Why Former Park Land is Attracting Data Center Buyers
Former park land is becoming a hotspot for data centers—here's what that means for the future of urban development.
The digital economy needs a place to thrive — and increasingly, that place used to be a public park.
A recent case out of DeKalb County, Georgia, illustrates exactly how far data center developers are willing to reach for viable land. Documents obtained by WABE show that potential buyers are eyeing former park land in the area, and that county officials are already aware of developer Millsap's interest. It's a story playing out not just in metro Atlanta, but across dozens of American cities where data center demand is colliding headfirst with a simple geographic reality: there's almost no good land left.
That collision is forcing a rethink of what land is "available" — and former green spaces are increasingly ending up in the crosshairs.
The Math Behind the Land Squeeze
Data centers are not subtle tenants. A single hyperscale facility can require 50 to 500+ acres, consume tens of megawatts of power (some pushing past 1 GW at the campus level), and demand proximity to fiber corridors, substations, and major transportation routes. All of that, ideally, within or near urban cores where enterprise clients, cloud tenants, and latency-sensitive applications are concentrated.
The problem is that land meeting all those criteria was largely built out a decade ago. Northern Virginia — still the world's largest data center market — is running so short on developable land that Loudoun County has seen industrial parcels trade at prices once reserved for commercial real estate. Markets like Phoenix, Dallas, and Atlanta moved into secondary corridors only to find those filling up too.
Atlanta, in particular, has become one of the Southeast's hottest data center markets. Low power costs, favorable tax incentives, and connectivity infrastructure have made it a magnet for colocation and hyperscale investment. DeKalb County sits adjacent to the city's core — close enough to matter, large enough to have parcels that haven't yet been spoken for. Former park land, especially land that may have been transferred, decommissioned, or otherwise removed from active public use, represents exactly the kind of opportunity developers are hunting for.
Why Park Land Specifically?
It's worth understanding what makes decommissioned or former park land attractive beyond simple availability.
First, there's the location premium. Parks were typically sited on land that was accessible, relatively flat, and buffered from incompatible uses — characteristics that translate almost directly into a data center site checklist. Flat terrain reduces grading costs. Existing road access cuts infrastructure timelines. Buffers from residential density reduce noise and visual impact complaints.
Former municipal land also tends to sit closer to existing utility infrastructure than raw greenfield sites on a city's exurban fringe — and for data centers, proximity to grid interconnection points isn't just a convenience; it's often the deciding factor between a viable project and a dead one.
Then there are the zoning dynamics. Land that was publicly held often exists in a kind of regulatory gray zone during transition. It may not yet carry the industrial or commercial zoning a data center needs, but local governments dealing with a decommissioned asset they need to monetize can be motivated partners in rezoning conversations. That's different from fighting NIMBYism in an established residential neighborhood or outbidding logistics developers for prized industrial-zoned parcels.
What Local Governments Are Actually Calculating
The DeKalb County situation is instructive because county officials appear to have been tracking the developer interest — not blocking it. That's telling.
Municipalities across the country are increasingly viewing data center development as a high-value, low-impact economic play compared to alternatives. A data center on 50 acres generates substantial property tax revenue, creates construction jobs, and — once operational — employs a relatively small but well-paid technical and facilities workforce. It doesn't generate truck traffic like a distribution warehouse. It doesn't bring retail foot traffic concerns. And unlike a residential development, it doesn't add school-age children to already-strained public school systems.
For a county holding a decommissioned park it can no longer maintain, a data center deal can look like a clean fiscal win on paper. The political calculus gets more complicated when residents learn that a green space they once used is being converted to a secured, fenced industrial facility. That's where community input processes become critical — and where developers who skip the engagement step tend to generate the loudest opposition.
The most sophisticated developers in this space now run parallel tracks: technical site qualification and community stakeholder outreach, simultaneously. The ones who treat community input as a checkbox to be completed after the deal is signed tend to face the consequences in permitting delays and local government cold feet.
Economic Impact: Beyond the Ribbon Cutting
The jobs-and-taxes pitch is real, but it deserves scrutiny.
Data center construction is genuinely labor-intensive. A major facility can generate 1,500 to 2,000 construction jobs over a multi-year build — electricians, ironworkers, HVAC specialists, concrete crews. Those are real, well-compensated jobs with meaningful local economic multipliers.
The operational employment picture is different. A 100MW data center might run day-to-day with 30 to 50 full-time employees. That's not a factory. But those employees tend to earn significantly above median wages, and the facility's infrastructure investment — upgraded substations, fiber, road improvements — often benefits the surrounding area long after the ribbon is cut.
The infrastructure upgrades that come bundled with data center development are underappreciated. In many secondary markets, a data center project is what finally triggers utility investment that the surrounding community has been waiting years for. Grid upgrades, fiber densification, improved road access — developers pay for improvements that stick around.
The flip side of park land conversion is obvious: once a green space is gone, it's gone. The environmental calculus — heat island effects, stormwater management, community health impacts from loss of green space — doesn't show up on a tax revenue spreadsheet. That's not an argument against data center development; it's an argument for honest accounting.
The Sustainability Question That Won't Go Away
Data centers and sustainability have a complicated relationship. Modern hyperscale facilities operated by the major cloud providers have made genuine progress on efficiency — Power Usage Effectiveness (PUE) ratios at leading facilities now approach 1.1, meaning almost no energy is wasted on overhead. Many are running on renewable energy contracts or on-site generation.
But the underlying energy consumption is massive and growing. AI workloads, in particular, are driving power demand that is straining regional grids in ways that weren't anticipated even three years ago. A single AI training cluster can consume more power in a week than a mid-sized hospital uses in a year.
Converting park land to data center use also replaces carbon-sequestering green space with impervious surfaces and significant mechanical infrastructure. For developers serious about sustainability credentials — and increasingly, their institutional investors and hyperscale tenants are demanding those credentials — that's a liability that needs to be offset.
Some developers are starting to address this directly: building vertical green infrastructure into facility design, funding off-site park improvements as mitigation, or pursuing brownfield sites adjacent to former industrial land before going after green space. The developers who figure out how to credibly offset park land conversion will have a distinct advantage as ESG scrutiny on infrastructure assets intensifies.
Where This Goes From Here
The DeKalb County situation is unlikely to be unique. As data center demand continues to grow — driven by AI, cloud adoption, and the general digitization of everything from healthcare to manufacturing — the pressure on urban and peri-urban land will only intensify. Former park land, decommissioned school sites, shuttered golf courses, and closed municipal facilities: all of it is being evaluated.
The developers who will win in this environment aren't necessarily the ones who find the most creative land sources. They're the ones who can move through community engagement, government partnership, and regulatory processes fast enough to stay ahead of competing uses — while building projects that leave the surrounding area measurably better than they found it.
That's a high bar. But in a market where land is the binding constraint, execution on that bar is what separates the deals that close from the ones that generate headlines for the wrong reasons.
**Explore more about data center opportunities and trends on the InfraSale Marketplace!**