Data Center Developer Acquires Land at Surprising Price

A data center developer in Hermantown just paid between 3 and 15 times the assessed value for a piece of property. Read that again. Not 30% over asking. Not even double. We're talking multiples — potentially an order of magnitude above what the county assessor thought the land was worth.

That gap tells you everything you need to know about where data center land prices are heading.

When Assessed Value Becomes Irrelevant

County assessors value land based on comparable sales, current use, and local market conditions. They're not in the business of pricing speculative infrastructure plays. So when a data center developer walks in and pays 3x to 15x what the assessor calculated, it's not because the developer miscalculated — it's because they're operating with entirely different information.

The assessed value of a parcel and its strategic value to a data center operator are two different numbers, and right now, the spread between them is enormous.

What drives that spread? Power. Fiber. Proximity to backbone infrastructure. Zoning that won't require a two-year fight with a planning commission. Cold climates that slash cooling costs. The Hermantown area in northern Minnesota checks several of these boxes — it sits near Duluth, has access to regional power infrastructure, and offers the kind of climate that data center operators dream about. Cooling accounts for roughly 30-40% of a hyperscale data center's operating costs. Every degree of free air cooling is money.

The assessor priced a piece of land. The developer priced an operations platform.

Why Developers Are Writing Checks That Make Assessors Blink

The demand driving these acquisitions isn't coming from traditional commercial real estate pressure. It's coming from AI infrastructure buildout, cloud expansion, and the simple math of data consumption doubling roughly every two years.

Hyperscalers — Microsoft, Google, Amazon, Meta — have announced hundreds of billions in data center capital expenditure over the next several years. They need land. Lots of it. And they need it in specific places: locations with reliable grid access, available power capacity, low latency to population centers, and room to build at scale. Those locations don't grow on trees. When one becomes available, developers pay what they have to pay.

The economics work even at a 15x land premium when the alternative is losing a $500M construction contract because you couldn't secure the site.

There's also a first-mover dynamic at play. In markets like Hermantown — regional, not yet saturated, but with genuine infrastructure advantages — being the first developer to control a prime parcel means you've potentially locked out competitors for years. The land cost becomes trivial relative to the competitive positioning it buys.

This isn't irrational exuberance. It's sophisticated real estate strategy dressed up in what looks, from the outside, like sticker shock.

What This Means If You're Looking at Data Center Real Estate as an Investment

The Hermantown transaction is instructive for investors, but not in the way you might expect. The temptation is to look at a 15x premium and think, "I need to buy land near data centers." The reality is more nuanced.

Location Specificity Is Everything

Not all land near data centers commands a premium. The premium attaches to parcels with specific characteristics: adequate acreage for large-footprint builds (think 50-200+ acres), proximity to high-voltage transmission lines, access to fiber infrastructure, and zoning flexibility. A field ten miles from a substation isn't worth 15x assessed value just because a data center opened nearby. Infrastructure investment returns are hyperlocal.

The Valuation Gap Creates Genuine Opportunity — and Genuine Risk

For sophisticated investors, the disconnect between assessed value and market value in emerging data center corridors represents a real opportunity. If you can identify where developers are likely to acquire next — based on power infrastructure maps, fiber routes, and utility expansion plans — you can potentially acquire land before the premium arrives.

The risk is timing and specificity. Data center site selection is notoriously difficult to predict from the outside. Developers spend months on power studies, environmental assessments, and grid interconnection queues before committing. Buying land speculatively in hopes of a data center acquisition is a concentrated bet with illiquid exposure.

Infrastructure investment at the land level rewards those who understand the technical site requirements, not just the zip code.

The institutional capital that has flowed into data center REITs and development partnerships over the past three years suggests the smart money has already figured this out. Individual investors looking to participate are generally better served by those structures than by direct land speculation.

What Landowners and Developers Should Take From This

If you own agricultural, industrial, or underdeveloped land in a region with strong power infrastructure and fiber access, the Hermantown transaction is worth paying attention to. Property value in these corridors is being repriced by infrastructure demand, not by traditional real estate metrics.

That doesn't mean every landowner should be calling data center developers. But it does mean that if you've been approached about a sale — or if your land sits near a substation, along a fiber route, or in a climate zone favorable to cooling — getting an independent assessment of your property's infrastructure value, not just its assessed value, is worth the conversation.

For developers, the Hermantown deal reflects the broader reality of land acquisition in 2025 and beyond: the sites that check every technical box are scarce, and the cost of losing them to a competitor exceeds the cost of overpaying. That calculus justifies premiums that look shocking on a tax record but make complete sense on a project pro forma.

There's also a strategic timing element. Developers who wait for prices to normalize are likely to keep waiting. The underlying demand drivers — AI compute growth, cloud infrastructure expansion, enterprise digitization — aren't cyclical. They're structural. Land prices in proven data center corridors aren't going to reset to assessed value.

Where Data Center Land Prices Go From Here

The short answer: higher, but unevenly.

Markets that have already absorbed major data center investment — Northern Virginia, Phoenix, Dallas, Columbus — are seeing land prices rationalize simply because there's less buildable land left and longer interconnection queues. The action is shifting to secondary and tertiary markets: the Hermantowns of the world, places that offer genuine infrastructure advantages but haven't yet been fully priced by the market.

That migration is already happening. Utilities in the upper Midwest, Pacific Northwest, and parts of the Southeast are fielding unprecedented volumes of large-load interconnection requests. Where the grid can support it and where land is available, developers are moving fast.

The wildcard is power availability. The United States is facing a genuine grid capacity constraint. Some analysts estimate data centers could consume 9-13% of U.S. electricity by 2030, up from roughly 4% today. Interconnection timelines are stretching to three, four, sometimes five years in congested markets. Land with existing or near-term power availability is effectively a scarce commodity, which means the premium developers are paying in Hermantown today could look conservative in three years.

The gap between assessed value and strategic value isn't closing. For anyone with a stake in infrastructure real estate — as a developer, investor, or landowner — understanding that gap is no longer optional.

Ready to explore the opportunities in data center real estate? Visit InfraSale Marketplace to learn more.