Will 24 MW Change How We Think About Data Center Power?

A single power agreement rarely makes headlines. But when a data center developer secures a will-serve commitment for 24 megawatts from a regional utility—in a location like northern Maine—it signals something worth paying attention to.

The Loring LiquidCool Data Center has locked in a will-serve agreement with Versant Power to deliver 24 MW of power to a warehouse facility leased by its developer. That's the core fact. What it means for data center power supply strategy, energy infrastructure planning, and the broader push toward efficient compute is the more interesting story.

What Loring LiquidCool Is Actually Building

Loring isn't a household name in the data center world—yet. The project is based at the former Loring Air Force Base in Limestone, Maine, a site that has spent decades searching for its post-military identity. Industrial reuse projects on former military land are notoriously difficult to execute. Infrastructure is aging, power access is uncertain, and the economic base needed to support large facilities often isn't there.

A 24 MW power commitment changes that calculus significantly.

Securing a will-serve agreement from Versant Power isn't just a procurement win—it's a site credibility event. It tells prospective tenants, investors, and co-location customers that the fundamental input—reliable, utility-backed electricity—is spoken for. In a market where data center developers routinely spend 18 to 36 months navigating interconnection queues and utility negotiations, having that agreement in hand compresses the risk timeline considerably.

The "LiquidCool" designation in the project name is also worth unpacking. Liquid cooling in data centers isn't new, but its adoption has accelerated sharply as AI workloads and high-density GPU clusters push rack power densities beyond what traditional air cooling can handle. Where a conventional enterprise data center might see 5–10 kilowatts per rack, modern AI inference and training environments regularly exceed 40–80 kW per rack. Air simply can't move heat fast enough at those densities. Liquid cooling—whether through direct liquid cooling, immersion, or rear-door heat exchangers—solves that problem at the physics level, not the engineering workaround level.

Versant Power's Role and What 24 MW Means in Practice

Versant Power serves roughly 165,000 customers across northern and eastern Maine. It's not a massive investor-owned utility, which is precisely what makes this agreement notable. Regional utilities of this scale don't casually commit 24 MW to a single customer—that's a meaningful slice of local grid capacity, and it requires confidence in both the project's viability and the infrastructure to support it.

For context: 24 MW is enough power to run approximately 8,000 average American homes simultaneously. For a data center, that same 24 MW translates into serious computational horsepower—particularly when the facility is designed around liquid cooling, which allows operators to extract more useful compute per watt than air-cooled alternatives.

The will-serve agreement is also a signal about where Versant Power sees growth coming from—and it's not residential rooftops.

Maine's grid has historically been shaped by paper mills, manufacturing, and residential demand. As those industrial anchors have declined, utilities like Versant have been quietly repositioning. A 24 MW data center customer represents a stable, long-duration load profile—the kind utilities love because it's predictable and doesn't disappear when commodity prices shift. From Versant's perspective, this is exactly the type of anchor load that justifies infrastructure investment and keeps rates stable for everyone on the system.

Why Data Center Efficiency Starts With Power Strategy

Here's the non-obvious angle most coverage misses: data center efficiency isn't primarily an engineering problem. It's a power strategy problem.

The most sophisticated cooling system in the world underperforms if the facility is drawing power from an unstable grid, paying peak demand charges that erode margins, or operating in a region where renewable energy access is limited. Where you plug in matters as much as how efficiently you run once you're plugged in.

Northern Maine has attributes that don't show up on typical site selection scorecards. The climate is cold—average temperatures that dramatically reduce the need for mechanical cooling, which in a liquid-cooled facility translates into lower Power Usage Effectiveness (PUE) numbers. The region also sits within reach of substantial hydroelectric and wind resources, both of which are increasingly important to hyperscale and enterprise customers with net-zero commitments.

A 24 MW facility operating at a PUE of 1.2 (achievable with liquid cooling in a cold climate) consumes 20 MW of IT load. Compare that to a facility in a warm climate relying on air cooling at a PUE of 1.6, which would need nearly 33 MW to deliver the same compute capacity. That 13 MW gap—at commercial electricity rates over a decade—is the difference between a profitable data center business and a marginal one.

Where Liquid Cooling Takes the Industry Next

The Loring project sits at the intersection of two trends that are reshaping data center infrastructure: the demand for AI-capable compute density and the pressure to decarbonize operations.

Liquid cooling enables both. Higher density means more revenue per square foot. Better thermal management means lower energy waste. And when you layer in a power supply agreement in a region with access to cleaner generation sources, the sustainability math starts working in your favor without heroic effort.

The industry is moving toward a world where data center power supply agreements are negotiated with the same rigor as real estate leases—because the two are equally foundational to project success.

What Loring LiquidCool is demonstrating, intentionally or not, is a model worth watching: identify underutilized infrastructure with good bones (a former air base with existing buildings and land), secure utility commitment before breaking ground, and design around cooling technology that scales with the workloads that actually matter in 2025 and beyond.

The developers who figure this out first—who treat power procurement as a first-principle design decision rather than a late-stage checkbox—will have a structural cost and reliability advantage that's very hard for competitors to replicate. Grid interconnection timelines are measured in years. You can't just add 24 MW when you need it.

The Loring agreement won't reshape the entire data center industry overnight. But it's a concrete, specific example of how the next generation of compute infrastructure gets built: deliberately, starting with power, and designed around the thermal realities of modern workloads rather than the assumptions of a decade ago.

Anyone building, investing in, or operating data centers should be paying close attention to how deals like this one get structured—and what comes next when the facility goes live.

Call to Action: Ready to explore the future of data center power? Visit InfraSale Marketplace for insights and opportunities.

[INTERNAL LINK: data center power supply strategy]

[INTERNAL LINK: liquid cooling technology]

[INTERNAL LINK: energy infrastructure planning]