMD is making a strategic equity investment of up to $5 billion in Anthropic. The companies will also conduct a multi-year engineering collaboration focused on using Anthropic’s Claude AI to optimize workloads on AMD Instinct GPUs and accelerate development of AMD’s ROCm software stack.

AMD will additionally deploy Claude internally across its own engineering and product development teams.

Tom Brown, co-founder and chief compute officer at Anthropic, said: “By partnering with AMD across the stack, we are securing the capacity we need and optimizing it for training and serving Claude.”

AMD CEO Lisa Su said the Anthropic deal “brings together Anthropic’s leadership in frontier AI with the full strength of AMD high-performance computing.”

Microsoft to Deploy Helios on Azure

AMD also expanded its partnership with Microsoft earlier this week. Under that agreement, Microsoft will deploy Helios on Azure to run frontier model inference workloads.

AMD expects to begin shipping Helios to customers, including Microsoft, in the second half of 2026. The Microsoft deployment adds another large customer commitment for AMD’s AI infrastructure roadmap.

In October 2025, AMD also announced a multiyear agreement with OpenAI that could generate tens of billions of dollars in annual revenue. That deal gives OpenAI the option to buy up to roughly 10% of AMD.

AMD’s data center business has been growing alongside demand for AI and server hardware. The company’s data center segment reported $5.78 billion in revenue in Q1 2026, up 57% from the same period a year earlier. AMD said growth was driven by demand for EPYC server processors and shipments of Instinct GPUs.

With Anthropic, Microsoft, and OpenAI committed to its next-generation AI hardware, AMD is positioning Helios as a larger challenge to Nvidia’s data center AI systems. Nvidia’s market lead remains substantial, but AMD’s latest launch and customer agreements mark a broader effort to expand its role in AI infrastructure.