By Jon Mool
Modernizing colocation and hyperscale facilities to meet AI demand starts with a strategic roadmap
Although the case for modernizing legacy data centers to keep up with AI workloads is clear, the roadmap to get there is often not. This is particularly true for colocation and hyperscale providers managing aging but still profitable data center fleets. For the uninitiated, retrofitting live data centers is a technically complex and operationally sensitive process.
Yet the scale of the retrofit opportunity is hard to ignore. Today, more than 70 percent of global data center capacity resides in existing buildings. A substantial portion of that capacity is underutilized or inefficient.
Despite the ongoing push toward the cloud, the Uptime Institute reports that nearly half of enterprise IT workloads remain in legacy data centers. At the same time, demand for low-latency inference and Edge workloads is growing exponentially.
According to Omdia, AI inference traffic is expected to grow sixfold by 2028. The legacy footprint – already embedded within metro networks and close to end users – represents an overlooked strategic asset for this next wave of infrastructure demand.
Overcoming organizational inertia
The greatest challenge to unlocking that value is not technical – it’s organizational. Some colocation and hyperscale providers are structurally biased toward new construction. Greenfield builds attract investor interest and make for compelling press releases.
In contrast, retrofits can be mistakenly relegated to the ‘deferred maintenance’ category, where they compete for capital with short-term repairs rather than being treated as forward-looking investments.
This bias for the new is reinforced by the way the supply chain is structured. Design firms, consultants, and OEMs are set up to deliver standardized packages. These cookie-cutter templates break down when applied to legacy environments.
OEMs often recommend full-system replacements rather than targeted upgrades, not necessarily because it’s the best option, but because it’s the one they’re set up to support.
The problem of misalignment
Within the data center operation itself, internal misalignment often stymies progress before it starts. Sales teams push for higher density and more sellable kW, while operations teams worry about the risk to uptime.
Finance lacks tools to model retrofit ROI across hybrid or phased deployments. Sustainability teams may not even be in the room. Meanwhile, executive leadership may lack the visibility or framework to compare opportunities across their portfolio.
Siloed decision-making means that even compelling retrofit opportunities can go unexamined. Without a unifying strategy, individual projects stall or are shelved in favor of more straightforward (if less efficient) alternatives.
What a strategic roadmap enables
A unified retrofit roadmap changes this equation. At its best, it serves as a cross-functional guide – one that aligns stakeholders, defines performance objectives, models investment returns, and sequences engineering efforts based on site-specific realities and portfolio-wide goals.
The process starts with internal discovery. This isn’t just about gathering data – it’s about bringing together stakeholders across departments to define shared KPIs. That might mean aligning around EBITDA uplift, sellable kW, or site-level carbon reductions. These conversations often require a third-party facilitator, particularly in organizations where departmental priorities have become entrenched.
Once alignment is in place, the heavy lifting begins: financial and technical scenario modeling. Operators can begin quantifying EBITDA gains from increasing IT load, projecting OPEX savings from modernizing cooling and electrical systems, and evaluating deferred CapEx by extending the life of legacy infrastructure.
These models also incorporate sustainability targets, such as reducing site PUE from, say, 1.8 to 1.4 – or cutting Scope 2 emissions through dramatically reducing consumption. The goal is to build a comprehensive case for transformation, one that integrates financial, technical, and ESG impacts.
From theory to execution
At this point, the roadmap begins to take shape. An example of a typical roadmap progression for an operating data center includes:
Phase 1: Quick wins – simple interventions like adding blanking panels, in-place retrofits of air conditioning units, retro-commissioning and control projects, or deploying in-row cooling systems that can reclaim five-to-ten percent of stranded capacity with minimal disruption.
Phase 2: Targeted capital investment – upgrading UPS systems and power distribution, or central cooling plants to drive returns and optimize capital deployment
Phase 3: Integrating rear-door or direct-to-chip liquid cooling to unlock densities suitable for AI inference clusters.
Every roadmap must also contend with real-world constraints. Lease durations, tenant churn risk, and the remaining lifespan of existing infrastructure all factor into prioritization. Equipment lead times – particularly for large-scale gear – can exceed 50 weeks.
Labor is another bottleneck. Nearly half of data center operators surveyed by the Uptime Institute cite staffing shortages as a primary barrier to retrofit execution.
Despite these challenges, the payoff can be significant. Retrofitted data centers often deliver 30-40 percent improvements in energy efficiency. The power saved can be repurposed for IT applications in these low-latency legacy facilities located in or near urban centers primed to serve AI inference needs.
Modern cooling strategies also reduce water consumption, while the reuse of existing infrastructure avoids the embodied carbon associated with new builds. These ESG gains often align directly with the sustainability mandates of institutional investors – but without a structured roadmap, they rarely make it into the business case.
Why it matters now
At Enabled Energy, we call these roadmaps NextField – a transition from brownfield. It’s a process-driven strategy for unlocking value in brownfield assets – not just to extend their useful life, but to reposition them for the AI-powered future.
NextField isn’t about chasing the latest technology for its own sake. It’s about careful, coordinated planning that respects the complexity of each site and the priorities of every stakeholder involved.
The result is a transformed facility with improved uptime, higher density, higher revenue, lower carbon intensity, and dramatically better economics. In a market where new capacity is increasingly difficult and expensive to bring online, retrofitting may be the highest-return investment available to digital infrastructure owners.
It’s time to stop viewing aging facilities as a liability. With the right roadmap, they can become your most strategic assets.
Learn more about how Enabled Energy can turn your brownfield facilities into a thriving NextField fleet.
This article was originally published on DataCenterDynamics on August 12, 2025.
EN 
FR