Originally posted on Datalec LTD.

Data centre leaders left ExCeL London earlier this month with one message ringing loud and clear: AI‑driven growth is accelerating, power is tight, and the choice of infrastructure partner is now business‑critical, not optional.

Against a backdrop of rapid hyperscale and colocation expansion, constrained power availability and rising energy scrutiny, the conversations at Data Centre World London 2026 underscored that operators need partners who can help them plan power‑first, deploy at speed, and operate reliably in high‑density environments.

For Datalec Precision Installations (DPI), DCW London was an opportunity to demonstrate exactly that kind of integrated, global capability, from modular data centre solutions through to facilities management, consultancy and lifecycle services. The questions operators brought to the stand were remarkably consistent, whether they were building in the UK, expanding in the Middle East, or planning their next phase of growth in APAC.

Below, we revisit three of the most important questions AI‑driven operators were asking in London and why they will matter even more as the industry converges on Singapore for DCW Asia later this year.

1. How quickly can you take me from secured power to live, AI‑ready capacity?

If there was one common theme at DCW London, it was that power availability has become the primary constraint on new data centre builds, not demand. Once operators have secured land and grid, the urgent requirement is simple: how fast can we safely turn that capacity into revenue‑generating, AI‑ready infrastructure?

This is where modular, pre‑engineered solutions dominated the conversation. Many visitors to the DPI stand wanted to understand how modular white space, plant and service corridors could compress design and construction timelines without sacrificing resilience or compliance. DPI’s next‑generation Modular Data Centre Solutions attracted strong interest because they are designed precisely for this challenge. They help clients move from planning to live halls at speed, whether that’s a new campus in a European hub, a hyperscale expansion in the Middle East, or an edge or colocation site in a fast‑growing APAC market.

To continue reading, please click here.

The post Three Questions Every AI‑Driven Operator Should Ask Their Infrastructure Partner appeared first on Data Center POST.

By Mike Hodge, AI Solutions Lead, Keysight Technologies

It’s the heart of the AI gold rush, and everyone wants to capitalize on the next big thing. Large language models, multimodal systems, and domain-specific AI workloads are moving from experimentation to production at scale. Across industries, enterprises are building their own proprietary models or integrating pre-trained ones to power applications spanning from video analytics to highly specialized inference services.

This shift has triggered a new wave of infrastructure investment. But while GPUs and accelerators dominate the conversation, scaling AI platforms has produced a less obvious constraint: front-end network performance. In increasingly distributed, multi-tenant AI environments, the ability to move data efficiently into (and across) platforms has become just as critical as raw compute density.

New AI platforms mean new expectations for infrastructure

AI infrastructure is no longer the exclusive domain of a handful of hyperscalers. A growing class of service providers has begun offering end-to-end AI platforms where compute, storage, networking, and orchestration are delivered as a service. Their value proposition is straightforward: customers bring data and models, while the platform handles the complexity of building, operating, and maintaining large-scale data center deployments.

Service models like these, however, place extraordinary demands on networking. Unlike traditional cloud workloads, AI jobs are defined by massive, sustained data movement and tight coupling between data pipelines and compute utilization. GPUs cannot perform at peak efficiency unless data arrives on time, in the right order, and at predictable speeds.

As a result, network performance is now one of the primary determinants of training, inference, and infrastructure efficiency.

The eye of the storm is moving from the fabric to the front end

AI infrastructure discussions often focus on back-end fabrics. Think about things like high-bandwidth, low-latency interconnects between GPUs, for example. However, while these fabrics are indeed essential, they are only part of the picture.

Before training or inference ever begins, data must first traverse the front-end network. This occurs in several ways, but some of the most common paths include:

• From remote object stores or on-premises repositories into the data center
• From ingress points into virtual machines or containers
• From storage into GPU-attached hosts

This is where north-south traffic (external to internal) intersects with east-west traffic (host-to-host and service-to-service). And in AI environments, these flows are not occasional spikes. They are sustained, high-throughput, latency-sensitive streams that run continuously throughout the lifecycle of a job.

When front-end networks underperform, the consequences are costly and immediate: idle accelerators, elongated training windows, unpredictable inference latency, and poor multi-tenant isolation.

Why traditional network validation falls short

Most cloud networks were designed around general-purpose workloads. Think about things like web services, databases, and transactional systems with relatively modest bandwidth demands and fluctuating traffic patterns punctuated by the occasional spike.

AI workloads, on the other hand, break these assumptions. On the front end, AI traffic is characterized by:

• Extremely large data transfers, often using jumbo frames
• Long-lived connections, sustained over hours or days
• Millions of concurrent sessions in multi-tenant environments
• Tight latency and jitter tolerances to avoid starving accelerators

Conventional network testing approaches — such as synthetic benchmarks, isolated link tests, or small-scale simulations — are unable to replicate this behavior. As a result, many issues only surface once customer workloads are already running, which also happens to be when the cost of remediation is highest.

The need for realistic workload emulation

Optimizing front-end AI networks requires the ability to reproduce real workload behavior at scale. That means emulating both north-south and east-west traffic patterns simultaneously, across distributed environments and under sustained load.

For north-south paths, this includes verifying that large datasets can be reliably pulled from diverse external sources into local storage. Moreover, the network must also be able to do so with consistent throughput, predictable latency, and no silent data loss. Transfers like these are essential, as any inefficiency propagates directly into longer training times and underutilized GPUs.

For east-west paths, the challenge shifts to connection density, latency, and scalability. Once workloads are running, virtual machines and services exchange data continuously. Sometimes within the same host, sometimes across racks, and sometimes across geographically separated data centers. Modern AI platforms increasingly rely on SmartNICs and offload technologies to make this feasible, so these components must also be validated under realistic connection rates and protocol behavior.

Without large-scale, workload-accurate testing, subtle bottlenecks — such as rule-processing limits, connection-tracking inefficiencies, or unexpected latency spikes — can remain hidden until production traffic exposes them.

Front-end optimization is a competitive differentiator

In response, the most advanced AI platform operators are shifting left: validating their front-end networks before customers ever deploy workloads. Along the way, their proactive approach is changing the economics of AI infrastructure.

Stress-testing networks under real-world conditions offers a range of benefits for network operators:

• Identifying performance cliffs at high line rates
• Understanding how different layers of the stack interact under load
• Resolving scaling limitations in NICs, virtual networking, or storage paths
• Delivering predictable performance across tenants and geographies

It’s not just about improving peak throughput. It’s about building confidence that platforms perform as expected under peak pressure. And in a market where AI workloads are expensive, time-sensitive, and strategically important, this confidence becomes a differentiator. Customers may never see the network directly, but they feel its impact in faster training cycles, lower inference latency, and fewer production surprises.

Looking ahead: front-end networks and the next generation of AI

AI workloads continue to evolve. Microservices-based architectures, distributed inference pipelines, and increasingly stateful services are placing even more emphasis on low-latency, high-availability front-end connectivity. At the same time, data is becoming more geographically distributed, pushing platforms to span multiple regions and network domains.

In this environment, front-end networks are no longer a supporting actor. They are a core component of AI system design. That means they must be engineered, validated, and optimized with the same rigor applied to compute and accelerators.

The lesson is clear: operators cannot optimize AI infrastructure by focusing on GPUs alone. The performance, efficiency, and reliability of tomorrow’s AI platforms will be defined just as much by how well they move data as by how fast they process it.

The post AI’s Overlooked Bottleneck: Why Front-End Networks Are Crucial to AI Data Center Performance appeared first on Data Center POST.

AI workloads are pushing data centre infrastructure towards higher rack densities, new cooling strategies and greater power demand. Jamie Darragh, Data Centre Director, Europe, at global data centre engineering design consultancy Black & White Engineering, examines the design implications for the next generation of facilities.

AI and high-performance computing are placing new demands on data centre infrastructure. Rack densities are increasing; facilities are being delivered at larger scale and operators are under pressure to support workloads that consume far greater levels of power and generate far higher heat loads than conventional cloud environments.

Independent forecasts underline the pace of expansion. Gartner estimates global data centre electricity consumption will rise from around 448TWh in 2025 to roughly 980TWh by 2030, driven largely by AI-optimised computing infrastructure. Within that growth, AI servers alone are expected to account for close to 44% of data centre power consumption by the end of the decade.

For our engineering teams, these workloads are altering the practical limits of traditional infrastructure design. Rack densities exceeding 100–200kW are now appearing in project specifications, particularly where large AI training clusters are planned. These loads influence every part of the building environment, from electrical distribution and cooling capacity to structural loading and cable management.

Designing for extreme density

Under these conditions, air cooling alone becomes difficult to sustain across entire facilities. Liquid cooling is therefore increasingly included in the baseline design of new data centres rather than introduced later as a specialist solution. This cooling method is becoming increasingly favourable due to its higher specific thermal capacity compared with air, which enables more efficient heat transfer and removal. Direct-to-chip and rack-level systems are being designed alongside air cooling so facilities can accommodate different densities and equipment types across the same site.

The introduction of liquid systems requires careful coordination between disciplines. Facilities must manage environments where air and liquid cooling operate together, supported by monitoring platforms, safety controls and operational procedures capable of supporting both approaches.

Some IT chips require different liquid cooling temperatures than those used in air-cooling systems, creating technical hurdles for the overall heat rejection system and requiring precise control of the cooling circuit temperature. Another engineering challenge lies in integrating these systems with power distribution, control platforms and maintenance strategies rather than selecting one cooling method over another.

Higher density also narrows operational tolerance. Commissioning becomes more demanding and redundancy strategies require more detailed modelling. Infrastructure must be capable of supporting peak compute demand while maintaining efficiency when loads are lower, placing greater emphasis on flexible electrical and mechanical systems.

The scale of development is also increasing. Buildings that once delivered a few megawatts of capacity are now part of campus-scale developments where multiple data halls contribute to facilities delivering hundreds of megawatts. data centres are increasingly planned and delivered as long-term infrastructure assets rather than individual projects.

This environment encourages repeatable design and industrialised delivery methods. Developers and investors expect predictable construction schedules and consistent performance across multiple sites. As a result, engineering teams are placing greater emphasis on modular infrastructure systems and digital design methods that allow mechanical and electrical systems to be configured and deployed repeatedly.

Power, control and operational intelligence

Power availability is also becoming a determining factor in project planning. In many regions, grid connection capacity is now one of the main constraints on new development. Gartner has warned that by 2027 as many as 40% of AI data centres could face operational limits because of power availability.

Developers are therefore engaging more closely with utilities during early feasibility stages and exploring complementary infrastructure such as on-site generation and energy storage. In some cases, data centres are also being designed to contribute to wider grid stability through demand response and energy management capability.

Artificial intelligence is also beginning to influence how facilities themselves are operated. Machine-learning systems are already being used in some environments to optimise airflow patterns, cooling plant performance and power distribution using live operational data.

The next stage will see more widespread use of integrated control platforms and digital twins capable of modelling facility behaviour in real time. These systems allow operators to simulate infrastructure performance under different load conditions, test operational changes and identify maintenance requirements before faults occur.

Environmental performance remains another constraint as compute density increases. Higher workloads place additional pressure on energy supply while raising questions around water consumption, construction materials and waste heat recovery. Planning authorities and investors are increasingly looking for measurable improvements in efficiency and carbon reporting before approving new developments. Sustainability therefore sits alongside power and cooling as a central engineering consideration rather than a secondary design feature.

Taken together, these conditions create a more complex design environment for data centre infrastructure. Higher compute densities, power constraints and new operational technologies require mechanical, electrical and digital systems to be considered together from the earliest design stages.

Facilities intended to support AI workloads must accommodate far greater performance requirements than earlier generations of data centres while remaining adaptable as infrastructure technologies and operating practices continue to develop.

# # #

About the Author

Jamie Darragh is Data Centre Director, Europe at Black & White Engineering. He leads the delivery of complex, mission-critical projects across the region, with a focus on technical quality, design coordination and strong client relationships. A Chartered Engineer and member of CIBSE and the IET, Jamie has worked across Europe, the Middle East and the UK since 2005. He brings a clear, practical approach to engineering challenges, combining technical expertise with commercial awareness. He is committed to developing teams that work collaboratively and perform at a high level. Jamie has received several industry awards, recognising both his technical capability and his impact on the built environment including ‘Engineer of the Year’ at leading Middle East industry awards.

The post AI Workloads and the Implications for High-Density Data Centre Design appeared first on Data Center POST.

Duos Technologies Group Inc. (Nasdaq: DUOT), through Duos Edge AI, Inc., has announced the deployment of its second Edge Data Center in the Amarillo, Texas market. The new carrier-neutral, SOC 2-compliant facility is located on Potter County land adjacent to the largest colocation facility in the Texas Panhandle, further strengthening digital infrastructure for carriers, healthcare organizations, enterprises, and public sector entities across the region.​

Building on the success of its initial Amarillo deployment, this latest installation expands Duos Edge AI’s footprint in the Panhandle and adds high-density, low-latency computing capabilities for real-time AI applications, enhanced bandwidth, and secure data processing.

“We are proud to deepen our commitment to the Amarillo market with this second deployment, building on the foundation established by our initial EDC, which brought high-performance computing directly to the heart of the Panhandle,” said Dave Irek, Chief Operations Officer of Duos Edge AI. “This expansion enhances capacity and capability in the region, and by partnering on Potter County land adjacent to a premier colocation hub, we are creating a robust, carrier-neutral ecosystem designed to support innovation, attract investment, and drive long-term economic growth.”​

The company said the deployment also helps reduce dependence on data centers located in tier one cities while supporting underserved and high-growth markets across Texas. Duos Edge AI’s broader Texas expansion includes recent installations in Lubbock, Waco, Victoria, Abilene, and Corpus Christi.​

Potter County Judge Nancy Tanner added, “This collaboration with Duos Edge AI represents a significant investment in our community’s future. Positioning this advanced, carrier-neutral data center on county land next to the Panhandle’s largest colocation facility will attract new businesses, improve connectivity for our residents and schools, and position Potter County as a leader in digital infrastructure.”​

The new EDC is expected to be fully operational in the coming months.

To learn more about Duos Edge AI, visit www.duosedge.ai.

The post Duos Edge AI Deploys Second Edge Data Center in Amarillo, Texas Market appeared first on Data Center POST.

How Centralized Infrastructure Intelligence Turns Emergency Replacements into Controlled Operations

Most infrastructure professionals spend their careers building for the planned: capacity expansions, technology refreshes, migration cycles that unfold over quarters or years. And then a Monday morning email changes everything.

A government agency bans equipment from a trusted vendor. A threat intelligence report reveals that a state-sponsored actor has been inside your network switches for eighteen months. A manufacturer announces that the platform running your entire campus backbone loses support in nine months. In each case, the same question emerges: how quickly can you identify every affected device across every facility, and how fast can you replace them without breaking what still works?

For a surprising number of organizations, the honest answer is: they don’t know. That gap between confidence in steady-state operations and readiness for unplanned mass replacement is where real risk lives.

The Forces That Turn Infrastructure Upside Down

Emergency hardware replacement at scale is not hypothetical. Recent years have produced real-world triggers across four broad categories, each with distinct operational implications.

Regulatory and geopolitical mandates. The federal effort to remove Chinese-manufactured telecommunications equipment from American networks—driven by the FCC’s Covered List and Section 889 of the National Defense Authorization Act—has forced carriers and federal contractors into wholesale infrastructure replacement on compliance timelines that don’t flex for budget cycles. The FCC has estimated the total program cost at nearly five billion dollars. Any organization touching federal dollars must verify its infrastructure is clean; if it isn’t, replacement is a compliance obligation, not a planning exercise.

Security crises that outpace patching. The Salt Typhoon campaign revealed that Chinese state-sponsored hackers had penetrated multiple major US telecommunications providers, maintaining persistent access for up to two years—exploiting legacy equipment, unpatched router vulnerabilities, and weak credential management. Investigators found routers with patches available for seven years that had never been applied. For affected carriers, the response demanded physical replacement of compromised infrastructure that could no longer be trusted regardless of patch status. When an adversary achieves sufficient persistence, patching becomes insufficient. Replacement is the only reliable remediation.

End-of-life announcements. Vendor lifecycle decisions create quieter but equally urgent pressure. An organization running multiple hardware platforms faces different end-of-support timelines for each, and dependencies between them mean replacing one can cascade into forced changes elsewhere. Without a consolidated view of what is running, where, and when it loses support, these effects are invisible until they cause failures.

Architectural shifts. Zero trust adoption, SASE frameworks, and cloud-delivered security are rendering entire categories of on-premises equipment architecturally obsolete—not because they’ve failed, but because the security model has moved on. The question is not whether legacy VPN appliances and perimeter firewalls will be replaced, but how quickly, and whether the organization has the visibility to execute in a controlled manner.

Why Standard Processes Break Down

Every mature IT organization has IMAC processes: Install, Move, Add, Change. These handle the predictable rhythm of infrastructure life. Emergency replacement programs share almost none of their characteristics.

They are triggered externally. Their scope is massive—hundreds or thousands of devices across multiple sites. They arrive without allocated budgets or pre-positioned inventory, carrying compliance deadlines indifferent to resource constraints.

The organizations that handle these events well recognize them for what they are: standalone programs needing their own governance, funding, and dedicated teams—and their own information infrastructure. That last requirement is where centralized infrastructure management becomes not a convenience but a prerequisite.

What Centralized Infrastructure Intelligence Must Deliver

Four questions—answered immediately.

What is affected, and where is it? When a regulatory notice references a specific manufacturer, or a security advisory identifies a particular hardware model and firmware version, the operations team needs a definitive count within hours, not weeks. Organizations maintaining a continuously updated centralized inventory—capturing hardware models, firmware versions, physical locations, logical roles, and contractual associations—can answer by running a query. Organizations relying on spreadsheets and periodic audits cannot. The difference in response time is typically measured in weeks, and in a compliance-driven scenario, weeks are what you don’t have. Equally important is dependency mapping: understanding that replacing a core switch will affect upstream routers, downstream access switches, and out-of-band management paths. Without it, a replacement that looks straightforward on paper can produce cascading outages in execution.

What is the replacement path? A legacy switch may need to be replaced by different models depending on port density, power constraints, and compatibility with adjacent equipment. Workflow-driven execution ensures every replacement follows the same approval steps, documentation requirements, and validation procedures—preventing errors that compound in programs spanning hundreds of sites.

Where are we right now? Leadership needs a live view of progress—which sites are lagging, where tasks are stalled, which teams are hitting milestones. This enables resource reallocation, timely escalation of procurement bottlenecks, and an auditable record for regulators. It also surfaces patterns previously invisible: a region that consistently runs behind, or an approval step adding days of unnecessary latency.

What did we learn? Emergency replacements are no longer rare—any organization operating at scale should expect one every few years. Those that conduct structured post-project reviews build a compounding advantage: better scoping templates, more accurate resource models, and pre-validated replacement mappings that make the next response faster.

Building Readiness Before the Next Crisis

Emergency replacements cannot be made painless—they are disruptive, expensive, and stressful regardless of preparation. But the difference between an organization that navigates one in three months and one that takes twelve is almost entirely a function of work done before the trigger.

That preparation has three dimensions: information readiness (a continuously updated inventory with hardware identity, location, firmware status, and dependency relationships), process readiness (defined workflow-driven procedures that activate quickly rather than being reinvented under pressure), and organizational readiness (governance, budget authority, and executive sponsorship that allows an emergency program to stand up as a dedicated initiative).

The organizations best positioned for the next regulatory mandate, zero-day disclosure, or end-of-life cascade are investing in that readiness today—not because they know what the trigger will be, but because they’ve built a discipline prepared for all of them.

# # #

About the Author

Oliver Lindner has over 30 years of experience in IT and the management of IT infrastructures with a focus on data centers. He has worked for many years at FNT Software, a leading provider of integrated software solutions for IT management. In his current position as Director of Product Management, he is responsible for the strategic direction and continuous improvement of the software products for data centers. The aim is to support customers in the efficient and transparent design of their IT infrastructure.

Oliver Lindner attaches great importance to customer focus, innovation and quality. His expertise also includes the development and provision of Software as a Service (SaaS) solutions that offer customers maximum flexibility and efficiency. To this end, he works closely with his own team, partners and customers to create sustainable and innovative software solutions.

The post When Your Data Center Becomes a Liability Overnight appeared first on Data Center POST.

The global data center HVAC market was valued at USD 13.7 billion in 2025 and is estimated to grow at a CAGR of 9.8% to reach USD 36 billion by 2035, according to recent report by Global Market Insights Inc.

Growth in the global data center HVAC industry is being fueled by rising computing intensity, expanding AI-driven workloads, and the continued development of hyperscale and enterprise facilities. As server densities increase and high-performance computing environments generate greater thermal loads, advanced cooling infrastructure has become essential to maintain operational stability and uptime. Research and development efforts across the HVAC industry are increasingly focused on liquid cooling technologies and next-generation thermal management systems capable of handling elevated power densities.

At the same time, stricter regulatory oversight related to energy consumption and environmental performance is encouraging operators to enhance system efficiency and reduce carbon output. ESG-focused initiatives and net-zero commitments are prompting facility upgrades aimed at optimizing Power Usage Effectiveness and lowering operating expenses. Improvements in airflow engineering, adoption of sustainable refrigerants, and integration of energy-efficient cooling architectures are reshaping infrastructure strategies. As regulatory expectations and energy costs continue to rise, demand for intelligent, high-efficiency HVAC solutions in data centers is expected to accelerate significantly.

Rising load capacities, sustainability targets, and regulatory compliance requirements are creating pressure for compact, scalable, and adaptable HVAC systems. Industry participants are responding by designing modular cooling platforms that can operate effectively across diverse geographies while maximizing space utilization and energy performance.

The data center HVAC market from solutions segment accounted for 76% share in 2025 and is forecast to grow at a CAGR of 8.9% from 2026 to 2035. Advanced monitoring tools equipped with artificial intelligence enable predictive maintenance, improve airflow management, and reduce unnecessary power consumption. Increased adoption of liquid-based cooling technologies is supporting high-density server environments while enhancing reliability and extending equipment lifespan through energy-conscious design.

The air-based cooling technologies segment held a 50% share in 2025 and is projected to grow at a CAGR of 8.8% during 2026-2035. Enhanced airflow optimization systems, variable-speed fan configurations, and intelligent environmental controls are improving thermal consistency and minimizing energy waste. Economizer-enabled designs are facilitating greater use of ambient air, while modular cooling units support scalability across both hyperscale and edge environments. Growing server power density is also accelerating interest in direct cooling and immersion-based methods supported by advanced coolant formulations that enhance heat transfer efficiency.

United States data center HVAC market reached USD 4.7 billion in 2025. Increasing cloud integration and AI-intensive applications are driving demand for more efficient cooling architectures. Investments are being supported by electrification incentives and decarbonization initiatives, encouraging broader adoption of intelligent HVAC controls and energy-optimized systems. Integration with smart building platforms and grid-responsive technologies is enabling facilities to manage peak loads, reduce demand charges, and incorporate renewable energy sources.

Key companies operating in the global data center HVAC market include Vertiv, Schneider Electric, Carrier Global, Daikin Industries, Trane Technologies, Johnson Controls, STULZ, Alfa Laval, Danfoss, and Modine Manufacturing. Companies in the global market are strengthening their competitive position through continuous innovation, strategic partnerships, and geographic expansion. Leading players are investing heavily in research and development to enhance liquid cooling efficiency, improve airflow intelligence, and integrate AI-driven monitoring systems. Collaborations with cloud service providers and data center developers are enabling customized cooling deployments for high-density environments. Firms are also expanding manufacturing capacity and regional service networks to support rapid infrastructure growth. Sustainability-focused product development, including low-global-warming-potential refrigerants and energy-efficient system architectures, is becoming a central competitive differentiator.

The post Data Center HVAC Market to Surpass USD 36 Billion by 2035 appeared first on Data Center POST.

As global investment in AI infrastructure, power, and advanced manufacturing accelerates, a critical constraint is coming into sharper focus—project execution.

A newly announced $25 million Series A funding round for Foresight underscores a broader industry shift: while capital continues to flow into large-scale infrastructure, delivering these projects on time and on budget remains a persistent challenge.

The current wave of infrastructure investment is unprecedented in both scale and complexity. Hyperscale data centers, energy systems, and advanced industrial facilities are being developed simultaneously across global markets, often with overlapping supply chains and tight delivery timelines.

However, execution has emerged as a systemic issue.

Research indicates that nearly 90% of large-scale infrastructure projects are completed late or exceed budget expectations. In the context of AI infrastructure, delays can have cascading effects—impacting capacity availability, increasing financing costs, and delaying revenue generation.

Industry observers note that as demand for compute continues to surge, particularly for AI workloads, the margin for error in delivery timelines is shrinking.

A Shift Toward Predictive Delivery Models

Foresight, which positions itself as a predictive project delivery platform, is part of a growing cohort of technology providers aiming to address these execution challenges through data and automation.

The company’s platform is designed to move beyond traditional project management approaches—often reliant on static schedules and retrospective reporting—by introducing continuous validation of project progress and early identification of risk factors.

According to the company, its system enables infrastructure owners to establish baseline schedules more quickly, integrate data across stakeholders, and forecast potential delays before they materialize. Early adopters report improvements in forecast accuracy and reductions in cost overruns.

While such claims reflect a broader trend toward digitization in construction and infrastructure delivery, they also point to a deeper industry need: greater predictability in increasingly complex builds.

Why Execution Matters More in the AI Era

For data center developers and operators, execution risk is becoming more consequential.

Unlike previous infrastructure cycles, AI-driven demand is both immediate and rapidly evolving. Delays in bringing capacity online can result in missed opportunities, strained customer relationships, and competitive disadvantages in key markets.

At the same time, projects are becoming more interdependent. Power availability, equipment procurement, and site development must align precisely—leaving little room for disruption.

This dynamic is prompting a reassessment of how infrastructure projects are planned and managed, with greater emphasis on real-time data, cross-functional visibility, and proactive intervention.

Expanding Beyond Data Centers

Although the initial focus is on sectors such as hyperscale data centers, the challenges associated with project execution are not unique to digital infrastructure.

Foresight plans to expand its platform into adjacent industries, including energy, defense, and advanced manufacturing—areas that share similar characteristics: large capital commitments, complex supply chains, and high sensitivity to delays.

The company’s recent funding, led by Macquarie Capital Venture Capital, reflects investor interest in solutions that address these systemic inefficiencies.

An Industry Inflection Point

The emergence of predictive project delivery tools signals a broader transformation in how infrastructure is built.

For years, innovation in the data center sector has centered on compute performance, cooling technologies, and energy efficiency. Increasingly, attention is shifting toward the process of delivery itself.

As infrastructure programs continue to scale, the ability to execute with precision may become a defining factor in project success.

In an environment where demand is high and timelines are compressed, the question facing the industry is evolving—from whether projects can be financed to whether they can be delivered as planned.

The post Foresight Raises $25M to Tackle Infrastructure Execution Risks in the AI Era appeared first on Data Center POST.

Looking back on a career in IT, I wanted to reflect on the 20-plus years I spent working in and running data centers for Fortune 500 companies in the New York and New Jersey area. This was an exciting time leading both large and small teams through some of the most complex transformations in IT infrastructure. That included designing a trading floor infrastructure for a major bank that was implemented globally, overseeing the merger of two banks with very different IT backbones, driving a mainframe-to-open-systems modernization effort, managing a data center consolidation, and establishing global IT standards.

Today, the challenges to the job are even more profound than transitioning from mainframes to the Internet, digital, mobile, and cloud world. With the advent of AI and explosive data growth from so many more devices and applications, IT infrastructure leaders must rewrite their stories to keep pace.

After moving to the vendor side several years ago and working as a Senior Solutions Architect at Komprise, I get to work with IT leaders daily.  I see just how much the role of the infrastructure or data center director has changed. Here’s how I see the shift with some tips for IT infrastructure directors and executives to stay relevant in their organizations while navigating these cataclysmic shifts in technology and work.

A Shift Toward Complexity and Constant Adaptation

The job of managing data centers and infrastructure has become more multi-faceted. It is no longer just about uptime and physical infrastructure. Directors are now expected to understand a rapidly expanding universe of technologies. There is increased separation of duties and new responsibilities that did not exist 10 years ago. Add in constant security threats, cloud optimization demands, and the exponential growth of unstructured data which requires ensuring that it is accessible where needed, but in a safe, secure manner and the scope of the role expands fast. And while all of this happens, IT budgets are being squeezed. The mandate remains the same: do more with less.

The Unstructured Data Growth Challenge

A resounding pressure point today is storage and the relentless growth of unstructured data. Recent estimates from IDC show that over 80 percent of enterprise data is unstructured, and that volume is expected to reach 291 zettabytes by 2027.

How do you back it all up in a timely way? How do you replicate it for disaster recovery? How do you ensure protection and accessibility? How do you efficiently prepare it for AI ingestion? It has really come down to understanding that all data is not the same, and you must treat data differently so that you can be efficient in your management of the data. Knowing what data you have, where it lives, and what value it offers is now a core competency for any infrastructure leader.

Hybrid IT and Simplification as a Strategy

Over the past few years, I have seen storage and infrastructure strategies shift significantly. The old model of managing everything the same way is obsolete. My approach has always been to keep environments as simple and basic as possible to reduce unnecessary complexity. In today’s typical hybrid IT landscape, that means using tools that are vendor-agnostic, that work across on-prem, outsourced, and cloud environments, and that give you a single dashboard to make informed decisions.

AI, Cost Cutting, and Evolving Job Roles

There is a lot of noise about AI taking over roles in IT. I do not believe that infrastructure managers, storage engineers, or data center professionals should fear for their jobs. However, relying on the status quo is not a strategy. The one thing that I have seen as a necessity for IT personnel is the ability to adjust and evolve as changes have appeared in the IT arena.

One thing is certain; AI is becoming ingrained across the business, and IT must be able to support it across every function. Nearly 90% of enterprises report regular AI use in at least one business function, compared with 78 percent in 2024, according to 2025 research from McKinsey. Learning how to work with AI, understanding its use cases and business applications, and knowing how to prepare the right data for it are key new skills. Equally important is staying current with cloud technologies and security best practices.

Balancing Cost, Security, and AI Readiness

IT leaders are being asked to walk a tightrope. On one side is the need to control cost and ensure security. On the other side is the drive to make data accessible and ready for AI. Yet these demands are interlinked. Cost control and security are critical to ensure that AI ambitions don’t fail or stall. Without security, AI becomes a liability rather than an advantage. The question facing today’s IT directors is along the lines of: “How do we make data more accessible without increasing risk or cost?” Success will come from integrating these requirements, not prioritizing one at the expense of the other.

Why It Is Still an Exciting Time to Work in IT Infrastructure

There is such a tremendous amount of growth in the amount of data being generated, and data has moved from a support function to a true driver of decisions, products, and strategy. Data is now central to every organization, from predicting outcomes, automating decisions, and personalizing experiences in real-time. Add to the fact that both AI and ML have accentuated the value of data, and there’s a lot of opportunity in this area for people who want to grow their careers and remain in IT infrastructure.

The ability to efficiently and strategically manage data and build the right environment for cost control along with flexibility and innovation is a huge need for the enterprise. In our recent industry survey (link) we found that AI data management is a top desired skillset, and organizations are prioritizing hiring individuals who can confidently lead the AI infrastructure discipline.

What’s Ahead for 2026 and Beyond

Looking ahead, I expect infrastructure directors to move beyond managing infrastructure to leading transformation. This means aligning technology with business strategy in areas such as AI integration, cybersecurity, cost control, and workforce development. AI is moving beyond the hype; it’s becoming increasingly relevant in production workflows. Security will continue to be a priority and will need to be addressed. Lastly, bridging the talent gap and reskilling existing workforces should be a focus.

Five Tips for Adapting as a Modern Infrastructure Leader

1. Treat data differently
   Stop managing all data the same way. Understand what is valuable, what is redundant, what is creating undue risks, and what needs to be accessible. Prioritize accordingly.
2. Focus on vendor-agnostic tools
   Choose solutions that work across vendors, technologies and architectures and reduce lock-in. This simplifies operations, reduces cost and delivers better agility.
3. Invest in learning AI concepts
   You do not need to be a data scientist. But you should understand how AI uses data, and how to prepare infrastructure to support it with proper governance.
4. Stay current with security developments
   Security threats evolve constantly. Keep up with best practices and build security into every aspect of data and infrastructure management. Partner with the CSO.
5. Use simplicity as a guiding principle
   Complexity creates risk and inefficiency. Whenever possible, simplify tools, processes, and architectures.

Final Thoughts

The infrastructure director’s role is not what it used to be, and that is a good thing. The scope has grown, the influence has deepened, and the strategic value of IT is clearer than ever. While the challenges are many, so are the opportunities. Those who can adapt, simplify, and lead through change will continue to be essential to their organizations.

# # #

About the Author: 

Paul Romano is a Senior Solutions Architect at Komprise. He has 25 years’ experience at Fortune 100 companies, possessing significant expertise in setting IT direction and policies, data center build outs and migrations, IT architecture, server and endpoint security, penetration testing, establishing productions support standards and guidelines, managing large IT projects and budgets, and integrating new technologies/technology practices into existing environments.

The post The New Demands on Data Center and Storage Leaders appeared first on Data Center POST.

Rob Thornton, President & CEO, International District Energy Association (IDEA)

As the number of data centers grows, so do concerns about location, power access, and grid capacity, especially as AI and cloud computing drive surging electricity demand. Yet, data centers hold an unexpected solution: the waste heat they generate can be harnessed for community benefit.

Captured through district energy systems, this heat can be transformed into a valuable community resource that provides low-carbon warmth, improves grid stability, and redefines data centers as energy partners.

The Power Behind the Numbers

In 2023, data centers accounted for roughly 4.4% of total U.S. electricity use, a share projected to rise to as much as 12% by 2028. As utilities and developers scramble to expand clean generation and transmission, waste heat reuse offers an immediate, scalable way to reduce carbon intensity and ease grid stress.

How Heat Reuse Works

Servers generate heat, which can be captured and directed into district energy networks—insulated pipes transporting hot or chilled water—supplying heat to nearby buildings. This approach reduces the electricity needed for heating and cooling, improving overall efficiency and cutting emissions. In essence, the data center becomes part of a shared local energy ecosystem.

Some add combined heat and power (CHP) systems that produce electricity and heat simultaneously. CHP can increase efficiency for large or urban centers. Two deployment models stand out:

• Urban data centers (10–20 MW): Linked to city energy networks for efficient heat export.
• Large, remote sites (100 MW–1 GW): Feature CHP-based microgrids to serve multiple facilities.

Cities Leading the Way

Areas with dense data center development, such as Northern Virginia’s “Data Center Alley,” are exploring new district heating networks to link excess data center heat with community energy needs. Several pioneering projects in Canada illustrate the potential.

• Markham, Ontario: An Equinix data center retrofitted for heat recovery now warms local condos, a university, schools, and recreation facilities, creating community benefits.
• Toronto, Ontario: Enwave Energy connects Telehouse Canada’s data centers to its system using deep-lake water cooling and waste-heat recovery. This model reduces resource use, enhances cooling, and supports city climate goals.

From Grid Burden to Energy Partner

Heat reuse fundamentally shifts the purpose of data centers from major power consumers to vital contributors in a circular energy economy. By sharing surplus heat, these facilities support decarbonization, reliability, and resilience, and these solutions can be achieved faster than large-scale infrastructure investments.

How Operators Can Get Started

For operators and planners evaluating heat reuse, three clear steps can set the foundation for success:

• First, thoroughly assess the site-level heat export potential for both new builds and retrofits by analyzing available waste heat, proximity to potential heat users, and compatibility with local district energy infrastructure.
• Second, proactively engage municipalities and district energy providers early. This means initiating discussions to align on infrastructure design needs, available incentives, and long-term energy offtake agreements.
• Third, explore hybrid system options—such as pairing CHP, thermal storage, and advanced cooling technologies—for maximum operational flexibility, especially when grid interconnections may be delayed. Evaluate each technology’s potential to complement site-specific requirements and constraints.

As the data economy grows, speed, sustainability, and resilience must move forward together. Waste heat has the potential to be much more than a byproduct; it can become a resource that positions data centers as active agents in community well-being. In the era of AI, shared energy is truly smart energy.

# # #

About the Author:

Rob Thornton is President & CEO of the International District Energy Association (IDEA), a global nonprofit founded in 1909 that advocates for efficient, resilient, and sustainable district energy systems. Under his leadership, IDEA works with public and private partners worldwide to advance energy efficiency, decarbonization, and community-scale thermal networks.

The post From Server Heat to City Warmth: Data Centers’ Hidden Energy Advantage appeared first on Data Center POST.

Duos Technologies Group, Inc. (Nasdaq: DUOT), a leader in intelligent technologies and digital infrastructure, has signed a non-binding letter of intent (LOI) with Hydra Host to deploy a high-density NVIDIA GPU cluster for a leading global technology customer. The project supports a GPU-as-a-Service (GPUaaS) partnership expected to generate approximately $176 million in revenue over a 36-month term, with gross margins exceeding 80% and projected annual EBITDA of more than $40 million.

“We are thrilled to partner with the Duos team on this opportunity,” said Aaron Ginn, CEO and Co-Founder of Hydra Host. “Their ability to deliver immediate access to power combined with an industry-leading deployment speed makes them a standout in the market. We see significant runway ahead as we look to expand our collaboration around colocation and Duos’ High-Power EDC model, which we believe is purpose-built to address a market where demand for AI compute capacity is fundamentally outpacing the speed at which traditional data center supply can be delivered.”

Complementing this milestone, Duos has appointed Doug Recker as Chief Executive Officer, effective April 1, 2026, as the company accelerates its transformation into a focused Edge AI and digital infrastructure platform. Mr. Recker succeeds Chuck Ferry, who will continue to serve on the board of directors.

“This initial customer marks a pivotal step in accelerating the buildout of Duos Edge AI,” said Doug Recker, Chief Executive Officer. “We are now entering an exciting phase of execution, further reinforced by our recently announced LOI with Hydra Host, which underscores growing third-party demand for our distributed AI infrastructure model and validates the scalability of our platform. With secured power, rapid deployment capabilities, and expanding strategic partnerships, we believe Duos is well positioned to pursue high-value infrastructure opportunities. Our focus remains on disciplined expansion, capital-efficient growth, and delivering sustainable long-term value for our shareholders.”

Beyond GPUaaS revenue, the collaboration creates a pathway for approximately $25 million in incremental colocation revenue over the same term, validating Duos’ High-Power Edge Data Center (EDC) business line. The company has also signed a non-binding LOI for a ground lease in Iowa with access to up to 10MW of utility power, advancing its long-term goal of building up to 75MW of distributed capacity.

To learn more about Duos Technologies Group, Inc., visit www.duostechnologies.com.

The post Duos Technologies Signs ~$200M LOI and Appoints Doug Recker as CEO appeared first on Data Center POST.

Datalec Precision Installations (DPI) has introduced its next-generation Data Centre Modularisation Solution, targeting operators that need to add capacity quickly without sacrificing control, reliability or lifecycle value.

Developed in response to surging demand for rapid capacity expansion, the new solution is designed to compress delivery timelines while maintaining full flexibility over configuration, performance and long-term scalability. Each system is precision engineered and manufactured by Datalec to ensure compatibility across structural, mechanical and electrical systems, helping to reduce onsite risk and integration challenges.

Datalec’s modular approach combines pre-engineered design principles with tailored manufacturing, enabling customers to adapt deployments to specific site conditions, operational requirements and growth strategies, including AI-intensive workloads. By shifting more work offsite into a controlled manufacturing environment, the solution minimises disruption associated with traditional construction-led projects and supports safer, more succinct installations and a faster speed to market.

“With organisations under pressure to scale quickly while managing capital expenditure and quality, this launch marks a pivotal shift in how data centre capacity can be delivered,” said John Lever, Director of Modular Solutions at Datalec. “Our modular solution brings these priorities together, giving customers the confidence and agility to develop at the pace their business requires.”

By emphasising reliability, engineering excellence and lifecycle value, Datalec’s new Modularisation Solution reinforces the company’s role in delivering robust, scalable infrastructure for today’s data-driven enterprises and AI-led digital transformation. More information on Datalec’s modular critical infrastructure solutions is available at www.datalecltd.com/critical-infrastructure/modular.

The post Datalec Unveils Next-Generation Modular Data Centre Solution to Accelerate Deployment appeared first on Data Center POST.

The digital landscape is undergoing a fundamental shift as computing power moves closer to the source of data generation. Distributed edge data centers are becoming the backbone of smart cities and critical infrastructure, providing the low latency and high bandwidth required for real-time applications. However, this decentralized model introduces a significant operational challenge. Unlike traditional centralized facilities, these edge sites are often small, unmanned, and located in diverse environments prone to temperature fluctuations and humidity.

A recent milestone in this sector highlights how large-scale organizations are addressing these complexities. A major Mexican municipality recently implemented a Smart IoT data center solution to oversee its growing network of facilities. This initiative serves as a blueprint for how urban centers can maintain resilient infrastructure without the need for constant on-site personnel. By adopting a proactive management strategy, the municipality ensured that its critical data services remain operational regardless of environmental stressors.

The Imperative for Remote Oversight

Data centers of all sizes require perfect working order to support cloud services and business operations. When facilities are distributed across a wide geographic area, the risks associated with equipment failure or unauthorized access increase. Traditional manual inspections are no longer sufficient or cost-effective. An ideal solution must instead rely on a network of sensors that monitor environmental conditions and potential risks in real time.

Environmental factors such as excessive heat, moisture, or even dust can lead to catastrophic hardware failures. Without automated oversight, a minor leak or a failing cooling fan can escalate into a major outage before it is even detected. Consequently, the industry is moving toward a framework that prioritizes predictive maintenance and real-time visibility.

Core Components of a Modern Monitoring Framework

To achieve true resilience, an ideal remote management system should integrate several key technological pillars.

First, connectivity must be both reliable and simple to deploy. Utilizing Long Range Wide Area Network (LoRaWAN) technology allows for long-range, low-power communication between sensors and the IoT gateway that delivers the data to and from the management platform. This approach eliminates the need for complex wiring or extensive new network infrastructure, significantly reducing setup costs and operational overhead.

Second, the sensor array must be comprehensive. Monitoring temperature and humidity is a baseline requirement, but true protection involves detecting a broader range of anomalies. Effective systems incorporate water leak detectors to protect liquid cooling systems and prevent moisture damage. They also utilize vibration sensors to identify mechanical failures in fans or servers before they cease functioning. Air quality and dust sensors are equally vital for maintaining the integrity of cooling systems over the long term.

Third, physical security must be integrated into the environmental monitoring platform. Automated access control sensors and motion detectors allow for the tracking of authorized personnel while immediately alerting operators to unauthorized entries.

From Data Collection to Actionable Intelligence

The value of an IoT solution lies in its ability to transform raw data into immediate action. A centralized dashboard should provide a clear overview of all conditions across the infrastructure. Rather than overwhelming users with information, alerts should be categorized by severity to allow for the quick resolution of the most critical issues.

Customizable thresholds enable users to define the exact parameters for their specific hardware requirements. When these limits are exceeded, the system should trigger instant notifications, allowing for an incident response that prevents impact on operations. Furthermore, maintaining detailed logs of all events is essential for troubleshooting and ensuring compliance with industry regulations.

Ultimately, the goal of modern data center management is to create a system that is as scalable as the data it processes. As organizations expand their footprint, they should be able to integrate additional sensors and facilities seamlessly into their existing monitoring architecture. This intelligent, proactive approach is the future of maintaining reliable and efficient digital infrastructure.

# # #

About the Author

Bjørn Bæra is RAD’s IoT solution manager. In his networking and Industrial IoT career, he served as a product manager for Nvidia’s Spectrum silicon and prior to that as a solution engineer at Cisco, specializing in internet, switching, routing, and management.

The post The Evolution of Remote Data Center Management: Insights from a Modern Municipality appeared first on Data Center POST.

Author: Paulo Campos, President, R&M USA Inc.

U.S. data centers are moving quickly from 100G/200G to 400G and 800G, while preparing for 1.6T. The main driver is AI: training and inference fabrics generate huge east-west (server-to-server) traffic, and any network bottleneck leaves expensive GPUs/accelerators underutilized. Cisco notes that modern AI workloads are “data-intensive” and generate “massive east-west traffic within data centers”.

This step-change is now viable because switching and NIC silicon can deliver much higher bandwidth density. Broadcom’s Tomahawk 5-class devices, for example, support up to 128×400GbE or 64×800GbE in a single chip, enabling higher-radix leaf/spine designs with fewer boxes and links. Optics are improving cost- and power-efficiency as well; a Cisco Live optics session highlights a representative comparison of one 400G module at ~12W versus four 100G modules at ~17W for the same aggregate bandwidth.

In parallel, multi-site “metro cloud” growth is increasing demand for faster data center interconnect (DCI). Coherent pluggables and emerging standards such as OIF 800ZR are making routed IP-over-DWDM architectures more practical for metro DCI.

What this changes

As data centers move to 400G/800G+, the physical layer shifts toward higher-density fiber with tighter loss budgets and stricter operational discipline:

• Parallel optics increase multi-fiber connectivity. Many short-reach 400G links (e.g., 400GBASE-DR4) use four parallel single-mode fiber pairs with 100G PAM4 per lane, which increases the use of MPO/MTP trunking, polarity management and breakout harnesses/cassettes over simple duplex patching. VSFF connectors (for example MMC/SN-MT) are currently becoming an alternative to familiar MTP/MPO connectivity.
• PAM4 is less forgiving. Operators typically specify lower-loss components, reduce mated pairs, and enforce more rigorous inspection and cleaning to protect link margin.
• Single-mode (OS2) expands inside the building. New builds often standardise on OS2 for spine/leaf and any run beyond in-row distances, while copper is largely confined to very short in-rack DACs (with AOCs/AECs or fiber used as lengths increase).
• DCI emphasizes single-mode duplex LC with coherent optics/DWDM, where fiber quality and minimal patching become critical.

The pre-con solution

Pre-connectorized (pre-terminated) cabling systems – including hardened variants – fit current U.S. requirements for speed, performance and repeatability:

• Faster deployment and predictable performance: factory-terminated “plug-and-play” trunks and panels reduce on-site termination, minimize installer variability, and help teams hit tight loss budgets at 400G/800G and beyond.
• Higher density and simpler change control: preterm MPO/MTP trunks with modular panels/cassettes pack more fibers into less space and make adds/changes faster with less disruption.
• Alignment to standards and repeatable architectures: ANSI/TIA-942 defines minimum requirements for data-center infrastructure, while ANSI/BICSI 002-2024 provides widely used best-practice guidance for data-center design and implementation – both encouraging well-defined pathways and modular, repeatable approaches.
• Resilience for harsh pathways: between buildings, in ducts, and at the edge (modular/outdoor DCs), hardened features such as robust pulling grips and improved protection against water/dirt can reduce rework during construction.

As U.S. data centers push into 400G/800G and prepare for 1.6T, pre-connectorized fiber helps deliver deployment speed, high-density layouts, and repeatable, testable performance – often with less reliance on scarce specialist termination labor.

# # #

References

1. Cisco. “AI Networking in Data Centers.” Cisco website. (Accessed Jan 2026).
2. Cisco Live 2025. “400G, 800G, and Terabit Pluggable Optics” (BRKOPT-2699).
3. OIF. “Implementation Agreement for 800ZR Coherent Interfaces (OIF-800ZR-01.0).” Oct 8, 2024.
4. Semiconductor Today. “OIF releases 800ZR coherent interface implementation agreement.” Nov 1, 2024.
5. Ciena. “Standards Update: 200GbE, 400GbE and Beyond.” Jan 29, 2018.
6. TIA. “ANSI/TIA-942 Standard.” TIA Online.
7. BICSI. “ANSI/BICSI 002-2024: The Standard for Data Center Design.” BICSI website.

The post Pre-Connectorized Fiber for 400G/800G and Beyond: Implications for U.S. DCs appeared first on Data Center POST.

Direct-source cooling moves from niche to necessity as AI-era thermal limits collide with traditional airflow design

For decades, server and IT device cooling has followed a predictable playbook: move enough air, manage hot and cold aisles, and rely on increasingly sophisticated fans and facility-level HVAC to keep silicon within tolerance. That model is now approaching its limits.

The rise of AI workloads–characterized by dense computing, high-bandwidth memory, and sustained 24/7 utilization–is forcing a rethink of how heat is removed from systems. The industry is shifting from generalized airflow toward direct-source cooling: targeted, device-level technologies designed to eliminate localized hot spots before they degrade performance or reliability.

2026 will mark a notable turning point as OEM roadmaps, AI-driven performance expectations, and the physical limits of traditional fans converge, making new thermal approaches not optional but inevitable. It will be a pivotal year for system design evolution, with a growing number of manufacturers aligning their roadmaps around architectures that must deliver very high compute horsepower and memory bandwidth to support AI workloads.

As a result, advanced thermal management is emerging as a critical enabler of performance, reliability, and product differentiation across the IT sector.

The Problem: AI Is Breaking the Thermal Envelope

AI-era servers and PCs don’t just run hotter — they also run continuously. Unlike bursty enterprise workloads of the past, AI inference and training systems push CPUs, GPUs, and memory at sustained utilization levels. Heat becomes the dominant constraint.

In practice, this manifests as thermal orphans: localized pockets of trapped heat inside a server or rack that traditional airflow simply can’t reach. When those pockets overheat, the system responds the only way it can: by throttling performance. For data center operators, throttling is not a thermal issue; it’s a business problem. It means paid-for silicon isn’t delivering paid-for performance.

From Airflow to Direct-Source Cooling

The industry needs to supplement, not replace, existing cooling with direct-source airflow applied exactly where heat accumulates. Ventiva’s approach is to add compact ionic modules near problem components, creating just enough directed airflow to clear thermal orphans without redesigning the whole chassis.

Rather than spinning fans faster or redesigning entire racks, system designers can use solid-state, ionic cooling-based solutions that sit close to heat-generating components. These solutions each create airflow by charging ions and using their motion to pull air through a targeted zone.

The result is modest but decisive: 2 to 3 cubic feet per minute (CFM) of airflow, precisely applied, is enough to push trapped hot air out of isolated pockets and back into the main airflow path. That small amount of airflow can be the difference between sustained full performance and permanent throttling.

How Ionic Cooling Works and Why It Matters

With ionic cooling technology, a current is passed through an emitter that ionizes molecules in the surrounding air. Those ions are attracted to an oppositely charged collector, and their movement creates airflow — without any mechanical parts. This has implications enterprises should care about:

• No moving parts means fewer mechanical failures and longer operational life.
• Dust-aware sensing allows the system to detect contamination and trigger automated cleaning, addressing a common failure mode in fans.
• Consistent airflow over time prevents the gradual thermal degradation that shortens component lifespan.

Heat is the fastest way to degrade electronics. By keeping memory and processors within optimal temperature ranges, direct-source cooling doesn’t just improve performance — it improves system longevity.

Performance First, Not Just Efficiency

While energy efficiency is often part of cooling conversations, performance stability is also a key concern. In AI-heavy environments, the worst outcome isn’t higher power draw; it’s unpredictable performance. There are a lot of issues around how high you can go with performance and still deal with the thermal envelopes, such that your system is reliable and can run as required.

By ensuring thermal stability, direct-source cooling allows systems to run at full bore, 24/7, without throttling. For enterprises, this reframes the ROI discussion. Cooling is no longer a facilities cost to be minimized; it’s a performance enabler that protects compute investment.

Fans Are Hitting Their Design Limits

Traditional fan technology is mature, and that’s part of the problem. Incremental gains are getting harder, while fan-based designs face inherent trade-offs. These are: a) higher RPM increases noise and power consumption; b) mechanical wear limits reliability; and c) airflow paths struggle to reach dense, obstructed layouts.

Cold plate and liquid cooling approaches address some of these challenges but add complexity, cost, and service requirements. Ionic cooling occupies a different niche: solid-state, targeted, and augmentative.

Ionic cooling technology isn’t a replacement for fans or liquid cooling. Instead, it fills the gap where traditional methods fail. These include hot spots, edge deployments, and compact systems.

Edge and Client Devices: The Steeper Hill

Ironically, qualifying new cooling technology for laptops and edge devices is more difficult than for data centers. Constrained spaces, lack of physical supervision, dust exposure, and high reliability expectations make these environments unforgiving.

Because there is so much more room in a data center, there’s much more volume of air moving, so you’re not necessarily going to contaminate your data center with, say, pet dander, fibers, or other kinds of dust that you would with a mobile (PC) unit. Edge devices also fall into this category.

Ionic cooling technology has proven particularly well-suited here. Edge devices often run unattended, making mechanical reliability critical. Mini-data center form factors, such as compact AI systems, combine high compute density with limited airflow. Client devices are becoming AI-aware, running inference locally and behaving more like servers than PCs.

As edge systems increasingly process AI workloads on-device, rather than in centralized clouds, they inherit data center-class thermal challenges without data center-class infrastructure.

2026: Why the Timing Matters

2026 is when multiple forces will align. Here’s the evidence:

• OEM “AI-ready” commitments. Major OEMs are locking product release schedules around AI capability. That means more memory, more compute, and higher sustained power.
• Thermal headroom is gone. Existing designs have little margin left. Incremental fan improvements won’t close the gap.
• Market realism. Data center managers are no longer asking if AI workloads will strain cooling but how to prevent performance collapse when they do.

CTO Choices: What to Evaluate Now

For IT and infrastructure buyers planning 2026 and beyond, the cooling decision tree is changing. Key questions include the following:

• Where do performance bottlenecks originate — facility-level airflow or device-level hot spots?
• Is throttling already occurring under sustained AI load?
• Do edge or compact systems lack serviceability or supervision?
• Can targeted airflow extend system life without redesigning the entire rack?

Direct-source ionic cooling technologies such as Ventiva’s don’t replace existing infrastructure, but they can delay costly redesigns, protect performance, and extend hardware ROI.

The Bigger Shift

The transition from fan-centric cooling to hybrid, direct-source approaches mirrors earlier infrastructure shifts. Just as AI forced a rethink of networking, storage, and compute architectures, it is now reshaping thermal design. In that sense, cooling is no longer a background concern. It is becoming a first-class architectural decision–one that will increasingly differentiate AI-ready systems from those that merely claim to be.

2026 is now here, and enterprises that treat cooling as a strategic lever and not an afterthought will be better positioned to extract real value from their AI investments.

# # #

About the Author

Dr. Brian Cumpston is Director of Application Engineering at Ventiva, where he leads the integration of advanced thermal management technologies into consumer electronics and computing platforms. With 25+ years of experience spanning multiple industries, he specializes in the commercialization of disruptive technologies that redefine performance and efficiency standards.

Brian brings a deep background in system architecture and a nuanced understanding of power and performance tradeoffs. He partners with OEMs to solve complex design challenges across acoustics, form factor, and energy efficiency, helping to unlock new possibilities for AI-enabled devices and next-generation platforms.

Brian holds a B.S. in Chemical Engineering from the University of Arizona and a Ph.D. in Chemical Engineering from the Massachusetts Institute of Technology.

The post Why 2026 Will Be a Turning Point for Server Cooling – and What Enterprises Should Know About It appeared first on Data Center POST.

The global data center liquid cooling market was valued at USD 4.8 billion in 2025 and is estimated to grow at a CAGR of 18.2% to reach USD 27.1 billion by 2035, according to a recent report by Global Market Insights Inc.

Rising energy costs, coupled with stringent sustainability requirements, are accelerating the adoption of liquid cooling technologies across data centers. Liquid cooling systems offer significantly lower Power Usage Effectiveness (PUE) ratios ranging from 1.05 to 1.15 compared to 1.4-1.8 for traditional air-cooled facilities, which directly lowers electricity consumption and reduces carbon emissions. Regulatory mandates, including the EU Energy Efficiency Directive, Germany’s Energy Efficiency Act targeting PUE 1.3 by 2027, and California’s energy efficiency standards, are pushing operators toward advanced cooling solutions.

Furthermore, the ability of liquid cooling systems to recover waste heat for district heating or industrial processes transforms data centers into contributors to circular energy economies, supporting corporate net-zero initiatives and enhancing operational sustainability. North America continues to lead the data center liquid cooling market, driven by a dense concentration of hyperscale cloud operators, semiconductor manufacturers, and systems integrators deploying high-density AI and HPC infrastructure.

The solution segment held a 71% share in 2025 and is forecast to grow at a CAGR of 15% from 2026 to 2035. Direct-to-chip cooling is the fastest-growing technology, employing cold plates and micro-channel coolers attached directly to processors, GPUs, and memory to remove 60-80% of heat before it enters the air. These systems circulate coolants such as water with inhibitors or glycol mixtures across chip surfaces, achieving thermal resistances as low as 0.01-0.05°C/W.

The single-phase liquid cooling systems segment reached USD 3.1 billion in 2025. These systems maintain coolant in liquid form throughout the cycle, transferring heat via conduction and convection without phase change. Coolants circulate through cold plates, immersion tanks, or heat exchangers at 18-50°C, depending on design, while facility chillers, dry coolers, or towers remove heat from the loop.

U.S. data center liquid cooling market captured USD 1.29 billion in 2025. Federal initiatives, including AI and HPC programs, semiconductor funding under the CHIPS Act, and defense modernization projects incorporating AI, are key drivers of liquid cooling adoption in public sector data centers.

Leading companies in the data center liquid cooling market include Alfa Laval, Asetek, Boyd, CoolIT Systems, Green Revolution Cooling, LiquidStack, Rittal, Schneider Electric (Motivair), Stulz, and Vertiv. Key strategies adopted by companies in the market focus on technological innovation, such as developing high-efficiency immersion and direct-to-chip cooling solutions for next-generation processors and GPUs. Firms are forming strategic partnerships with hyperscale cloud providers, semiconductor manufacturers, and HPC integrators to expand deployment. Investments in R&D for energy-efficient, modular, and scalable systems strengthen product differentiation. Companies are also emphasizing geographic expansion into emerging markets, supporting sustainability initiatives, and integrating IoT-enabled monitoring tools to optimize performance, enhance reliability, and maintain long-term client relationships.

The post Data Center Liquid Cooling Market to Surpass USD 27.1 Billion by 2035 appeared first on Data Center POST.

The digital infrastructure industry is accelerating rapidly.  New facilities are being built at record speed, and hyperscale customers continue to push for faster deployment timelines.  According to Uptime Institute’s 2024 Staffing Survey, ‘data centers continue to expand their hiring, with 35% reporting more new hires in 2024 compared to 2023.’

Competition has intensified, with Uptime also reporting that ‘57% of organizations increased salary spending in 2024, while only 6% reduced it.’  This aligns with the reality many operators experience: talented technicians routinely receive multiple offers, often at significantly higher pay.

AFCOM’s 2024 State of the Data Center Report underscores how market expansion fuels staffing instability.  The report states, ‘There are nearly 10,000 colocation and wholesale data center facilities across North America, and this number is expected to rise dramatically in the next three years.’

This growth isn’t limited to footprint.  AFCOM adds: ‘New data center builds are expected to multiply sixfold over the next three years,’ contributing to rising wage pressure and aggressive hiring tactics among competitors.

Synergy Research Group provides additional context, noting: ‘The number of hyperscale data centers surpassed 1,000 in early 2024,’ and ‘total hyperscale capacity has doubled in the past four years and is expected to double again in the next four.’

These massive industry shifts increase the demand for qualified operators and heighten turnover risk.  But retaining talent is not solely about matching salary offers, it requires intentional leadership and a strong, supportive culture.

Technicians value trust, recognition, inclusion, and connection.  These emotional drivers often outweigh financial incentives.  Many organizations underestimate their power.  For example, simple cultural investments; providing lunches, snacks, or occasional team entertainment, can materially improve team morale and connection.  These small enhancements help build a positive and inclusive environment.  I have personally turned down external offers because of the strong sense of belonging and support at my current company.

A critical industry-wide shift is also needed.  As operators, we should work collaboratively to reduce extreme market volatility.  Personnel movement will always be part of business but offering 20–30% above market rates just to secure staffing is not sustainable.  More responsible, proactive hiring practices, such as beginning recruitment earlier during buildout and commissioning phases, can help stabilize wage expectations and normalize staffing patterns across the industry.

In one practical example, an engineer on my team appeared increasingly disengaged.  Through direct conversation, I learned he had received an external offer.  Because our leadership team proactively monitored regional wage trends, we were prepared to offer a competitive adjustment and a clear development path.  This combination of financial acknowledgement and emotional investment resulted in his decision to stay, and he eventually became one of our strongest leaders.

Sustainable retention requires long-term systems: continuous market analysis, consistent leadership visibility, structured development pipelines, and meaningful cultural investment.  These practices signal to employees that they are valued, and that their future is worth building inside the organization.

The data center workforce landscape will continue to evolve.  Retention is no longer a transactional process but a strategic, people-centered leadership responsibility.  Organizations that anticipate change, invest in people, and create environments where employees feel valued will be the ones best positioned to thrive in the years ahead.

# # #

About the Author

Tim Shoemaker is a data center operations leader with extensive experience managing mission-critical environments, operational teams, staffing strategies, and organizational change. He has overseen teams through rapid growth cycles, market wage pressures, and major buildouts across multiple data center facilities.  Tim is passionate about developing strong operational cultures that retain talent and reinforce reliability.

The post Leading Through Change: How Data Center Leaders Can Retain Talent in a Competitive Market appeared first on Data Center POST.

Data Center POST had the opportunity to connect with Carlo Malana, President and CEO of STT GDC Philippines, which is a joint venture among Globe Telecom, Ayala Corporation and ST Telemedia Global Data Centres. The company provides secure, reliable, and sustainable data centers to enable digital transformation for global and local businesses. With more than two decades of diverse leadership experience in the ICT industry, his background includes strategic roles at AT&T and as CIO for Globe. He earned a double degree from the University of California at Berkeley and an MBA from Southern Methodist University.

With over 20 years in Information Communications Technology (ICT) including roles with AT&T, across the United States, Mexico, and the Philippines, he has led both technology and business organizations in such diverse areas as strategy, program management, merger integration, retail, finance, customer operations, and sales.

The interview information below has been summarized to provide readers with clarity into who STT GDC Philippines is, what they do and the problems they are solving in the industry.

What does STT GDC Philippines do?  

ST Telemedia Global Data Centres (STT GDC) Philippines empowers business digital transformation through a service model integrating Colocation, Cross connect, and Support Services. We provide Colocation via scalable, sustainable, and secure infrastructure operated to strict global standards, a commitment recently validated by our flagship 124MW STT Fairview Data Center Campus, achieving the IDCA G2 Design Certification, and our STT Cavite 1 data center earning the Uptime Institute Tier III Design Certification. While our Interconnect & Connectivity solutions provide a carrier-neutral platform optimized for seamless access to hybrid and multi-cloud environments, our Support Services complement this technology as your extended technical team, managing critical facility operations so you can focus exclusively on your core business performance.

What problems does STT GDC Philippines solve in the market?

STT GDC Philippines addresses the critical shortage of high-quality digital infrastructure in Southeast Asia (SEA) by replacing outdated systems with massive, scalable facilities built for the future. We solve the capacity shortfall by delivering hyperscale-ready infrastructure, such as our 124MW STT Fairview campus, designed to meet the rigorous TIA-942 Rated 3 and Uptime Institute Tier III standards for concurrent maintainability. We specifically address the urgent demand for AI and high-performance computing by building AI-ready facilities equipped with high power density and advanced liquid cooling support. Most importantly, we eliminate downtime concerns by providing SLA-backed availability, ensuring your mission-critical business operations remain secure and stable 24/7 with a sustainable environment. Finally, we remove connectivity restrictions through our carrier-neutral ecosystem, providing a resilient platform that offers customers superior network choice and the flexibility to connect with the partners that best serve their requirements.

What are STT GDC Philippines’s core products or services?

Our core services are colocation, cross connect, and support services.

What markets do you serve?

ST Telemedia Global Data Centres (STT GDC) Philippines is a leading carrier-neutral provider dedicated to supporting the high-density requirements of Hyperscalers, AI companies, and large enterprises in the banking, financial services, and telecommunications sectors.

As a joint venture between Globe Telecom, Ayala Corporation, and STT GDC, we enable digital transformation by offering scalable, sustainable, and secure infrastructure designed for mission-critical applications. Our facilities are specifically optimized for high-performance workloads, leveraging strategic partnerships with industry leaders and partners to deploy advanced solutions such as liquid cooling for AI-driven demands.

Our data centers provide a flexible technology foundation with direct access to major global cloud platforms and a diverse ecosystem of connectivity partners. This carrier-neutral approach ensures optimal connectivity for hybrid and multi-cloud environments, while our strict operational excellence and 24/7 on-site technical expertise deliver industry-leading uptime. By integrating these best-in-class partnerships, we allow your organization to rely completely on our reliable infrastructure while you focus on driving your core business growth.

What challenges does the global digital infrastructure industry face today?

The industry is currently facing a massive energy and power crisis, where securing reliable electricity has become significantly harder than finding physical land. Because AI operations consume vast amounts of energy, they place an immense strain on local power grids, making it difficult for operators to find suitable locations while sticking to green energy goals.

Secondly, the rapid adoption of AI has created a thermal management challenge; the extreme heat generated by modern high-performance chips exceeds the limits of traditional air cooling, forcing a pivot toward advanced liquid cooling methods even as universal standards remain undefined.

Finally, geopolitical instability and supply chain disruptions are acting as a major brake on progress. Rising global tensions are complicating where secure networks can be built, while acute shortages of essential equipment, like high-voltage transformers and backup generators, are delaying construction and preventing the infrastructure from keeping pace with global demand.

How is STT GDC Philippines adapting to these challenges?

STT GDC Philippines is adapting by building flexible, high-capacity infrastructure, such as the 124 MW STT Fairview Data Center Campus, that is fully ready for AI and liquid cooling but remains adaptable to changing technology rather than being limited to a single purpose. We are addressing the energy challenge by committing to 100% renewable energy for our operations. To navigate global instability, we maintain a fairly neutral position as a carrier-neutral platform, ensuring resilience and open choices for all networks.

What are STT GDC Philippines’s key differentiators?

Our key differentiators begin with our adherence to global standards, ensuring that every facility in our portfolio operates with the same rigor and reliability found across our international platform. This foundation allows us to provide the most extensive capacity in the region, highlighted by the 124MW STT Fairview Data Center Campus, the largest, most interconnected carrier-neutral, and sustainable data center in the Philippines. Our commitment to international, sustainability-driven design is evident in our LEED Gold and TIA-942 Rated 3 certifications, as well as our “AI-ready” infrastructure that supports liquid cooling to reduce environmental impact.

Beyond physical assets, we prioritize our talent through the DC Power Up program, a milestone initiative that trains and certifies the next generation of data center professionals to ensure a future-ready workforce. Our operational excellence is the heartbeat of our business, utilizing advanced automation and AI-powered cooling to maintain peak efficiency 24/7. Finally, we leverage deep local expertise through our powerful partnership with Globe and Ayala, combining the country’s leading telecommunications reach and corporate heritage to provide customers with a seamless, trustworthy gateway into the Philippine digital economy.

What can we expect to see/hear from STT GDC Philippines in the future?  

STT GDC Philippines is focused on rapidly scaling its delivery capabilities, a goal already in motion as we begin operating with our first customers at STT Fairview 1. This marks a significant milestone for what will be the largest and most AI-ready data center campus in the Philippines, featuring infrastructure specifically engineered for high-density computing and advanced liquid cooling. Our commitment to innovation is further showcased at our AI Synergy Lab, where we demonstrate the future of thermal management and high-efficiency power solutions. To support this growth, we are accelerating partnerships across the ecosystem by  recently onboarding key connectivity partners to ensure our facilities serve as the premier, carrier-neutral gateway for Southeast Asia’s digital future.

What upcoming industry events will you be attending? 

We are excited to represent STT GDC Philippines at two of the most influential technology gatherings in the region and the world this year. This February, our team will be in Jakarta for APRICOT 2026, the Asia Pacific region’s premier internet operations and networking summit. This event is a critical forum for us to collaborate with network engineers and policymakers to strengthen the digital fabric of Southeast Asia. Following this, we will be attending NVIDIA GTC in March in San Jose, California. Often called the “Super Bowl of AI,” GTC is where we engage with the latest breakthroughs in AI infrastructure and high-performance computing, ensuring that our data centers remain at the cutting edge of the global AI revolution.

Do you have any recent news you would like us to highlight?

We are excited to share several major milestones that underscore our rapid growth and commitment to the Philippines’ digital future. Most recently, in October 2025, we announced the onboarding of our first connectivity partners at our flagship STT Fairview Data Center campus. These partnerships are significant for our carrier-neutral ecosystem, providing customers with diverse network choices and the resilience needed for AI-powered growth. Additionally, the 124MW STT Fairview Data Center campus recently achieved the prestigious IDCA G2 Design Certification, recognizing its world-class N+1 design and operational excellence. On the sustainability front, we are proud to have transitioned to 100% renewable energy across all our operational data centers as of early 2025.

Is there anything else you would like our readers to know about STT GDC Philippines and capabilities?

Finally, we want your readers to know that STT GDC Philippines is actively pioneering the future of high-performance computing through our AI Synergy Lab. Launched in collaboration with industry leaders, the lab allows enterprises to run actual AI workloads in a controlled environment, providing a live showroom for high-density computing solutions that are essential for modern digital transformation. By bridging the gap between theoretical AI potential and real-world deployment, the AI Synergy Lab ensures that our partners can optimize their hardware configurations for maximum performance and efficiency. This initiative reinforces our commitment to making the Philippines a premier hub for AI innovation in Southeast Asia, providing the specialized environment required to support the next generation of intelligent computing.

Where can our readers learn more about STT GDC Philippines?  

Readers can learn more on our company website, www.sttelemediagdc.com/ph-en.

How can our readers contact STT GDC Philippines? 

You can contact us through Facebook, Linkedin, or our website.

# # #

About Data Center POST

Data Center POST provides a comprehensive view of the digital infrastructure landscape, delivering industry insights into the global data center ecosystem. As the industry’s only peer-contributed and online publication, we offer relevant information from developers, managers, providers, investors, and trendsetters worldwide.

Data Center POST works hard to get the most current information and thought-provoking ideas most apt to add relevance to the success of the data center industry. Stay informed, visit www.datacenterpost.com.

If you are interested in contributing to Data Center POST, contact us at contributions@datacenterpost.com or submit your article here.

Want more digital infrastructure news? Stay in the know and subscribe to Data Center POST today!

The post Company Profile: STT GDC Philippines on Building the Philippines’ Largest AI-Ready Data Center Campus appeared first on Data Center POST.

A happy month of February to you! The big players have dominated the HPC-AI news front of late, here’s a fast (8:44) recap of recent developments, including: Microsoft Maia 200 chip for AI inference, Nvidia H200 (not H20) ....

The post HPC News Bytes 20260202: Microsoft’s Inference Chip, H200 (Not H20) GPUs for China, Mega AI Data Center Deals appeared first on Inside HPC & AI News | High-Performance Computing & Artificial Intelligence.

💾

Renewables and storage could reliably power data centers, but success requires active grids, coordinated planning, and the right mix of technologies. Hitachi Energy CTO, Gerhard Salge, tells pv magazine that holistic approaches ensure technical feasibility, economic viability, and energy system resilience.

As data centers grow in size and complexity, supplying them with cheap and reliable power has never been more pressing. Gerhard Salge, chief technology officer (CTO) at Hitachi Energy, a unit of Japanese conglomerate Hitachi, shed light on the relationship between renewable energy and data center operations, noting that while technically feasible, success requires careful planning, the right infrastructure, and a holistic approach.

“When we look at what's happening in the grids, then renewables are an active element on the power generation side, and the data centers are an active element on the demand side,” Salge told pv magazine. “What you need in addition to that is in the dimensions of flexibility, for which we need storage and a grid that can actively act also here in order to bring all these elements together.”

According to Salge, the key is active grids, not passive systems that simply react to conditions. With more renewables, changing demand patterns, new load centers, and storage options like batteries and existing facilities such as pumped hydro, it is crucial to coordinate these resources actively to maintain supply security, power quality, and cost optimization.

“But when you talk about the impact and the correlation between renewables and data centers, you need always to consider this full scope of the flexibility in a power system of all the elements—demand side, generation side, storage side, and the active grid in between,” he said, noting that weak or congested grids would not serve this purpose.

AI data centers

Salge warned that not all data centers are the same. “There are conventional data centers and AI data centers,” he said. “Conventional data centers are essentially high-load systems with some fluctuations on top. They contain many processors handling requests—from search engines or other applications—so the workload is distributed stochastically across them. This creates a baseline load with random ups and downs, which is the typical load pattern of a conventional data center.”

AI workloads, in contrast, rely heavily on GPUs or AI accelerators, which consume significant power continuously. Unlike conventional data centers, AI data centers often run at sustained high load, sometimes close to maximum capacity for long periods.

“AI data centers are specifically good in doing parallel computing,” Salge explained. “So many of them are triggered with the same demand pattern at the same time, which creates these spikes up and down in the demand profile, and they come in parallel all together.”

These fluctuations challenge both the power supply and the voltage and frequency quality of the connected grid. “So, you need to transport active power from an energy storage system or a supercapacitor to the demand of the AI data center. And that then needs to involve really the control of the data center’s active power. What you need is the interaction between the storage unit and then the AI data center to provide active power or to absorb it afterwards when the peak goes down. That can be also done by a supercapacitor.”

Batteries can store much more energy than supercapacitors, but the latter can ramp smaller energies more frequently. “However, if you put a battery that is smaller than the load, and you really need to cycle the battery through its full capacity, the battery will not survive very long with your data center, because the frequency of these bursts is so high, then you are aging the battery very, very quickly, yeah, so supercapacitors can do more cycles,” Salge emphasized.

He also noted that batteries and supercapacitors are both mature technologies, but the optimal setup—whether one, the other, or a combination with traditional capacitors—depends on storage size, number of racks, voltage levels, and overall system design.

Managing AI training bursts

Salge stressed the importance of complying with grid codes across geographies. “You need to become a good citizen to the power system,” he said. “You have to collaborate with local utilities to make sure that you are not infringing the grid codes and you are not disturbing with the data center back into the grid. A good way to do this, when renewables and data centers are co-located, is to manage renewable energy supply already inside the data center territory. Moreover, having a future-fit developed grid is a clear advantage. Because you have much more of these flexibility elements and the active elements to manage storage and renewable integration and to manage the dynamic loads of the data centers.”

If the grid is not future-fit with modern, actively operating equipment, operators will see significantly more stress. “With holistic planning, instead, you can even use some of the data center flexibility as a controllable and demand response kind of feature,” Salge said, adding that data center operators could coordinate AI training bursts to periods when the power system has more available capacity. This makes the data center a predictable, controllable demand, stressing the grid only when it is prepared.

“In conclusion, regarding technical feasibility: yes, it’s possible, but it requires the right configuration,” Salge said.

Economic feasibility

On economics, Salge believes solar and wind remain the cheapest power sources, even when accounting for the grid flexibility needed to integrate them with data centers. Solar is fastest to deploy, wind complements it well, and both can be scaled in parallel.

“Any increase in data center demand requires investment, whether from renewables or conventional power. Economics depend on the market, and market mechanisms, regulations, and technical grid planning are interconnected, influencing energy flow, pricing, and system stability,” he said.

“We recommend developers to work with all stakeholders—utilities, technology providers, and planners—from the start to ensure reliability, affordability, and social acceptance. Holistic planning avoids reactive fixes and leads to better long-term outcomes,” Salge concluded.

AI is increasingly steering the data center industry toward new operational practices, where automation, analytics and adaptive control are paving the way for “dark” — or lights-out, unstaffed — facilities. Cooling systems, in particular, are leading this shift. Yet despite AI’s positive track record in facility operations, one persistent challenge remains: trust.

In some ways, AI faces a similar challenge to that of commercial aviation several decades ago. Even after airlines had significantly improved reliability and safety performance, making air travel not only faster but also safer than other forms of transportation, it still took time for public perceptions to shift.

That same tension between capability and confidence lies at the heart of the next evolution in data center cooling controls. As AI models — of which there are several — improve in performance, becoming better understood, transparent and explainable, the question is no longer whether AI can manage operations autonomously, but whether the industry is ready to trust it enough to turn off the lights.

AI’s place in cooling controls

Thermal management systems, such as CRAHs, CRACs and airflow management, represent the front line of AI deployment in cooling optimization. Their modular nature enables the incremental adoption of AI controls, providing immediate visibility and measurable efficiency gains in day-to-day operations.

AI can now be applied across four core cooling functions:

• Dynamic setpoint management. Continuously recalibrates temperature, humidity and fan speeds to match load conditions.
• Thermal load forecasting. Predicts shifts in demand and makes adjustments in advance to prevent overcooling or instability.
• Airflow distribution and containment. Uses machine learning to balance hot and cold aisles and stage CRAH/CRAC operations efficiently.
• Fault detection, predictive and prescriptive diagnostics. Identifies coil fouling, fan oscillation, or valve hunting before they degrade performance.

A growing ecosystem of vendors is advancing AI-driven cooling optimization across both air- and water-side applications. Companies such as Vigilent, Siemens, Schneider Electric, Phaidra and Etalytics offer machine learning platforms that integrate with existing building management systems (BMS) or data center infrastructure management (DCIM) systems to enhance thermal management and efficiency.

Siemens’ White Space Cooling Optimization (WSCO) platform applies AI to match CRAH operation with IT load and thermal conditions, while Schneider Electric, through its Motivair acquisition, has expanded into liquid cooling and AI-ready thermal systems for high-density environments. In parallel, hyperscale operators, such as Google and Microsoft, have built proprietary AI engines to fine-tune chiller and CRAH performance in real time. These solutions range from supervisory logic to adaptive, closed-loop control. However, all share a common aim: improve efficiency without compromising compliance with service level agreements (SLAs) or operator oversight.

The scope of AI adoption

While IT cooling optimization has become the most visible frontier, conversations with AI control vendors reveal that most mature deployments still begin at the facility water loop rather than in the computer room. Vendors often start with the mechanical plant and facility water system because these areas present fewer variables, such as temperature differentials, flow rates and pressure setpoints, and can be treated as closed, well-bounded systems.

This makes the water loop a safer proving ground for training and validating algorithms before extending them to computer room air cooling systems, where thermal dynamics are more complex and influenced by containment design, workload variability and external conditions.

Predictive versus prescriptive: the maturity divide

AI in cooling is evolving along a maturity spectrum — from predictive insight to prescriptive guidance and, increasingly, to autonomous control. Table 1 summarizes the functional and operational distinctions among these three stages of AI maturity in data center cooling.

Table 1 Predictive, prescriptive, and autonomous AI in data center cooling

Most deployments today stop at the predictive stage, where AI enhances situational awareness but leaves action to the operator. Achieving full prescriptive control will require not only a deeper technical sophistication but also a shift in mindset.

Technically, it is more difficult to engineer because the system must not only forecast outcomes but also choose and execute safe corrective actions within operational limits. Operationally, it is harder to trust because it challenges long-held norms about accountability and human oversight.

The divide, therefore, is not only technical but also cultural. The shift from informed supervision to algorithmic control is redefining the boundary between automation and authority.

AI’s value and its risks

No matter how advanced the technology becomes, cooling exists for one reason: maintaining environmental stability and meeting SLAs. AI-enhanced monitoring and control systems support operating staff by:

• Predicting and preventing temperature excursions before they affect uptime.
• Detecting system degradation early and enabling timely corrective action.
• Optimizing energy performance under varying load profiles without violating SLA thresholds.

Yet efficiency gains mean little without confidence in system reliability. It is also important to clarify that AI in data center cooling is not a single technology. Control-oriented machine learning models, such as those used to optimize CRAHs, CRACs and chiller plants, operate within physical limits and rely on deterministic sensor data. These differ fundamentally from language-based AI models such as GPT, where “hallucinations” refer to fabricated or contextually inaccurate responses.

At the Uptime Network Fall Americas Fall Conference 2025, several operators raised concerns about AI hallucinations — instances where optimization models generate inaccurate or confusing recommendations from event logs. In control systems, such errors often arise from model drift, sensor faults, or incomplete training data, not from the reasoning failures seen in language-based AI. When a model’s understanding of system behavior falls out of sync with reality, it can misinterpret anomalies as trends, eroding operator confidence faster than it delivers efficiency gains.

The discomfort is not purely technical, it is also human. Many data center operators remain uneasy about letting AI take the controls entirely, even as they acknowledge its potential. In AI’s ascent toward autonomy, trust remains the runway still under construction.

Critically, modern AI control frameworks are being designed with built-in safety, transparency and human oversight. For example, Vigilent, a provider of AI-based optimization controls for data center cooling, reports that its optimizing control switches to “guard mode” whenever it is unable to maintain the data center environment within tolerances. Guard mode brings on additional cooling capacity (at the expense of power consumption) to restore SLA-compliant conditions. Typical examples include rapid drift or temperature hot spots. In addition, there is also a manual override option, which enables the operator to take control through monitoring and event logs.

This layered logic provides operational resiliency by enabling systems to fail safely: guard mode ensures stability, manual override guarantees operator authority, and explainability, via decision-tree logic, keeps every AI action transparent. Even in dark-mode operation, alarms and reasoning remain accessible to operators.

These frameworks directly address one of the primary fears among data center operators: losing visibility into what the system is doing.

Outlook

Gradually, the concept of a dark data center, one operated remotely with minimal on-site staff, has shifted from being an interesting theory to a desirable strategy. In recent years, many infrastructure operators have increased their use of automation and remote-management tools to enhance resiliency and operational flexibility, while also mitigating low staffing levels. Cooling systems, particularly those governed by AI-assisted control, are now central to this operational transformation.

Operational autonomy does not mean abandoning human control; it means achieving reliable operation without the need for constant supervision. Ultimately, a dark data center is not about turning off the lights, it is about turning on trust.

---

The Uptime Intelligence View

AI in thermal management has evolved from an experimental concept into an essential tool, improving efficiency and reliability across data centers. The next step — coordinating facility water, air and IT cooling liquid systems — will define the evolution toward greater operational autonomy. However, the transition to “dark” operation will be as much cultural as it is technical. As explainability, fail-safe modes and manual overrides build operator confidence, AI will gradually shift from being a copilot to autopilot. The technology is advancing rapidly; the question is how quickly operators will adopt it.

The post AI and cooling: toward more automation appeared first on Uptime Institute Blog.