a map line is not powered land

Six hundred contiguous acres, a 345 kV line clipping the northeast corner, a listing that says data center land. That parcel might be worth $30,000 an acre or $3,000, and the line on the map tells you almost nothing about which.

The line is the visible part. What decides the deal is invisible: whether the utility can deliver firm capacity to that point, on what schedule, and at whose cost.

Data center land is a parcel underwritten on power deliverability first — a documented interconnection path, substation headroom, a phased load plan, and a written answer to who pays for network upgrades — and only then on acreage, fiber, water, and entitlements. Proximity to transmission is a screening input, not a qualification.

A map line is not enough.

What follows is the screen that separates strategic inventory from speculative inventory. If you allocate capital to power-adjacent land, it is the screen you should be running. If you own ground near a substation, it is the screen being run on you — worth reading before a broker calls.

the screen: strong signal vs speculative signal

Developers and infrastructure funds look at thousands of parcels and underwrite very few. The filter is not acreage. It is whether power optionality is credible enough to survive a utility engineer's questions.

Land layout is the second filter, and it is more mundane than the brochures suggest. A campus needs contiguous acreage for pads, its own substation, stormwater, equipment yards, laydown space, parking, and security setbacks. Hines uses roughly 3 to 4 MW per acre as a planning ratio — about 3.5 on average — which is an estimate for sizing conversations, not a design spec.

Run that arithmetic and the constraint becomes obvious. A 300 MW campus needs on the order of 85 acres of power-dense footprint. Acres are almost never the scarce thing. Watts are.

Then come the gates that kill projects after the power question looks solved:

• Water. Basin stress and cooling technology decide both feasibility and permit politics. Roughly 43% of data centers sit in high water-stress areas, per S&P Global. The depth is on data centers drink water — ensurance refills the glass.
• Fiber. Diverse physical paths and latency to carrier hotels. Usually screened alongside power, occasionally forgotten on rural sites.
• Community and zoning. Moratoria, farmland identity, noise, and grid cost pass-through now kill projects that pass every power test. Industry trackers count local opposition among the fastest-growing causes of cancellation, though there is no clean national tally and anyone quoting one should show you the methodology. That fight has its own hub: why communities oppose data centers.
• Environmental. Wetlands, listed species, floodplain. This gate does not usually deny a project. It adds years, and years are the expensive part.

the queue is the clock

Ask ten people how long interconnection takes and you get answers between five and ten years, because they are counting different clocks. Two of those clocks are publicly documented — but the caveat has to come first, because it is the one people drop when they quote the numbers.

The well-documented duration data describes generation interconnection queues: how long a power plant waits to connect, not how long a data hall waits to be served. Load queues are newer, less standardized, and much less transparent. So treat what follows as the tempo of the machine your project has to move through, not as your campus's schedule. Quoting it as a campus wait is the same error as quoting a map line as deliverability.

Lawrence Berkeley National Laboratory's Queued Up: 2026 Edition reports that the median generator reaching commercial operation in 2025 took 61 months from interconnection request to operation, up from 36 months in 2015 and 22 months in 2008. For generators, the interconnection process alone — request to executed agreement — ran a median of 45 months. Of everything that requested interconnection between 2000 and 2020, about 19% had reached operation by the end of 2025.

In PJM, projects entering service in 2025 averaged more than seven years: roughly three years to reach an interconnection service agreement, then about four more years to energize, per PJM figures reported by Data Center Knowledge. PJM's cluster-based reform is expected to bring new Cycle 1 agreements down to one or two years.

The second waiting period is the one that got worse recently, and it is the reason the first number misleads. Approval is not power. After an agreement is signed, the site still needs transmission upgrades, a substation, and equipment. Wood Mackenzie's analyst Ben Boucher, quoted by Data Center Knowledge, puts substation transformer lead times at roughly 140 weeks in 2023, about 150 in 2025, and above 160 weeks — more than three years — in 2026.

Which is why "we have an interconnection agreement" is a permission, not a delivery date. PJM's own position is that a project is finished with PJM the moment that agreement is signed; everything after it is construction, procurement, and transmission — none of which the grid operator controls.

Worth naming who is not the villain here. The utility inherited its queue, buys transformers in the same starved market you do, and cannot reprice service outside a formal case. That is exactly why the utility conversation, not the map layer, is the thing you are underwriting.

flexible interconnection is a plan b, not magic

The workaround is real and getting formalized. On June 18, 2026, the Federal Energy Regulatory Commission issued Section 206 show-cause orders to all six RTOs and ISOs, directing each to justify or rewrite how large loads connect — including transmission service built for flexible large loads that can curtail on command, terms for co-located and behind-the-meter generation, and transparency intended to stop cost shifting onto other customers. FERC preliminarily defined a large load as a new commercial or industrial customer at a single site with peak load of 50 MW or greater, interconnecting above 69 kV; if you have been calling your project "a big load" informally, it now has a definition and a docket. Queue-reform depth belongs on entitled is not energized — for the screen, what matters is how much of the clock a fast track actually removes.

Less than the headlines suggest. FERC accepted Southwest Power Pool's High Impact Large Load process on January 14, 2026: for a qualifying load, SPP issues a delivery-point study report within 90 calendar days of an executed study agreement, and a separate expedited process gives generation built specifically to serve that load its own 90-day system impact study. That compresses the study — not the interconnection agreement, not the network upgrades, not the energization date. A faster answer is worth real money. It is not a faster hookup.

Then the limits, because this is where underwriting goes soft:

• Flexible service means you have agreed to be curtailed. That comes with telemetry, automated control, and enforceable operating commitments — and a load factor your financial model may not have assumed.
• On-site generation has its own queue. Turbine order books, gas interconnects, and air permits are not faster by nature; they are just a different line.
• An air permit hearing looks a great deal like a zoning hearing. Behind-the-meter gas can convert a power problem into a community problem.

Plan B moves the schedule. It does not remove the politics.

cost allocation is an underwriting variable

It used to be a footnote. Now it moves basis. The March 2026 Ratepayer Protection Pledge has hyperscalers committing to fund the generation and delivery upgrades their load requires; states are creating separate large-load rate classes with take-or-pay floors; FERC's own orders name cost shifting as a thing to prevent.

Whether that lands as a signed tariff or a press release changes your basis. We wrote the full version — the wholesale evidence, the residential evidence, the tariff mechanics — here: data centers and your electric bill. One paragraph is all it gets on this page.

the half of the screen nobody scores

Look back at the environmental gate. Wetlands delineation, a listed species, a floodplain, the basin the cooling draws from, the working farm across the road that shows up at the hearing. Watersheds, working farmland, and living cover on and beside a power corridor exist whether or not anyone options an interconnect — and the power screen books every one of them as risk, never as something it could fund.

Instruments for holding living systems are not new. Conservation easements, mitigation banks, and land trusts have done this work for decades, and done it well. What is missing is any of it priced into the land deal's basis: the remainder enters the file as a promise made at a hearing rather than a funded line in the underwriting, and promises are what get projects denied. ensurance is that line item — funding named living systems now, in the same basin, while the queue does what the queue does. A certificate funds a specific natural asset; an agent is the onchain account that receives those proceeds and routes them to the place.

Two live agents, and small volumes — we will say so plainly. rural-open-space.ensurance and inland-wetlands.ensurance represent two of the ecosystem types — stocks, in our accounting — that turn up most often on power-adjacent land. Certificate volumes today are small. The mechanism works; the scale is early.

Neighboring farmland has its own post — data centers don't have to eat the farm — and the permit-risk argument lives on the cheapest insurance against a denied permit.

what we can and cannot do

Worth being explicit about our own limits, since the entire post is about not mistaking proximity for capability.

We can: talk the gate — what the screen looks like, which corridors are under conversion pressure, how the living half of a parcel gets funded and held. We find, broker, option, entitle, and subdivide infrastructure land, and route the natural remainder to a place-based agent.

We cannot: sell you a queue position, move an interconnection study, lower a wholesale price, or energize a site. Those belong to the utility, the RTO, and FERC. We do not build hyperscale pads or halls.

We also do not run a production substation-proximity scoring layer. So when anyone — us included — puts a transmission heat map in front of you, the useful question is which utility conversation sits behind each pin. That question is the whole post.

frequently asked questions

what makes land power-ready?

Power-ready land has a documented path to firm electricity: substation adjacency with real headroom, an interconnection study underway or complete, a named upgrade scope, a phased energization plan, and a written answer on who funds network upgrades. Fiber, contiguous acreage, water, and entitlements matter — but they are graded after power, not instead of it.

is land near a transmission line good for a data center?

Sometimes, and proximity alone never proves it. A 345 kV line crossing your property tells you nothing about available capacity at the nearest substation, your position in the utility's study process, or the cost of the upgrades your load would trigger. Land two miles from a substation with headroom and a cooperative utility beats land under a line in a constrained pocket. Proximity is not deliverability.

what is power deliverability?

Power deliverability is the utility's ability to actually serve a specific load at a specific point on a specific schedule — capacity, transmission, upgrade path, energization date, and cost allocation together. It is the first of the three variables infrastructure investors underwrite, alongside development execution and commercial durability of tenant demand. It is a commitment, not a distance.

next steps

If you take one thing from this: grade the parcel on watts, schedule, and cost allocation before you grade it on acres. Then grade the living systems on the same parcel, because they will show up either as a funded commitment or as a hearing.

• Read next: the acre next to the substation — what the million-dollar bid does to a farm, and what a conversion clock actually is.
• Start with the definition: what powered land actually is.
• If you allocate capital here: the infrastructure investor view — including the half of the parcel your power screen skips.
• If you are siting or defending a campus: the data center view — permit friction, water, and the living commitments that hold up at a hearing.