Site Selection Strategy & the Reordered Criteria Hierarchy

For roughly two decades, data center site selection was a settled discipline with a stable ranking of criteria. You screened for latency to population centers and peering hubs, then for land (cheap, flat, contiguous, zoned), then for fiber routes and carrier diversity, then for fiscal incentives (sales-tax exemptions, property-tax abatements), with power treated as a utility commodity you simply ordered. That hierarchy produced Northern Virginia — the largest data center cluster on earth, carrying a large, often-cited (though unverifiable) share of internet traffic — precisely because it optimized the old criteria: dense fiber, proximity to the federal market, generous Virginia incentives, and, once upon a time, abundant cheap power.

The AI era made megawatts energized by a date a binding input wherever grid delivery lags the workload’s need; chips, land and fiber still have independent pass/fail conditions. US generator interconnection queues now hold on the order of 2,000+ GW of requested capacity (LBNL Queued Up: >2,060 GW active end-2025, down from ~2,290 GW end-2024), and on the load side ERCOT's large-load queue alone went from ~63 GW at end-2024 to ~233 GW at end-2025 and ~474 GW of tracked requests by August 2026, about 90% of it data centers. PJM's application-to-commercial-operation timeline stretched from under two years in 2008 to more than eight years by 2025. When firm power is the scarcest, slowest input, investigate it early to avoid spending on sites that cannot energize; workload connectivity, water and lawful land use remain mandatory gates, and only their eligible survivors reach the weighted comparison.

What follows sets out the constrained search: early screening of the longest unresolved dependency, how the workload archetype (carried in from Chapter 1.1) drives the siting search, the single-campus vs distributed-fabric fork, the four build modalities and what each one buys, and the site-selection funnel as mandatory pass/fail/unknown tests, with diligence effort sequenced by cost and the binding need date. A slab cannot be moved once it is poured, so an ordering mistake made here compounds for the life of the asset.

The reordering is not a re-weighting of the same list — it is a change in what gates the deal. In the old model, every criterion was a continuous score and you summed a weighted matrix; a site weak on power could compensate with strong fiber and incentives. In the new model, speed-to-power is a mandatory gate whose evidence should be acquired early when grid delivery controls the need date, and a site that cannot show a credible path to firm capacity by your in-service date is eliminated before its fiber, land, or tax score is ever computed. The matrix still exists — but only for sites that pass power, network, cooling, land and legal gates.

The break is discontinuous rather than gradual because the AI campus load is enormous, dense, and growing faster than transmission can be built. A single hyperscale AI building now draws 100+ MW; a gigawatt campus draws as much as a mid-sized city. The grid cannot absorb that on the old timeline, so power moved from a commodity you order to a multi-year, attrition-ridden process you must win — and the process, not the construction, is the long pole. AI data center construction itself takes only 18–24 months (a single hall can be faster); the grid path behind it takes 3 to 7+ years, and longer in the densest hubs (→ Chapter 3.2). That gap defined 2024–2026 siting, and it is why 'bring-your-own-power' (behind-the-meter gas, batteries, fuel cells, eventually SMRs) moved into first-power comparisons where fuel, permits and availability can close the gap — roughly 90 GW of BTM gas had been announced cumulatively by 2026.

Speed-to-power is the gate because speed is worth more than almost anything else the site decision controls. The arithmetic is direct: a gigawatt of AI capacity generates on the order of $12–13B of revenue per year (SemiAnalysis — a contested, single-source figure; the revenue math lives in Chapter 1.8), so accelerating 200 MW into service six months early is worth roughly $1.2–1.3B of incremental revenue you would otherwise never earn — against a GPU fleet that is depreciating against a contested 2–3-year bear-case economic clock the moment it is racked. A site that is cheaper on power but two years slower to energize is not cheaper; it is a different, worse business.

Workload then shapes how you satisfy the gate. Training-shaped load is flexible and curtailable in ways that shorten the path to energization: a synchronous training job already checkpoints and resumes, so it tolerates non-firm or curtailable interconnections, bridge power, and phased energization in ways an always-on inference SLA cannot. That flexibility is the accelerant — utilities and ISOs are far more willing to fast-track a load that will curtail during system stress (the Duke Nicholas Institute modeled ~98 GW of US headroom available at just 0.5% annual curtailment). Batch inference shares training's throughput-first flexibility and can consume curtailable power. Latency-sensitive online/edge inference needs the firmness its serving SLO demands because unserved load breaches that SLO and loses revenue, which is why it pays the proximity-and-firmness premium. The siting consequence is direct — training and batch inference can use speed-to-power tricks that latency-sensitive serving cannot, so the same megawatt is easier to land for a flexible cluster than for an always-on serving fleet. → mechanics of the queue and flexible interconnection in Chapter 3.2; the energy-supply strategy that backs the gate in Chapter 3.4 and Chapter 3.5.

Once the workload sets the funnel, the next structural fork is concentration vs distribution: do you fight for one enormous interconnection at a single gigawatt campus, or spread the same megawatts across multiple smaller, faster-to-energize sites stitched together by long-haul fiber? The choice sets the network architecture, the redundancy posture, and which workloads the facility can run at all.

A single campus is the natural home of frontier pre-training. One tightly-coupled supercomputer wants its GPUs in one place because synchronous all-reduce collectives run on every training step; test any cross-site design against step time, collective bytes, topology, available bandwidth, and added RTT, then reject it when measured goodput misses the target. Concentration also wins on cost: one substation, one fiber build, one security perimeter, one shared cooling and power yard. The price is concentration risk — you are betting the whole campus on a single interconnection fight, a single grid node's reliability, and a single permitting jurisdiction's goodwill. If that one node slips two years, the entire investment slips with it.

A distributed fabric spreads the bet: several 100–300 MW sites, each with its own faster interconnection, aggregated to gigawatt scale. Multiple medium interconnections in less-constrained markets can energize faster in aggregate than one mega-interconnection in a saturated hub, and the fabric diversifies jurisdiction, grid, and weather risk. But distribution only works for workloads that tolerate the inter-site latency: inference geo-distributes naturally (independent requests), and asynchronous or hierarchical training schemes (multi-datacenter training) can tolerate bounded staleness across sites. Run a strictly-synchronous pre-training job across a distributed fabric and you pay a goodput tax on every step that can erase the speed advantage you bought. The fork therefore loops back to the workload: distribution is a power-and-speed strategy that only synchronous-tolerant or inference workloads can spend.

The third fork is how you take the site to operating capacity — and it is the lever that most directly trades speed against control and capital. The procurement archetypes from Chapter 1.6 map onto siting as four delivery modalities, each landing at a different point on the speed-control-capital surface.

Greenfield self-develop is buying raw land and originating the entire stack yourself — interconnection, substation, shell, power, cooling, fit-out. It gives total control over density, voltage architecture, and cooling modality, and the lowest long-run unit cost at scale, at the price of the longest, queue-gated schedule to a live cluster, the deepest capital commitment, and ownership of every permitting and interconnection risk. It is the right call only for a durable, well-forecast workload where you intend to operate at scale for years. Build-to-suit (BTS) hands the development and construction risk to a developer who builds to your spec on a long (often 15-year) lease — you get a purpose-built liquid-ready hall without fronting the land-and-shell capital, but you inherit the developer's site, including its interconnection position, and you pay a credit-tenant lease premium. Powered shell is the speed-and-optionality middle: a developer delivers an energized, weather-tight building shell with power and water brought to the boundary, and you complete the IT fit-out. It preserves your fit-out and density decisions while letting someone else absorb the slowest, most uncertain part — the interconnection and the shell — and it is the modality that most cleanly converts an irreversible decision into a deferrable one. Colocation (wholesale or retail) rents someone else's already-powered hall, buying a live cluster on the operator's schedule at capex-light terms, at the cost of a shared shell and the least control over the underlying power and cooling design.

The decision is rarely all-or-nothing. The dominant 2026 pattern is a hybrid: anchor durable base load in a self-develop or BTS campus, take a powered shell to preserve a density ramp you have not committed to, and rent colocation or neocloud for burst and bridge capacity while the owned interconnection matures. → the build-vs-buy-vs-rent NPV is scored in Chapter 1.8; the long-lead equipment that gates every modality is in Chapter 2.3.

A defensible site-selection program is a funnel of ordered gates, not a single weighted spreadsheet. The ordering matters as much as the criteria, because the point of a funnel is to fail candidates cheaply on the binding constraint before spending diligence dollars on the rest. In 2026 the binding constraint is almost always firm-power-by-a-date, so it gates first; check the other mandatory constraints in parallel rather than after it. Run the funnel in the old order — score everything, then check power last — and you will spend months of geotech, fiber, and incentive diligence on sites that were never going to energize in time.

• Power — required capacity by its need date (pass/fail/unknown). Is there a credible path to your required firm (or curtailable, for training) megawatts by your in-service date? This means a real read on the local interconnection queue, substation headroom, transmission proximity and voltage class, and any BYOP bridge option. Sites without a path die here, regardless of every other virtue.
• Delivered bill — affordability and contracted structure. Among sites that can energize, what is the all-in cost: nodal/LMP pricing, congestion exposure, capacity and demand charges, curtailment terms, and who pays for network upgrades? In the cited Epoch case, energy is the largest OpEx category, so this is the dominant economic screen among survivors. → Chapter 3.3.
• Network — workload and route-failure budgets, including training and batch. Does the site meet the workload's documented percentile, TTFT, TPOT/inter-token latency, output-length, concurrency, replica-routing, and network-RTT requirements? Those requirements — not a generic radius — decide which power-rich basins remain viable. → Chapter 3.6.
• Water & climate — lawful cooling service. Can you cool here without a water fight? In water-stressed regions this is a gate; model site WUE and annual/peak makeup water, then choose dry/non-evaporative or hybrid/adiabatic heat rejection to meet the site's water budget. → Chapter 3.7.
• Land, geotech, flood, zoning — a buildable, lawful layout. Enough contiguous, developable, appropriately-zoned acreage (GW campuses want 500–1,000+ acres, only ~30–40% built) on bearable soil outside the floodplain. Rarely the binding constraint, but a hard stop when it bites. → Chapter 3.8.
• Fiber, fiscal, social license — service and approval gates before incentives. Carrier-diverse service and the operative community/permitting conditions must pass; incentives enter the survivor comparison using dated qualification, repeal and recapture cash scenarios. → Chapter 3.9, Chapter 3.10, Chapter 3.11.

A failed must-have removes a site; an unknown condition holds the irreversible release until its evidence arrives. The full weighted scoring matrix, hard pass/fail templates, and the stage-gated desktop→field→binding diligence sequence are built out in Chapter 3.13; this chapter identifies the constraints and sends their evidence to that single decision method.

Regional markets package these constraints differently, and the power-first screen is visibly redrawing the map. Capital is moving away from the constrained legacy primaries — Northern Virginia (Dominion/PJM shortfall, zoning reform ending by-right development) and California — toward markets that can actually energize: Texas/ERCOT, where ERCOT’s December 2025 forecast carries data-center load to ~28 GW in 2028 and ~40 GW in 2030 on a single-ISO regulator, the first 765 kV lines in ERCOT (PUCT-approved in April 2025 for the Permian Basin import paths) and SB6’s large-load framework; US secondary markets (Columbus, Salt Lake City, San Antonio, Reno, Indianapolis, the Permian) that offer headroom the primaries have exhausted; the Nordics (firm renewables, average annual temperatures below 10 °C giving up to ~8,000 free-cooling hours a year and facility PUE as low as ~1.09, heat reuse into district heating); and the Gulf (sovereign capital and abundant gas, with FTI Consulting’s August 2025 forecast taking GCC capacity from just over 1 GW to 3.3 GW by 2030).

The power read is geographically specific and time-sensitive in both directions. A market that was ‘full’ for power last year may have opened through a 765 kV line or a bring-your-own-power regime, and a market that looked open may have closed: Dublin’s connection moratorium ran from CRU’s November 2021 direction until the December 2025 policy that now requires matching on-site or local generation, and EirGrid still lists Greater Dublin as fully constrained (May 2026); Singapore paused new data centers from 2019 to 2022 and has reopened only through capped calls (at least 300 MW under the May 2024 Green Data Centre Roadmap). The survivor set moves faster than a five-year siting study can track. Use Chapter 3.13 for the cluster deep-dives and to choose which authority and service territory to investigate; use Chapters 3.2–3.12 to test the actual parcel and contracts. A regional pipeline figure reserves no power for a particular project.

Site selection contains the most irreversible decisions in the entire lifecycle, so the discipline from Chapter 1.1 applies with full force: over-build or hedge the irreversible decisions now; defer the reversible ones and keep them cheap to change.

Irreversible (decide once, at siting): the site itself — you cannot move a slab; the interconnection capacity and voltage class — the queue slot is the single scarcest asset in the project and you do not get it back; the structural and water provisioning of the substrate — floor loading for ~3,000–5,000 lb wet racks, facility water headroom, and pipe-rack space for a density ramp you have not committed to; and the macro climate-and-water basin you have planted in. These are where you spend the option premium — provisioning headroom you may not use, because retrofitting it mid-life is prohibitively costly or impossible.

Reversible (defer, re-decide cheaply): the build modality (a powered shell or colo lease preserves an exit and a fit-out re-decision); the accelerator generation within a fixed power/cooling envelope; the workload mix ratio within an archetype; and the energy-supply tactic layered on top of a secured interconnection (grid-only vs grid-plus-bridge vs hybrid — re-decidable as the BYOP and PPA markets move). The strategic move is to convert irreversible decisions into reversible ones wherever the premium is cheap: a powered shell instead of a full self-build preserves IT-fit-out optionality, and reserving floor loading and water turns an irreversible density ceiling into a deferrable choice.