AMD and Antropic to Push for 2GW AI Infrastructure Construction
Published2026.07.23 12:18AMD and generative AI company Anthropic have entered into a strategic partnership to build gigawatt (GW)-class AI infrastructure. The two companies establish a long-term cooperation framework that includes building next-generation AI learning and inference environments, software optimization, and investment cooperation.
AMD and Antropic announced on the 22nd plans to deploy AMD Helios rack-scale solutions based on AMD Instinct MI450 series GPUs with a capacity of up to 2GW.
According to both companies, the construction of the first 1GW infrastructure is scheduled to begin in the first half of 2027.
The core of this collaboration is the introduction of AMD's next-generation AI platform, Helios.
Antropic plans to build an AI infrastructure based on AMD Instinct MI455X GPUs, AMD EPYC 'Venice' CPUs, AMD Pensando networking, and ROCm software.
AMD explained that Helios is a platform designed for large-scale AI training and inference workloads, providing performance, power efficiency, and scalability.
This project is an extension of the AMD Instinct MI355X-based environment currently utilized by Antropic.
The two companies are expanding cooperation not only in hardware supply but also in the software sector.
We plan to jointly pursue workload optimization for AMD Instinct GPUs and ROCm software development using Antropic's AI model, Claude.
AMD also plans to expand the use of Claude across its engineering and product development organizations.
AMD also announced that it plans to pursue a strategic equity investment of up to $5 billion in Antropic in the future.
Lisa Su, CEO of AMD, said, “By combining Antropic’s AI technology with AMD’s high-performance computing capabilities, we will accelerate the building of next-generation AI infrastructure.”He said, "dquo;".
Tom Brown, co-founder and CCO of Antropic, stated, “Securing sufficient computing capabilities is crucial for maintaining Claude’s competitiveness,” adding, “We are now able to operate infrastructure suitable for each workload by utilizing various hardware.”
Through this collaboration, the two companies plan to jointly pursue the construction of a large-scale AI data center and the development of a next-generation AI infrastructure platform.
관련뉴스
.jpg)
AMD, “'X100' High-End Embedded Platform Targets Physical AI and Robotics”
AMD's Naveen Bohra, Product Manager, stated during his keynote speech at the 'AMD x86 Embedded Solution Day' on the 10th that "x86 embedded is not simply an extension business, but a core strategic area in which we are investing long-term," and added, "We will pursue both rapid product release cycles and expanded long-term roadmaps in parallel."
2026.06.10by 배종인 기자
AMD Emphasizes 'Lab-Level CPU Performance' for Agentic AI… Presents EPYC-Based Infrastructure
AMD has provided guidance on the direction of data center infrastructure changes resulting from the spread of agentic AI. It emphasized that moving beyond a simple GPU-centric architecture, CPU performance and processing capability at the rack level are becoming increasingly important.
2026.06.11by 배종인 기자
.jpg)
AMD, “Robotics: ‘Response over Speed’ – Distributed AI and Heterogeneous Semiconductors Key”
AMD Systems Design Specialist Jung-hoon Jeong presented the architecture of robotic systems and the appropriate semiconductor architecture direction at the 'AMD x86 Embedded Solution Day' on the 10th, emphasizing that "robots are systems that must simultaneously satisfy not just computational performance, but 'latency, reliability, and power efficiency,'" and noted that as the era of so-called 'Physical AI'—including robotics, autonomous driving, and industrial automation—is in full swing, semiconductor structures fundamentally different from the existing data center-centric AI are now required.
2026.06.11by 배종인 기자
Nvidia is the 'Brain' and AMD is the 'Nervous System': The Physical AI Semiconductor War Begins in Earnest
The center of gravity in AI semiconductor competition is shifting beyond data center GPUs to 'Physical AI' such as robots, autonomous vehicles, and smart factories. Nvidia, led by CEO Jensen Huang, is seeking to reorganize the entire computing architecture by leading with its own Arm CPU and full-stack platform, while AMD, led by CEO Lisa Su, is emphasizing customer choice by leveraging x86 compatibility and an open ecosystem.
2026.06.11by 배종인 기자
AMD and Rackspace Sign Agreement to Build Phase of 30MW AMD AI Infrastructure
AMD와 랙스페이스 테크놀로지가 6월 17일 초기 30MW 규모의 전용 AI 컴퓨팅 구축 계약을 체결했다. 지난 5월 MOU를 이행하는 것으로, 2026년 말부터 2028년까지 랙스페이스 글로벌 데이터센터에 AMD 인스팅트 GPU(MI355X·MI350P)와 에픽 CPU가 배치된다. 양사는 데이터 주권과 규제가 엄격한 의료 등 산업의 임상 AI·대규모 추론 워크로드를 겨냥해, 거버넌스가 내재된 관리형 엔터프라이즈 AI 인프라 시장을 공동 구축한다.
2026.06.17by 배종인 기자
