As AI compute density climbs faster than anyone planned for, the heat and power bottlenecks that have dominated data center roadmaps are creating a new problem: protecting the very equipment built to overcome them. A fully loaded high-density AI rack can top $3 million, replacement lead times are measured in quarters, and a coolant leak at the server level can ruin hardware within seconds and snowball down a rack. With direct-to-chip liquid cooling quickly becoming standard, comprehensive monitoring of coolant lines and UPS battery strings is no longer optional.
Ahead of SC26 (Supercomputing 2026) in Chicago, VMblog spoke with Joe Arena, Chief Commercial Officer at Parameter Technologies, the company behind RLE, Cellwatch, and NDSL. Arena discusses the debut of the XD Series for RLE Leak Detection, built with feedback from NVIDIA, Google, and Vertiv, and explains how Parameter fits into the BMS and DCIM environments operators already run. He also shares what attendees can see at booths 727 and 4200 on the McCormick Place show floor, including a live leak isolation demo with partner Rotork, and offers a look ahead to SC27, where retrofit opportunities, grid-participating battery assets, state-level disclosure rules, and insurance pricing could reshape how the industry thinks about monitoring.
++
VMblog: If you were giving an SC26 attendee a quick overview of Parameter, what would you say? How has Parameter evolved since SC25, and what has been the biggest change in your business over the past twelve months?
Joe Arena: We are the monitoring and detection layer for mission-critical operations. Our products continuously monitor for leaks, environmental conditions, and battery health to help technicians identify and resolve operational threats before they occur. We’ve been the industry leader in those applications for over 40 years. Whether they know us as Parameter, RLE, Cellwatch, NDSL, or one of our many excellent channel partners, most operators building the world’s largest AI infrastructure have installed our products somewhere in their facility.
The AI race catapulted compute density so fast that virtually everyone started pushing up against the same heat and power bottlenecks. The heat bottleneck was addressed first, as liquid cooling started becoming standard in new development plans. That put us in high demand, our production scale has doubled several times in the last year to match the order influx for leak cable. And with the power bottleneck being tackled in force, the demand for comprehensive battery systems has never been higher, so we don’t expect that growth trend to weaken any time soon.
But these new data centers don’t just need more leak detection, they need more complex leak detection. That’s why we built the XD Series for RLE Leak Detection, which debuted this last week. The XD Series is built from the ground up for liquid-cooled data centers, with higher scalability and precision than any other solution on the market. It’s a big goal to create a future-proof solution for the fastest-developing industry in history, but we’re very proud of what we’ve come up with.
VMblog: What’s your big message at SC26 this year? Everyone’s AI budget is under a microscope and hardware isn’t getting cheaper. What do attendees say to their CFO and CIO on the morning after the event?
Arena: Protect your most critical assets, that’s the bottom line. Memory and GPU supply is tight. A fully loaded high-density AI rack can top $3 million, and replacement time is measured in quarters, not months. Direct-to-chip cooling is quickly becoming standard, and a coolant leak at the server level will ruin equipment within seconds and snowball down a rack. Response time is critical for AI infrastructure, so whether you’re monitoring coolant lines or UPS battery strings, comprehensive monitoring is non-negotiable.
Some of the latest figures put unplanned downtime at up to $15,000 per minute for Global 2000 enterprises, roughly $600 billion a year. Uptime Institute’s 2026 outage analysis found that 57% of operators said their most recent major outage cost more than $100,000, with one in five above $1 million, and power (including battery backup failures) remains the leading cause. When it comes to downtime prevention, the budgetary concern should be on what you stand to lose.
VMblog: This year’s theme is ‘HPC Unites.’ Where does Parameter XD fit alongside the BMS, DCIM, and cooling systems operators already run, and which partners are part of that?
Arena: “HPC Unites” actually names one of the challenges we solve for. Parameter has made DCIM integration a priority for decades, and we were the first to introduce universal protocol conversion across all our control units, so any Parameter system can seamlessly communicate with the BMS through Modbus, BACnet, or SNMP.
Nowadays, the focus is on centralized monitoring. For years, operators built compute infrastructure to function, then added monitoring and alert systems à la carte, meaning each system had a different alarm, a different protocol, and interfaced with the DCIM differently. Facility management is limited by facility visibility; you can’t fix what you can’t see, and monitoring that reports in silos just recreates the fragmentation operators came to this show hoping to escape.
Parameter’s goal is to be a centralized monitoring platform that provides comprehensive, continuous visibility across all areas of your operation.
VMblog: Nearly every vendor now claims to be ‘AI-ready.’ What are your provable differentiators, and which production deployments or field results back them up?
Arena: We could point to our outstanding reputation for reliability or our history of industry-leading hardware and software innovations, but the crown jewel on this topic is our new XD Series. The system is built specifically for AI data centers, and over the last two years we’ve been developing it with direct feedback from AI giants like NVIDIA, Google, and Vertiv.
As a direct result of that input, XD Series boasts several unique innovations we see as crucial for leak detection in Next-gen AI systems. Right now, XD Series has the highest scalability and highest resolution of any solution on the market; it is the only leak detection system that supports PoE and the only one to natively support MQTT protocols. Beyond that, XD Series uses quad-channel sensors with relay outputs at every level, so a single sensing unit can monitor an entire rack and send a signal to isolate power and coolant flow the instant it detects a leak, reducing leak exposure by over 97%.
VMblog: How can attendees find you at SC26? What hands-on demos, live benchmarks, or interactive experiences will you have at your booth, and what’s the one thing you’d tell someone not to miss?
Arena: You’ll find us at two booths on the McCormick Place show floor. At booth 727, you’ll see automated server-level leak isolation in action with our strategic partner Rotork. XD Series sensors pick up a coolant leak and trace it to the affected server, then Rotork’s electrically actuated valves automatically shut off the flow within seconds.
You’ll find Parameter at booth 4200, where you can see live demos of XD Series, along with products and demos for Cellwatch Battery Monitoring, RLE Environment Monitoring, and RLE Leak Detection. I’ll be at the booth with our engineering team, and we’re happy to talk about where liquid cooling is headed and what it means for the people running these facilities.
As some added fun, we’ve also got an arcade machine in the booth. Top scores take home some pretty nice prizes, so swing by if you’ve got a competitive streak.
VMblog: Rack densities are climbing fast, and 800 VDC power and liquid cooling are now on nearly every roadmap. What are you showcasing for power, cooling, and performance per watt, and how do you help customers retrofit existing facilities rather than only build new ones?
Arena: To be clear, we don’t sell power or cooling equipment. Our job is protecting it. Higher densities mean more coolant in the rack, more stored energy in battery backup, higher DC voltages and less margin for error everywhere. At SC26, our focus is the XD Series, designed for liquid cooling at any scale and density, alongside Cellwatch battery monitoring for VRLA and lithium-ion, including string-level thermal runaway protection. Cellwatch works across battery chemistries and string configurations, so a change in battery vendor doesn’t force a change in monitoring vendor.
Most operators are retrofitting live halls, and we built our solutions with that in mind. Part of our design philosophy is that simplified deployment reduces installation time, minimizes failure points and technician error, and significantly increases system longevity. That’s one reason Parameter products are preferred for retrofit projects. XD Series is modular, and one controller covers over 8,000 zones – so teams can expand coverage rack by rack without adding controllers. You can install or remove Cellwatch hardware without ever disconnecting critical batteries. Our WiNG wireless sensors reach up to 600ft, last 12 years on a battery, and support auto discovery, so installation is as simple as sticking a sensor to any magnetic surface.
While our products offer plenty of convenience features to improve quality of life, the main draw we’re seeing retrofits care about is the rock-solid reliability of our hardware. We don’t move performance per watt directly, but operators push density harder when they know they can find and contain a leak at the rack in seconds.
VMblog: Looking toward SC27, what do you anticipate will be some of the challenges or opportunities infrastructure providers will face a year from now?
Arena: The industry has been building so fast that monitoring and redundancy are deferred as routine. Facilities going vertical right now are being commissioned with instrumentation suited to a 40 kW rack, not the 130 kW rack that will land in them. Gaps like this aren’t seen or addressed during commissioning and surface 18 months later as unexplained loss events, putting many of them squarely in the SC27 window. Enormous retrofit opportunity, and almost nobody is positioning for it.
UPS and BESS assets will move from backup to revenue-generating grid participation through programs like DCFlex. This will expand battery health from a reliability metric into a revenue metric.
Disclosure rules keep spreading at the state level, too. Pennsylvania already has water and power reporting on the books, and facility telemetry becomes a compliance artifact that has to survive an audit.
Finally, there’s insurance. Underwriters are starting to create data center lifecycle programs, and pricing premiums based on demonstrated monitoring coverage is the obvious next move, just as property insurers price sprinkler systems. Once instrumentation appears as a line item in a renewal, this stops being a facilities debate and becomes a finance decision.
