The Data Center Frontier Show

Artificial intelligence will continue to transform how data centers are designed, built, and operated, placing new demands on energy systems, infrastructure, and reliability. As AI workloads grow more intensive and always‑on, meeting these challenges will require a coordinated, systems‑level approach. In this episode, Patrick Hughes, SVP of Technical and Industry Affairs at the National Electrical Manufacturers Association (NEMA) will explore the AI Data Center Energy Performance Framework, developed in collaboration with ASHRAE and Pacific Northwest National Laboratory. The Framework provides practical, expert‑driven guidance to help owners, operators, engineers, and policymakers navigate the evolving AI landscape. He will provide an overview of why the Framework was created and how it is intended to be used. With thousands of data centers already operating—and many more planned—AI will drive higher load densities and increase pressure on both facilities and the grid. The Framework offers a shared foundation to align energy performance, reliability, and resilience across the full lifecycle of a data center. The conversation will also highlight NEMA’s role in ensuring electrical systems are fully integrated into data center design. Power distribution, safety, and infrastructure will need to work seamlessly with cooling and thermal management to avoid operational risks and support long‑term performance. A key theme of this Framework is collaboration. By bringing together NEMA’s leadership in electrical infrastructure, ASHRAE’s expertise in building systems, and PNNL’s energy research capabilities, the Framework will bridge traditional silos and promote a more integrated approach. We will also discuss how the Framework supports both new builds and existing facilities, helping organizations modernize infrastructure to meet AI demands. As a living, evolving resource, it will adapt alongside rapid changes in technology and energy needs. He’ll also explore what it means for communities and policymakers as data center growth accelerates—offering a path to balance innovation with reliability, efficiency, and long‑term infrastructure planning.

As AI infrastructure scales from megawatts to gigawatts, liquid cooling is rapidly becoming a foundational technology rather than a specialized option. In this episode of the Data Center Frontier Show podcast, recorded at Motivair's headquarters and manufacturing facility in Buffalo, New York, DCF Editor in Chief Matt Vincent sits down with Motivair CEO Rich Whitmore to discuss the evolution of liquid cooling from its roots in high-performance computing to its central role in today's AI data centers. Whitmore explains how Motivair's decade-plus experience supporting supercomputing environments helped position the company for the current AI boom, which he describes as the commercialization of traditional HPC at unprecedented scale. The conversation explores how liquid cooling products are developed years ahead of silicon roadmaps, why manufacturing discipline and testing standards have become competitive differentiators, and how global production capacity is increasingly essential as AI deployments accelerate worldwide. The discussion also examines one of the industry's emerging technical debates: whether ever-larger "facility-scale" coolant distribution units are the best answer for AI infrastructure. Whitmore offers a unique perspective on the realities of thermal management, noting that while AI workloads can change almost instantaneously, mechanical cooling systems must still operate within the physical constraints of pumps, valves, and fluid dynamics. The interview was recorded during a Schneider Electric global media event that included a tour of Motivair's Buffalo manufacturing operations and the nearby 750 MW TeraWulf Lake Mariner AI campus. There, Motivair liquid cooling technologies—including CDUs, in-rack manifolds, and ChilledDoor rear-door heat exchangers—are helping support one of North America's most ambitious AI infrastructure developments. As Whitmore explains, the question facing the industry is no longer whether liquid cooling will become mainstream. That transition is already underway. The challenge now is executing at scale—and building the manufacturing, supply chain, and engineering capabilities required to support the next generation of AI infrastructure.

Recorded live at Data Center World 2026, Data Center Frontier Editor in Chief Matt Vincent sits down with Phillip Koblence, COO of NYI and co-founder of Nomad Futurist, for the latest installment of Nomads at the Frontier. The conversation explores the accelerating realities of AI infrastructure buildouts, the industry’s growing focus on community engagement, workforce shortages, and the shift toward inference-driven deployments following NVIDIA GTC 2026. Koblence discusses why major interconnection hubs and edge-adjacent urban facilities may become increasingly important in the inference era, the operational realities of deploying AI infrastructure in legacy carrier hotels like 60 Hudson Street, and why the industry can no longer remain invisible to the communities where it builds. Additional topics include: The continuing surge in digital infrastructure demand Why conference attendance reflects sustained industry expansion Power constraints and energy storage discussions emerging at Data Center World AI factories and the evolving economic role of data centers Workforce shortages across engineering and skilled trades Nomad Futurist’s workforce development initiatives with Infrastructure Masons and I Am The Armed Forces The growing complexity and diversity of the data center ecosystem “Every element of everything within the data center has a full sub-vertical industry associated with it,” Koblence says during the discussion. “People would be surprised how large of an ecosystem is involved in creating the digital economy that exists today.” Listen now for a candid, fast-moving conversation on the state of AI infrastructure and the future of digital infrastructure development.

The AI infrastructure buildout has a gating problem, and it isn't megawatts. It's certainty of delivery. In this episode, Data Center Frontier Editor-in-Chief Matt Vincent sits down with Jim Summers, CEO of GPC Infrastructure, to examine what large-scale power delivery actually requires in today's market. Summers argues that hyperscalers are no longer shopping for energy. They're buying speed to market, guaranteed timelines, and risk transfer. Utilities, hamstrung by interconnection queues and uncertain delivery dates, increasingly can't provide those things. The conversation covers the full picture: why on-site natural gas has moved from bridge solution to permanent architectural layer, how battery systems have become essential infrastructure for managing AI's volatile load profiles, and what the supply chain — not energy policy — now governs project timelines. Summers also walks through GPC's mobile PPA structure, designed to give operators long-term cost amortization without locking equipment in place, and makes the case that waste heat capture will eventually become standard practice. The broader theme is risk. On-site generation shifts capital and operational responsibility to the developer. But it also hands them something utilities can't offer: direct control over their cost exposure, in a commodity market that is liquid and hedgeable. Power in the AI era, Summers concludes, is no longer a utility assumption. It is a negotiated outcome.

As AI workloads continue to scale, data centers are facing a new class of electrical challenges—ones driven not by total energy demand alone, but by how quickly that demand can change. AI training environments, particularly those built around dense GPU clusters, can cause rapid and unpredictable swings in power consumption. These fast load changes place stress on power systems that were originally designed for steadier, more predictable behavior. In the podcast, we explore why traditional approaches to power stabilization may not fully address the demand of AI-driven variability. While these approaches can absorb momentary spikes, they may fall short when it comes to sustained smoothing or supporting broader system stability. This becomes even more complex as many data centers are powered by on-site generation before transitioning to utility grid connections later in their lifecycle. The conversation highlights how newer energy storage strategies are evolving to meet these demands. Advanced battery-based systems, when paired with more adaptive control strategies, are designed to respond rapidly to load changes while operating effectively across different grid conditions. Rather than reacting only after voltage or frequency disturbances occur, these systems can proactively manage fluctuations at the point of interconnection, helping protect generation assets, improve power quality, and facilitate faster project timelines. As AI continues to push infrastructure into unfamiliar territory, the industry will need flexible, high-speed solutions that work across both islanded and grid-connected environments. Technologies designed with this adaptability in mind are quickly becoming a key enabler for the next generation of AI-ready data centers. These capabilities described herein reflect general technology characteristics and may vary based on system configuration, site conditions, and grid environment.

In today’s mission-critical supply chains, downtime is not an inconvenience—it’s a crisis. Whether supporting manufacturing, fabrication, integration or construction, warehouse management systems (WMS) have evolved from simple inventory tools into the digital backbone of high-stakes logistics environments. Today, Jarrett Atkinson, Vice President of Supply Chain for BluePrint Supply Chain explores how modern WMS platforms are redefining resilience, visibility, and performance in mission critical construction supply chains where failure is not an option We dive into what separates a standard WMS from one engineered for high-availability operations supporting multi-site deployment and specialized handling of large-scale gear. We will also discuss critical KPIs, reporting and visibility—how a WMS unlocks critical business insights that can improve efficiency, reduce costs, and eliminate project obstacles. Beyond technology, we also address implementation risk and examine the innovations poised to shape the next five years of mission-critical logistics.

As AI data center campuses scale toward gigawatt capacity, the industry is confronting a new kind of bottleneck. Not just how to generate power, but how to move it efficiently across increasingly complex environments. In this episode of the Data Center Frontier Show Podcast, MetOx CEO Bud Vos outlines why traditional copper-based power distribution may be approaching its limits, and how high-temperature superconducting (HTS) wire could offer a fundamentally different path forward. “When you start looking at gigawatt-type campuses, you find three fundamental constraints—the grid interconnect, campus distribution, and delivery inside the data hall,” Vos explains. At each layer, scaling with copper drives exponential increases in materials, infrastructure, and complexity. HTS technology changes that equation. By delivering roughly 10x the power density of copper, superconducting cables can dramatically reduce the physical footprint of power infrastructure, replacing dozens of conventional cables with just a few, while also cutting material use and simplifying system design. The technology also reverses a key trend in data center power architecture. Instead of pushing voltage higher to compensate for copper limitations, superconductors enable higher current at lower voltage, potentially simplifying electrical systems across the facility. Just as importantly, superconductors are effectively lossless. “They don’t generate heat as part of the power delivery infrastructure,” Vos notes, a property that could reshape how operators think about thermal management in high-density AI environments. While HTS systems require cooling with liquid nitrogen, that requirement may align with the industry’s broader shift toward liquid cooling. Beyond engineering, HTS could also play a role in easing permitting and community opposition by reducing the physical footprint of power infrastructure. Narrower rights-of-way and fewer materials translate into less visible impact—an increasingly important factor as data center development faces growing scrutiny. Crucially, superconducting systems are not theoretical. They have already been deployed in utility environments, providing a track record of reliability that may help accelerate adoption in the data center sector. As onsite and behind-the-meter generation become more common, HTS is particularly well-suited to moving large amounts of power across multi-building campuses and into high-density data halls. At the same time, the technology offers a potential alternative to strained supply chains for copper and traditional electrical equipment. Looking further ahead, superconductivity’s role may extend even deeper, with HTS materials also serving as a foundation for emerging fusion energy systems, hinting at a future where power generation and data center infrastructure are more tightly linked. For now, Vos sees the industry at the beginning of an adoption cycle. “We’re deploying, testing, and then innovating on top of that,” he says. As AI infrastructure enters its execution phase, superconductivity may move from a niche technology to a core component of how the next generation of data centers is powered.

Subzero Engineering is pleased to announce the acquisition of the Dissolvable Air Barrier (DAB) Panels product line from Cambridge R&D, further expanding Subzero’s portfolio of data center containment solutions and reinforcing its commitment to safety, performance, and turnkey system delivery. DAB Panels are a unique overhead containment solution designed to provide effective airflow separation during normal data center operation while dissolving within seconds when exposed to water during sprinkler activation. This dissolvable design helps eliminate falling panel hazards and supports safer fire suppression outcomes—addressing a critical challenge found in traditional rigid overhead containment systems. “With this acquisition, we’re strengthening our ability to deliver truly integrated, safety-driven containment solutions,” said Shane Kilfoil, President of Subzero Engineering. “DAB Panels complement our existing containment portfolio and give our customers another proven option to address airflow management and fire safety without compromise.” DAB Panels are engineered for both hot aisle and cold aisle containment applications and offer a combination of airflow performance, safety, and installation flexibility. Made from EPA-certified, plant-based cellulose materials, the panels achieve Class A fire and smoke performance, producing low heat and minimal smoke while maintaining visibility for emergency personnel. Despite their dissolvable design, DAB Panels remain durable during normal operation—withstanding high static air pressure and maintaining airflow separation where it matters most. Panels can be easily modified in the field to accommodate varying cabinet heights and existing infrastructure, eliminating the need to relocate sprinkler heads and reducing installation time and cost. DAB Panels integrate seamlessly across Subzero’s full portfolio of data center containment products, including aisle frames, doors, roofs, and airflow management systems. This unified approach enables Subzero to deliver turnkey containment solutions engineered for performance, safety, and long-term scalability—backed by a single partner and a coordinated system designed to work together.

The AI infrastructure boom is rapidly reshaping how the data center industry thinks about power. What was once a relatively straightforward utility procurement exercise is evolving into a complex strategy spanning onsite generation, fuel logistics, financing, and system architecture. That reality framed a recent special edition of The Data Center Frontier Show Podcast, which recast and updated a pivotal DCF Trends Summit 2025 session: From Grid to Onsite Powering: Optimizing Energy Behind the Meter for Data Centers. Moderated by Fengrong Li, Senior Managing Director at FTI Consulting, the panel explored how operators are responding as interconnection timelines stretch and AI workloads surge. Li’s framing emphasized a core shift: onsite power is moving from contingency planning to critical-path infrastructure. From the OEM perspective, David Blank of Siemens Energy noted that behind-the-meter deployments have accelerated sharply over the past year as developers confront multi-year waits for firm utility capacity. “Everyone would prefer grid power,” Blank said. “But in many cases, reliable access isn’t available for five, ten, even ten-plus years.” Panelists agreed that AI’s scale and speed are driving a structural rethink. Brian Gitt of Oklo described the moment as a return to industrial roots, with large loads once again building dedicated generation to meet growth timelines. At the same time, new technical pressures are emerging. AI clusters can produce sharp load swings, forcing developers to deploy fast-response buffering technologies such as batteries, flywheels, and supercapacitors to maintain stability. Despite differing technology paths—including gas turbines, hydrogen fuel cells, and advanced nuclear—the panel aligned on one common theme: modularity. Phased power blocks increasingly mirror how AI campuses are actually built and financed. The discussion also highlighted the growing importance of contract structures. Long-term offtake commitments, capacity reservations, and credit support are increasingly required to unlock equipment queues and fuel supply. Other panelists included Marty Trivette of AlphaStruxure and Yuval Bachar of ECL. The event was hosted by Data Center Frontier’s Matt Vincent. The takeaway was clear: in the AI era, energy strategy has moved to the critical path—and for many operators, that path now runs behind the meter.

In the latest episode of The DCF Show Podcast, Data Center Frontier founder Rich Miller joins present DCF Editor in Chief Matt Vincent and Senior Editor David Chernicoff to examine where the data center industry stands as AI infrastructure moves from announcement to execution. Miller also discusses his new Data Center Richness podcast and Substack project, which explores how data center professionals consume content and learn about the rapidly evolving industry. With information overload now a reality, Miller’s goal is to distill the most important signals shaping infrastructure decisions. The conversation then turns to what defines 2026 for data centers: execution. After a year filled with megaproject announcements, the industry now faces the harder task of actually delivering campuses at AI scale—often under severe power constraints. With utilities struggling to keep pace, on-site generation is shifting from temporary solution to long-term strategy, as developers seek reliable ways to power projects while easing community concerns about grid impacts. Public resistance has also become a major factor. Miller notes that community opposition is now delaying or halting billions of dollars in projects, forcing operators to rethink how they engage with local stakeholders. Issues like power pricing and water usage are increasingly central to project approval. On the technology front, Nvidia’s roadmap continues to reshape infrastructure planning, with rack densities rising sharply, liquid cooling becoming standard, and new power distribution models emerging to support AI factories. At the same time, Miller expects the market to stratify, with some operators specializing in AI factories while others serve cloud and enterprise demand. The discussion also touches on nuclear power’s future role, with data centers positioning themselves as anchor customers, though meaningful SMR deployment remains years away. Ultimately, Miller argues that the industry is moving faster than ever, and 2026 will reveal how well today’s massive investments translate into real deployments. As he concludes: the next phase belongs to those who can deliver.

In the latest episode of the Data Center Frontier Show Podcast, Editor in Chief Matt Vincent speaks with Sailesh Krishnamurthy, VP of Engineering for Databases at Google Cloud, about the real challenge facing enterprise AI: connecting powerful models to real-world operational data. While large language models continue to advance rapidly, many organizations still struggle to combine unstructured data (i.e. documents, images, and logs) with structured operational systems like customer databases and transaction platforms. Krishnamurthy explains how vector search and hybrid database approaches are helping bridge this gap, allowing enterprises to query structured and unstructured data together without creating new silos. The conversation highlights a growing shift in mindset: modern data teams must think more like search engineers, optimizing for relevance and usefulness rather than simply exact database results. At the same time, governance and trust are becoming foundational requirements, ensuring AI systems access accurate data while respecting strict security controls. Operating at Google scale also reinforces the need for reliability, low latency, and correctness, pushing infrastructure toward unified storage layers rather than fragmented systems that add complexity and delay. Looking toward 2026, Krishnamurthy argues the top priority for CIOs and data leaders is organizing and governing data effectively, because AI systems are only as strong as the data foundations supporting them. The takeaway: AI success depends not just on smarter models, but on smarter data infrastructure. 🎧 Listen to the full episode to explore how enterprises can operationalize AI at scale.

Applied Digital CEO Wes Cummins joins Data Center Frontier Editor-in-Chief Matt Vincent to break down what it takes to build AI data centers that can keep pace with Nvidia-era infrastructure demands and actually deliver on schedule. Cummins explains Applied Digital’s “maximum flexibility” design philosophy, including higher-voltage delivery, mixed density options, and even more floor space to future-proof facilities as power and cooling requirements evolve. The conversation digs into the execution reality behind the AI boom: long-lead power gear, utility timelines, and the tight MEP supply chain that will cause many projects to slip in 2026–2027. Cummins outlines how Applied Digital locked in key components 18–24 months ago and scaled from a single 100 MW “field of dreams” building to roughly 700 MW under construction, using fourth-generation designs and extensive off-site MEP assembly—“LEGO brick” skids—to boost speed and reduce on-site labor risk. On cooling, Cummins pulls back the curtain on operating direct-to-chip liquid cooling at scale in Ellendale, North Dakota, including the extra redundancy layers—pumps, chillers, dual loops, and thermal storage—required to protect GPUs and hit five-nines reliability. He also discusses aligning infrastructure with Nvidia’s roadmap (from 415V toward 800V and eventually DC), the customer demand surge pushing capacity planning into 2028, and partnerships with ABB and Corintis aimed at next-gen power distribution and liquid cooling performance.

AI data centers are no longer just buildings full of racks. They are tightly coupled systems where power, cooling, IT, and operations all depend on each other, and where bad assumptions get expensive fast. On the latest episode of The Data Center Frontier Show, Editor-in-Chief Matt Vincent talks with Sherman Ikemoto of Cadence about what it now takes to design an “AI factory” that actually works. Ikemoto explains that data center design has always been fragmented. Servers, cooling, and power are designed by different suppliers, and only at the end does the operator try to integrate everything into one system. That final integration phase has long relied on basic tools and rules of thumb, which is risky in today’s GPU-dense world. Cadence is addressing this with what it calls “DC elements”: digitally validated building blocks that represent real systems, such as NVIDIA’s DGX SuperPOD with GB200 GPUs. These are not just drawings; they model how systems really behave in terms of power, heat, airflow, and liquid cooling. Operators can assemble these elements in a digital twin and see how an AI factory will actually perform before it is built. A key shift is designing directly to service-level agreements. Traditional uncertainty forced engineers to add large safety margins, driving up cost and wasting power. With more accurate simulation, designers can shrink those margins while still hitting uptime and performance targets, critical as rack densities move from 10–20 kW to 50–100 kW and beyond. Cadence validates its digital elements using a star system. The highest level, five stars, requires deep validation and supplier sign-off. The GB200 DGX SuperPOD model reached that level through close collaboration with NVIDIA. Ikemoto says the biggest bottleneck in AI data center buildouts is not just utilities or equipment; it is knowledge. The industry is moving too fast for old design habits. Physical prototyping is slow and expensive, so virtual prototyping through simulation is becoming essential, much like in aerospace and automotive design. Cadence’s Reality Digital Twin platform uses a custom CFD engine built specifically for data centers, capable of modeling both air and liquid cooling and how they interact. It supports “extreme co-design,” where power, cooling, IT layout, and operations are designed together rather than in silos. Integration with NVIDIA Omniverse is aimed at letting multiple design tools share data and catch conflicts early. Digital twins also extend beyond commissioning. Many operators now use them in live operations, connected to monitoring systems. They test upgrades, maintenance, and layout changes in the twin before touching the real facility. Over time, the digital twin becomes the operating platform for the data center. Running real AI and machine-learning workloads through these models reveals surprises. Some applications create short, sharp power spikes in specific areas. To be safe, facilities often over-provision power by 20–30%, leaving valuable capacity unused most of the time. By linking application behavior to hardware and facility power systems, simulation can reduce that waste, crucial in an era where power is the main bottleneck. The episode also looks at Cadence’s new billion-cycle power analysis tools, which allow massive chip designs to be profiled with near-real accuracy, feeding better system- and facility-level models. Cadence and NVIDIA have worked together for decades at the chip level. Now that collaboration has expanded to servers, racks, and entire AI factories. As Ikemoto puts it, the data center is the ultimate system—where everything finally comes together—and it now needs to be designed with the same rigor as the silicon inside it.