The Open Compute Project (OCP) Global Summit, held at the San Jose Convention Center, drew a record crowd of more than 11,000 participants, underscoring the event's growing importance in the global data center industry. Two major developments stood out during the keynote session: an expanded OCP board reflecting stronger collaboration across the ecosystem, and a strategic shift in technical focus from rack-level designs to power and connectivity beyond the data center.
OCP expands board to include Nvidia, AMD, and Arm
OCP CEO George Tchaparian announced the addition of three new board members: Rob Ober, Chief Platform Architect for Data Centers at Nvidia; Robert Hormuth, corporate vice president for Strategy and Platforms at AMD; and Mohamed Awad, senior vice president and general manager of Infrastructure Line of Business at Arm.
Previously, the six-member board included representatives from Meta, Google, Microsoft, Intel, and Arista Networks co-founder Andy Bechtolsheim. The inclusion of Nvidia, AMD, and Arm reflects a more diversified semiconductor representation, addressing the growing dominance of AI data centers and Intel's declining share in server CPUs.
The new composition broadens OCP's influence across the computing stack—from x86 and Arm CPUs to GPUs—and strengthens collaboration within the open hardware ecosystem. Notably, future board expansions could potentially include AWS, Broadcom, or even OpenAI, which is reportedly preparing massive AI data center investments starting in 2026.
Technical vision expands beyond racks to grid-level systems
Founded by Meta in 2011 and joined by Microsoft and Google in later years, OCP has evolved from defining rack-level standards to addressing AI compute modules, cooling systems, data center sustainability, and security frameworks.
From 2011 to 2016, OCP's standards focused on in-rack architectures, including servers, racks, power, networking, and storage. Between 2017 and 2024, the scope expanded to include AI accelerator modules (Accelerator & AI Compute), cooling and thermal management, data center facilities and sustainability (Rack & Power), as well as security and compliance.
In his keynote, Tchaparian outlined two key development paths for OCP: a topological expansion from hyperscale to AI and edge data centers, and a scale expansion from chips and systems to facilities and even power grid integration.
Speakers also highlighted the growing importance of power systems, thermal management, and modular data center design. Microsoft's Saurabh Dighe noted that after scale-out computing, the next frontier will be AI-wide area networks (AI WANs)—linking compute resources across regional data centers through next-generation telecom infrastructure.
These trends indicate that OCP's influence is moving beyond traditional hardware design, extending toward the broader infrastructure ecosystem that powers AI-driven computing.
Credit: DIGITIMES
Credit: DIGITIMES
Article edited by Jerry Chen
