It’s easy to define data centers solely by key metrics such as PUE, capacity, and time to compute. But as AI models grow more sophisticated and densities continue to rise, the idea of what makes up a modern data center must broaden, and attention must shift towards front-of-house performance.

Behind every expertly optimized white space lies a less celebrated, yet increasingly critical, layer of operational fabric: the gray space.

In a recent DCD>Broadcast, Pete Byrnes, industry marketing manager at Mitsubishi Electric Automation, unpacks why this often-overlooked element is becoming a strategic enabler of data center efficiency, resiliency, and speed of deployment.

Whether it's via smarter HVAC and backup systems, advanced controls, liquid-cooling integration, modular programmable logic controllers (PLCs), or built-in software intelligence, Byrnes argues that the gray space could hold the key to unlocking the next generation of facility optimization.

“People focus heavily on power access, PUE, and how fast they can deploy compute,” he says. “Those are important, but the small gains you can achieve inside gray space applications add up to a huge amount of efficiency for your final product. We’re working toward full-facility optimization – and gray space is a major part of that.”

Gray space, not a gray area

One of the fundamental hurdles in getting the gray space the attention and recognition it deserves lies in its definition. With so many different ideas surrounding what it constitutes, Byrnes sets the record straight:

“When we talk about gray space, we’re talking about the facility and environmental systems that support the core compute. HVAC solutions, backup systems, fuel pumps, generators, supporting circuitry for mission-critical security, and safety systems – that’s all gray space.”

Increasingly, this category also includes the software stack used to manage these systems – such as BMS and EPMS, or controls that sit between cooling infrastructure and compute. Within this context, Byrnes also places liquid cooling technology and CDUs under the gray space umbrella.

“This isn’t traditionally considered gray space, but because the controls and integration sit within this environment, we treat CDU implementation for immersion and direct liquid cooling as part of the gray space ecosystem.”

This broadened view reflects the growing complexity of today’s data center operations and architectures. With more interconnected components now required to manage and maintain critical uptime, gray space has evolved into a dense, data-rich layer with significant potential.

“There are a lot of really smart people who are too busy to think about the benefits you can gain from improving this layer,” says Byrnes. “That’s where we come in.”

Speed as the new foundation of efficiency

Gray space modernization goes beyond simply increasing efficiency of cooling or backup systems – it’s also about accelerating deployment, integration, and operational responsiveness.

“Efficiency comes from speed – both in implementing the compute or hardware you’ve selected, and then afterwards when the facility is up and running,” explains Byrnes.

A key example is Mitsubishi Electric’s deployment of CC-Link IE TSN – a time-sensitive networking protocol that Byrnes describes as huge a breakthrough:

“It’s powerful. We’re the first company to bring this technology to the US. Essentially, it reduces the requirement for cabling and speeds up implementation.”

The core innovation lies in synchronized clocks across all IoT-connected components, allowing different communications protocols to ride on the same cable in timestamped lanes.

“You’re able to communicate in different languages over one cable – essentially turning a single Ethernet line into a three-lane highway,” continues Byrnes. “And you can prioritize protocols based on different time signatures.”

Speed also comes from automation inside the equipment itself – such as auto-tuning within Mitsubishi Electric’s variable frequency drives (VFDs).

“Proportional-integral-derivative (PID) controls are more art than science, and they take a long time for facility engineers to implement,” says Byrnes. “Let us do that hard work for you and let the motor decide what’s best for the drive. It’s the most efficient way to drive any motor connected to one of our inverters.”

For operators looking to meet rapid deployment schedules and support high-density compute, these time savings translate directly into operational resilience and lower overheads.

High-availability control systems, without the wait

An additional pain point Byrnes highlights is within the supply chain. Many data center-specific components come with long lead times – presenting a major issue when build schedules are racing to keep pace with demand. Mitsubishi Electric Automation’s approach is different:

“We’re providing components that aren’t data center-specific SKUs,” says Byrnes. “That’s a significant benefit. We serve a range of industries, and we stock these components in large quantities at our facility in Illinois. For high run-rate systems, we’re leading in shipping times. We’ll beat your delivery time if you give us a chance.”

The company’s flagship iQ-R PLC – a high-performance, modular PLC platform designed for advanced factory automation – integrates multiple controls onto a single rack for superior speed, connectivity, and efficiency to deliver near-instantaneous failover.

“Switch times are ten nanoseconds if it’s under 90 meters of cable,” explains Byrnes. “And you can place units up to 550 meters apart. With these kinds of numbers, the only limitation on response time is the speed of light.”

Software intelligence

Shifts in the gray space aren’t exclusive to hardware. Instead, software orchestration plays an equal, or perhaps greater, role in how efficiently facilities operate. Byrnes shares two major areas where Mitsubishi Electric Automation is adding value: embedded logic and centralized software platforms.

Many of Mitsubishi Electric’s newest VFDs include a built-in simple PLC, allowing small applications to run without an overarching controller. This reduces cost, simplifies architecture, and cuts latency across control loops.

“In some cases, we’re removing PLCs entirely,” says Byrnes. “Or we’re doing more simple motions inside systems where others can’t. We can implement IO and smarten up the feedback loop in the fastest way possible.”

On the software side, Mitsubishi Electric’s GENESIS platform enables modern supervisory control and data acquisition (SCADA) functionality, digital twins, and operational intelligence. As Byrnes describes, the biggest initial benefit is simply connecting everything that can be connected:

“Most facilities we go into, not everything that’s IoT-enabled is actually hooked up,” he says. “We’re seeing an eight to ten percent increase in connected hardware once GENESIS is deployed. That gives higher visibility into energy consumption.”

For colocation providers, this means more accurate billing, and for operators, it means more accurate PUE – and, critically, a clear roadmap for improvements by identifying the squeaky wheels that need greasing.

The platform also includes a solutions accelerator – a productivity booster that preloads common graphics, layouts, and screens. “It saves ten to 15 hours of development time,” says Byrnes. “And it removes that blank-screen intimidation.”

Modularity and the future of CDUs

A major growth area for gray space optimization is liquid cooling, particularly in high-density AI-ready facilities. Here, Mitsubishi Electric Automation is working closely with CDU manufacturers to build systems with a modular design approach, to deliver built-in adaptability across multiple generations of GPU.

“The modular PLCs are going to be the path forward,” explains Byrnes. “We want to specify the perfect performance for each new generation of GPU. I envision a future where CDUs have longer lifespans because you can switch out individual components on an iQ-R-powered backplane. Hot-swappable, expandable input and output (IO).”

As cooling systems generate exponentially more data, expandable IO helps operators scale sensing and control without redesigning entire systems. This modularity also extends to motors and drives:

“When pumps stop pushing enough for the GPUs installed, you can switch out different types of motors,” says Byrnes. “Many of our drives can handle induction, IPM, or PM motors. If you want space savings, you can move to a PM motor without changing the drive.”

AI assistance

Beyond performance, Byrnes stresses the importance of maintainability – an area where gray space often becomes a bottleneck.

For example, Mitsubishi Electric’s GOT2000 series HMI interfaces allow operators to diagnose and debug VFDs or PLCs without opening the panel. “If it's a caustic environment or hard to reach, that’s a huge time saving,” notes Byrnes.

Inside Mitsubishi Electric’s FR Configurator software, AI-assisted fault analysis – powered by the company’s MaisartAI State of the Art Technology, and compact built-in AI systems – helps operators quickly identify issues and receive recommendations on how to improve system performance.

“AI is monitoring the data coming in and making suggestions,” says Byrnes. “It’s another way we help operators get more from their gray space systems.”

Combined with a 24-hour support line and impressive warranty, Mitsubishi Electric Automation is positioning itself as a long-term partner on the road to modernization.

The overlooked frontier

As data centers race to support AI and high-density compute on a global scale, industry-wide discourse has been dominated by white space design, power procurement, and efficiency metrics. Yet beneath these systems lies a complex and increasingly strategic operational layer that remains untapped.

“This area is often seen as someone else’s responsibility,” notes Byrnes. “We hear a lot of people say, ‘If it speaks the right communications protocol and shows up on time, I don’t care what the gray space control system is.’ We think that’s overlooking a huge benefit.”

Efficiency gains inside the gray space may not be at the very top of the priority list, but they compound into measurable improvements in uptime, energy performance, deployment speed, and operational clarity.

“We want operators to see gray space not as an afterthought, but as a strategic asset,” concludes Byrnes. “That’s where the next wave of optimization will come from.”

To hear more about how the gray space is changing, watch the full DCD>Broadcast episode with Mitsubishi Electric Automation’s Pete Byrnes here.