Why Natelli Investments Scrapped a 250MW Data Center

A quarter-gigawatt data center doesn't just walk away from a site without a compelling story behind it.

Natelli Investments LLC recently withdrew plans for a 250MW data center campus in Apex, North Carolina — a project that would have comprised four 200,000-square-foot buildings and represented one of the more substantial hyperscale-adjacent developments proposed for the Research Triangle region. The withdrawal is notable not just for its scale, but for what it signals about the friction now embedded in large-scale data center development across the United States.

This isn't an isolated stumble. It's a data point in a pattern that developers, municipalities, and infrastructure investors are all watching closely.

What Natelli Investments Was Building — and Where

Apex sits at an interesting intersection. It's a fast-growing suburb southwest of Raleigh that has benefited from the Research Triangle's tech economy without carrying the density or infrastructure congestion of Durham or Raleigh proper. For a developer eyeing a greenfield data center campus, the appeal is intuitive: proximity to fiber corridors, a growing regional power grid, and a business-friendly local government that has historically welcomed commercial development.

The proposed project — four buildings totaling roughly 800,000 square feet of data center space — wasn't a hyperscaler's owned campus. Natelli Investments is primarily a real estate development firm, suggesting this was likely structured as a build-to-suit or speculative colocation play, designed to attract large cloud tenants or enterprise anchor customers. At 250MW, the project would have ranked among the largest single campuses in the Southeast.

That scale matters: 250MW is enough to power approximately 185,000 average American homes, and in data center terms, it represents the kind of load that strains regional utility planning horizons.

The Withdrawal: Economic Pressures and Regulatory Headwinds

The source material doesn't enumerate a specific cause, which is itself telling. When developers pull the plug on nine-figure projects and decline to explain why publicly, the reasons are almost never simple.

What we do know from the broader data center development environment is that projects of this scale are running into a convergence of pressures that didn't exist at the same intensity even three years ago.

Power is the central bottleneck. Duke Energy, which serves the Apex area, has been managing an increasingly strained interconnection queue across the Carolinas. Large data center loads — particularly those above 100MW — now face multi-year interconnection timelines and, in some cases, demands for significant transmission infrastructure investment from the developer. A 250MW campus doesn't just plug into the existing grid; it requires dedicated substation infrastructure, transmission upgrades, and utility coordination that can add $50–150 million to project costs before a single server rack is installed.

Regulatory scrutiny has also intensified at the local level. North Carolina municipalities have grown more cautious about approving data center projects after watching neighboring jurisdictions grapple with the tax base implications. Data centers generate significant property tax revenue and construction jobs, but they employ relatively few permanent workers — a campus this size might sustain 50–100 full-time operations staff. That ratio has started to shift community calculus, particularly when the land in question could alternatively support residential development in a region experiencing a housing shortage.

The Permitting Environment Has Changed

Zoning approvals that once moved in 6–9 months are now taking 18–24 months in many jurisdictions, as planning commissions request more detailed environmental impact assessments, water usage studies, and grid impact analyses. For a project still in early planning stages, those timeline extensions directly affect financing terms, equity partner patience, and the ability to hold a site under option without carrying costs that erode returns.

None of this means Natelli's project was unfeasible. It means the margin between feasible and infeasible has compressed, and more projects are falling on the wrong side of that line.

What This Means for Data Center Development Broadly

The data center sector has attracted extraordinary capital over the past four years. Global investment in data center construction exceeded $40 billion annually by 2023, driven by hyperscaler expansion, AI infrastructure buildout, and enterprise cloud migration. That capital is real — but it's increasingly selective.

The projects that are moving forward share a common characteristic: they have power certainty before they break ground. Developers who have secured utility commitments, signed interconnection agreements, or acquired sites adjacent to existing high-voltage infrastructure are winning deals. Those approaching power as a problem to solve after site selection are discovering it's often unsolvable on a timeline that investors will accept.

Apex, NC isn't a uniquely difficult market. But it's also not Northern Virginia, where Dominion Energy and local governments have built a decade-long muscle memory for onboarding 100MW-plus facilities. The Research Triangle is still developing that institutional infrastructure — the utility relationships, the permitting precedents, the workforce pipelines — that makes hyperscale development reliably executable.

That gap matters for data center development news broadly: secondary and tertiary markets will continue attracting developer interest because primary markets like Loudoun County are genuinely capacity-constrained. But translating interest into operating megawatts in those markets requires a level of patience and pre-development investment that not every developer is structured to absorb.

The Local Equation for Apex

From Apex's perspective, losing a 250MW data center campus is a mixed outcome — and local officials would be wise to resist framing it purely as a loss.

The construction phase would have generated significant economic activity: hundreds of construction jobs over a multi-year build, substantial materials procurement, and local contractor engagement. The permanent employment picture is murkier. A campus this size, once operational, typically runs lean — specialized electrical and mechanical technicians, security staff, and a small operations team. It's not a manufacturing plant or a corporate headquarters.

The more significant long-term question is what the site becomes instead. If Natelli retains the land and pursues alternative development, the community may see a mixed-use or residential project that ultimately generates more sustained economic activity and tax revenue per acre. If the land sits idle while the developer reassesses, it's simply a deferred decision.

For communities evaluating data center proposals, the right framework isn't "yes or no" — it's understanding what you're trading and what you're getting in return.

Future projects in Apex or comparable Research Triangle communities would benefit from pre-negotiated development agreements that establish clear expectations around power infrastructure cost-sharing, construction workforce hiring commitments, and community benefit contributions. Several Virginia counties have moved in this direction, and it's made approvals faster, not slower.

Where Data Center Development Goes From Here

The withdrawal of Natelli's Apex project won't slow the industry. Demand for data center capacity is structurally driven by AI training workloads, inference infrastructure, and enterprise digitization at a pace that the existing supply base cannot satisfy. The question is where and how that supply gets built.

A few dynamics are worth watching as infrastructure trends continue to evolve:

Modular and distributed architectures are gaining traction as an alternative to the single-campus 250MW megaproject. Spreading load across multiple smaller facilities in different utility territories reduces both interconnection risk and permitting exposure — at some cost to operational efficiency.

Co-location of data centers with power generation — either utility-scale solar with storage, or in some cases natural gas peakers — is moving from concept to construction in several markets. Developers who can bring their own electrons to the site fundamentally change their negotiating position with utilities and municipalities.

Nuclear is becoming a serious part of the conversation. Microsoft's investment in restarting Three Mile Island's Unit 1 is the most visible example, but smaller developers are beginning to evaluate sites in proximity to operating nuclear plants specifically for data center development.

The projects that get built over the next five years will be the ones that solved power first, treated permitting as a pre-development workstream rather than an afterthought, and brought a community value proposition that extended beyond construction employment. Natelli's withdrawal in Apex is a reminder that the era of relatively frictionless data center development — when a well-located parcel and a letter of intent from a hyperscaler was enough to move a project — is over.

The developers who internalize that reality fastest will build the most megawatts.

Ready to explore more about data center developments and opportunities? Visit our marketplace at [InfraSale Marketplace](https://infrasale.com/marketplace) today!

[INTERNAL LINK: data center trends]

[INTERNAL LINK: infrastructure investment]

[INTERNAL LINK: regulatory challenges]