Introducing PCE: Why PUE No Longer Says Enough in the AI Factory Age
November 17, 2025
The New Data Center Blueprint: Why Modular Is Winning in the AI Era
November 20, 2025As artificial intelligence and high-performance computing (HPC) workloads continue to reshape the data center landscape, one thing is clear: traditional cooling methods can no longer keep up.
In this exclusive JSA TV interview, Tony Fischels, Vice President of PowerOne™ at AIRSYS Cooling Technologies, discusses how AIRSYS is tackling these challenges head-on with a new division and a groundbreaking approach to liquid cooling.
Meeting AI Workload Demands Through Innovation
“With the growth of AI, we’re seeing unprecedented demands on compute,” says Fischels. As data centers transition into the AI and HPC era, power density and heat loads are increasing dramatically, which is why all supporting infrastructure must evolve in parallel with compute technology.
This is the mindset that drove AIRSYS to develop PowerOne.
Introducing PowerOne: A New Era of Cooling Innovation
In the interview, Fischels shares what led to the creation of PowerOne: “PowerOne was really inspired by challenges we’re seeing in today’s data center environment. We’ve designed it to tackle the issues around access to stranded power and the growing energy demands of evolving infrastructure.”
The PowerOne division offers two distinct solution sets:
- Standard cooling systems: The tried-and-true methods used in most data centers today.
- LiquidRack™: A completely new category, referred to as a server-level liquid cooling unit.
Liquid Cooling Is the Future, and LiquidRack Is Built for What’s Next
LiquidRack is AIRSYS’s direct answer to the growing thermal demands of modern AI-driven workloads. As the industry recognizes that liquid cooling is the only sustainable path for supporting ever-increasing power densities, LiquidRack stands as a clear testament to how AIRSYS continues to evolve in step with data center technology.
Unlike conventional cold-plate systems, LiquidRack uses single-phase spray cooling technology applied directly to the chip, an approach that dramatically “allows for more heat rejection at the GPU level compared to traditional systems”.
This design not only boosts cooling efficiency and performance but also makes LiquidRack suitable for both new builds and retrofit applications. As Fischels notes, the result is a major leap forward in thermal management for high-density, AI-driven environments.
PCE: The Power of Power
What sets LiquidRack apart is the focus on accessing stranded power, recognizing that today’s data center operators and investors need clearer visibility into how power is allocated and utilized. As Fischels puts it, “power allocation decisions are now often made around compute, not mechanical systems.”
This shift in perspective led AIRSYS to introduce a new performance metric: Power Compute Effectiveness (PCE).
“PCE measures the ratio of power dedicated to compute versus the total power capacity of a data center,” Fischels explained. “It complements the traditional Power Usage Effectiveness (PUE) metric but reflects how modern data centers are being designed today.”
By emphasizing compute-driven energy efficiency, PCE reframes how operators measure success in a high-density, high-demand environment.
Innovation and Collaboration
Fischels credits innovation as the key to meeting the rising thermal and energy challenges brought by AI workloads.
He also highlighted the company’s partner division, Aegis, which focuses on two-phase cooling solutions. “They’re using a flow boiling approach and 3D-printed cold plates,” Fischels noted. “It’s exciting to see the level of advancement happening within the AIRSYS family.”
Start Your PowerOne Journey
For data center operators and engineers looking to learn more about PowerOne’s solutions, please contact us today and follow us on LinkedIn.
