TAIPEI, May 20, 2025 /PRNewswire/ — The Open Compute Project Foundation (OCP) is redesigning data center power architecture to support AI’s growing demands, introducing “1 Megawatt racks” that could slash energy losses from 40% to just 7%, according to Chief Innovation Officer Cliff Grossner.
With AI infrastructure stretching the limits of current systems, OCP’s community of over 400 corporate members is developing open standards to address critical bottlenecks in compute density, power distribution, and cooling.
“The industry is at an inflection point,” said Grossner. “We no longer assume that all processors can be generic. We need to create custom processors for specific workloads.”
Power Revolution
Current power conversion processes waste approximately 40% of energy when converting from high voltage to chip-level voltage—a significant inefficiency as facilities increasingly consume hundreds of megawatts.
OCP’s proposed “1 Megawatt racks” would move power supplies out of server racks into separate units. Eventually, power generation could move entirely outside computing floors, with facilities re-architected to supply 400 or 800 volt DC power.
Future data centers may incorporate on-site renewable energy generation through solar, wind, and potentially small nuclear reactors as part of an ecosystem approach to power management.
Environmental Focus
With data centers projected to consume 3% of global electricity by 2030, sustainability has become a priority. OCP is developing metrics beyond traditional Power Usage Effectiveness (PUE) to address water consumption and equipment efficiency.
“A data center is not a standalone object in the future,” Grossner noted. “It’s going to be part of an ecosystem that includes potentially a wind farm next to it, and proximity to buildings that can use the heat produced.”
The organization is also investigating alternative concrete formulations to reduce carbon emissions in construction and collaborating with Infrastructure Masons on standardized carbon accounting for IT equipment.
Cooling Transformation
Increasing computational density is forcing a transition from air to liquid cooling systems.
“The thermal management needs of current and next-generation silicon make it pretty clear that chips will be cooled either by single or two-phase liquid cooling,” Grossner said.
OCP is working to standardize liquid cooling delivery systems, including specifications for connectors, pressures, and fluid properties to ensure interoperability across vendors.
Silicon Innovation
OCP has established a marketplace with over 25 suppliers offering modular semiconductor components, positioning itself as the “front door to the open chiplet economy.” This approach enables specialized silicon components to be combined for optimal performance.
The organization envisions distinct roles in this ecosystem: companies creating chiplets, businesses building ASICs, and firms providing support through design tools and testing services.
Education Initiatives
OCP recently launched the OCP Academy to support adoption of its standards and provides access to hundreds of OCP-recognized products through its marketplace. OCP aims to remain the premier organization for AI infrastructure by standardizing silicon, power, cooling, and interconnects; supporting complete system development; and providing industry education.
As AI continues to redefine computing requirements, OCP’s role in fostering open, sustainable, and scalable infrastructure appears increasingly vital to managing the environmental impact of explosive computational growth.
Global Expansion
With 30% of its membership in Asia-Pacific (primarily Taiwan), OCP will hold its APAC Summit in Taipei on August 5-6, 2025. “Taiwan’s semiconductor industry is a key partner,” Grossner emphasized.
Event Link: https://www.opencompute.org/summit/2025-ocp-apac-summit
