Quick answer: Data center site selection and feasibility is an independent comparison of candidate locations on the factors that decide whether a data center can be built, energised and filled: power and grid connection timing, land, fiber, water, latency, tax incentives, permitting, tenant demand and risk. DC Market Insights scores every site on one scale, tests the leading site in a feasibility study, and runs it against a downside case before you commit capital. It is used by developers, colocation operators, hyperscalers, investors and the advisers who support them.
Choosing a site used to start with land and fiber. It now starts with power. The International Energy Agency expects data center electricity use to more than double, from 415 TWh in 2024 to around 945 TWh by 2030, and warns that around 20% of planned projects could be delayed by grid constraints. JLL reports that grid connection lead times now exceed four years in primary markets.
The grid backlog is visible in the numbers. Lawrence Berkeley National Laboratory’s Queued Up: 2026 Edition found more than 2,060 GW of generation and storage capacity waiting to connect to the US grid at the end of 2025, with a median of more than five years from interconnection request to commercial operation for projects built in 2025. In Great Britain, the grid operator’s connection reform cut a queue of more than 700 GW and moved data center projects into firm-capacity blocks extending mostly towards 2035 rather than before 2030, according to Data Center Dynamics.
Demand has not slowed to wait for the grid. Vacancy across North America’s primary markets held at 1.4% in the first half of 2026, and Northern Virginia stood at 0.2%, according to CBRE. A site with a credible power date is worth far more than one without, and a feasibility study measures that gap.
DC Market Insights is the data center practice of Credence Research, a research and consulting firm founded in 2015 with 200+ analysts and consultants and 450+ consulting projects a year.
What does data center site selection compare?
It compares nine factors that decide whether a site can be built, energised and filled, and power and grid connection timing comes first because it now sets the timetable for everything else. Each factor is scored on the same scale for every candidate, so the shortlist is a fair comparison rather than a set of separate stories.
1. Power and grid connection timing
Power is the factor that most often turns a good site into a late one. We look at the capacity the local network can deliver, the connection the site holds or has applied for, its place in the queue, and the substation and transmission work it depends on. We also look at realistic on-site and alternative supply, such as gas generation, battery storage or a private wire.
The output is a dated power curve: how many megawatts the site can expect at the meter, and when, with a confidence grade on each step. A connection offer is not the same as an energisation date, and a letter from a utility is not the same as a signed agreement. We separate the three clearly.
2. Land
We check whether the plot is large enough for the campus you plan, including future phases, cooling plant, substations and setbacks. We look at zoning, flood risk, construction access and your control over the land: ownership, option or lease.
3. Fiber and connectivity
We map the fiber routes near each site, the carriers within reach, the distance to the nearest internet exchange and the diverse routes a tenant could order.
4. Water
Cooling design decides how much water a site needs, and local rules decide how much it can take. We look at water supply, abstraction and discharge permits, drought risk and local concern over water use. Where water is scarce, we test whether the cooling design you plan is realistic at that site and what a lower-water design would change in power use and capex.
5. Latency
Latency matters more for some workloads than others. We measure each site’s network distance to the user populations, cloud regions and exchanges that matter for the tenants you want. We score latency against the tenant profile, not against a single standard.
6. Tax incentives
We set out the incentives each location offers: sales tax exemptions on equipment, property tax abatements, enterprise zone benefits and any conditions attached, such as job, investment or efficiency thresholds.
7. Permitting
We map the planning, environmental and building permits a project needs at each site, the authorities involved, the public consultation steps and the record of recent data center applications in that jurisdiction. Where moratoria or new rules on power and water have appeared, we show their likely effect on the timetable.
8. Demand
A site is only worth building if tenants want capacity there. We size demand within reach of each site: hyperscale and cloud expansion, AI training and inference, enterprise outsourcing and edge needs. We set it against operating, under-construction and planned supply in the same market, so you can see whether the gap will still be open when your building is ready.
9. Risk
We bring the remaining risks into one view: natural hazards, political and regulatory change, supply chain lead times, construction labour, and the risk that a competitor secures the same power first. Each risk is scored for likelihood and impact, and the largest ones feed straight into the downside case.
Why is power now the first test for any data center site?
Because demand for grid capacity is rising faster than grids can connect it: US data centers used about 4.4% of the country’s electricity in 2023, and that share could reach 6.7% to 12% by 2028, according to a Lawrence Berkeley National Laboratory report released by the US Department of Energy. That means data center use rising from 176 TWh in 2023 to between 325 and 580 TWh by 2028.
The pressure is not only American. In Ireland, data centres accounted for 23% of all metered electricity consumption in 2025, up from 5% in 2015, according to the Central Statistics Office. When one type of customer takes that share of a national grid, every new connection is examined more closely.
The table below shows the power signals we read for any market before we score individual sites.
| Signal | Recent evidence | What it means for site selection |
|---|---|---|
| Grid connection lead time | More than four years in primary markets (JLL, January 2026) | A site without a firm date is an option, not a project |
| US interconnection queue | More than 2,060 GW of generation and storage waiting at end-2025 (LBNL) | New supply to serve load also waits in line |
| Queue to operation | Median of more than five years for projects built in 2025 (LBNL) | Power plans must allow for long lead times |
| GB connection reform | Queue of more than 700 GW cut; data center projects moved mostly towards 2035 (DCD, December 2025) | UK timetables depend on where a project sits after reform |
| Share of national electricity | 23% of Irish metered electricity in 2025 (CSO) | Expect tighter rules and closer scrutiny |
| Projects at risk | Around 20% of planned projects could be delayed by grid constraints (IEA) | Every shortlist needs a power downside case |
How do we score and compare candidate sites?
We score every site on the same nine factors, on the same 1 to 5 scale, with weights agreed with you before any site is scored. Agreeing the weights first stops the scoring from being bent towards a favourite site after the fact.
The weights depend on what you are building. A hyperscale training campus puts most of its weight on power and land. An interconnection-led colocation site puts more weight on fiber, latency and nearby demand. An edge facility puts weight on latency and permitting. We set the weights with your team in the scoping call and record them in the scope.
| Factor | What scores a 5 | What scores a 1 |
|---|---|---|
| Power and grid timing | Signed connection with a firm date that fits your plan | No application, or a date that depends on unbuilt grid works |
| Land | Controlled plot that fits all planned phases | Plot too small for later phases, or no control |
| Fiber | Several carriers and diverse routes on site or close by | One route, long build to reach a carrier |
| Water | Supply and permits secure for the planned cooling design | Supply at risk, or permits unlikely |
| Latency | Within the network distance your target tenants need | Too far for the workloads you plan to host |
| Tax incentives | Stable incentives that cover your phases | No incentives, or incentives under review |
| Permitting | Clear route with a recent record of approvals | Moratoria, strong opposition or new restrictions |
| Demand | Clear gap between demand and planned supply | Market already well supplied or slowing |
| Risk | Low hazard and regulatory risk | High hazard, or rules likely to change |
Every score comes with its evidence, and any score that rests on an assumption says so on the scorecard.
Some factors work as gates rather than scores. A site that cannot secure power within your window, or cannot be permitted, drops out however well it scores elsewhere, so the shortlist holds only sites that could be built.
What is a data center feasibility study?
A feasibility study is a detailed test of whether the leading site can be built, energised, filled and paid for on the timetable and budget your plan assumes. Site selection tells you which site is best. Feasibility tells you whether the best site is good enough.
A feasibility study covers five questions:
- Power: what the dated power curve looks like, what it depends on and what alternatives exist if the main connection slips
- Demand and lease-up: how much capacity the market will absorb, from which tenant types, and how quickly your building could fill
- Price: what tenants are likely to pay per kW per month for your building’s size and specification, set against comparable facilities
- Build and design: what density and cooling the market now wants, and whether the planned design meets it
- Supply: who else is building or planning capacity in the same market, and when their space will arrive
Pricing varies widely between markets. CBRE’s Global Data Center Trends 2026 put asking rents for 250 to 500 kW requirements at USD 190 to 235 per kW per month in Northern Virginia and USD 235 to 265 in Frankfurt, while Singapore had the highest average asking rent, at USD 403. A feasibility study sets your building’s likely price against the right comparables, not a national average.
Design matters too. The Uptime Institute’s 2026 global survey found the most common rack density reached 11 kW, the highest on record and up from 9 kW in 2025, as reported by Data Center Knowledge. AI racks run far above that. So we test the planned design against the density target tenants actually ask for.
Why do we test the leading site against a downside case?
Because the factors that sink data center projects are rarely the ones in the base case: power arrives late, equipment slips, a competitor fills the market first, or rules change. JLL reports that more than half of projects saw delays of three months or more in 2025 and that equipment lead times have reached 33 weeks, 50% longer than before 2020.
Our downside case changes the inputs that matter most for your site, one at a time and then together:
- Power delay. The connection date slips, or the first phase arrives with less capacity than planned.
- Slower lease-up. Demand in the market grows more slowly, or a large tenant chooses another site.
- Lower price. New supply in the same market arrives faster than expected and pushes rents down.
- Higher build cost. Equipment, labour or grid works cost more, or take longer to deliver.
- Rule change. A permit, incentive or connection rule changes before a later phase.
For each case we show the effect on timetable and revenue, and the mitigations that would protect the project: an alternative power source, a pre-let, a phased design or a reserve site.
Who is site selection and feasibility for?
It is for anyone committing land, power or capital to a new data center location, and the scope changes with the client.
| Client | The question they bring | What the study focuses on |
|---|---|---|
| Developers | Which site can we energise and sell or lease first? | Power timing, land, permitting, demand |
| Colocation operators | Where should our next building go, and will it fill? | Demand, supply, price, connectivity |
| Hyperscale and cloud teams | Where can we get power and land in time? | Power timing, land, incentives, risk |
| Investors and lenders | Does the site support the plan we are funding? | Downside case, lease-up, price |
| Landowners and energy companies | Is our land or power position attractive to data center buyers? | Power, demand, comparable sites |
Developers and operators
Developers and colocation operators need a site that can be delivered on a date tenants will commit to. Over 80% of the capacity under construction in North America’s primary markets was already preleased in the first half of 2026, according to CBRE, so tenants are committing early. A developer that can show a credible power date wins those commitments. See also our page on consulting for operators and developers.
Hyperscale and cloud teams
Hyperscale and AI platforms are limited less by demand than by how quickly a region can deliver power, land and permits. We compare regions and sites on deliverable capacity and timing. See our page for hyperscalers and cloud providers and our work on market entry and power strategy.
Investors and lenders
Funds and banks backing a new build need an outside view of whether the site supports the plan, with a downside case a credit committee can defend. Where the decision is an acquisition rather than a new build, our commercial due diligence service tests the asset in the same way. See also our page for investors.
How does a site selection engagement work?
It runs in five steps, from a scoping call to a findings session with your team, and each step produces something you can use before the next one starts.
- Scope the decision. A call to agree what you plan to build, the regions or candidate sites in play, the weights for each factor, the gates a site must pass and your deadline. You receive a written scope and a quote.
- Screen the long list. Our analysts screen candidate regions or sites against the gates, using our market models, utility and planning records, carrier maps and supply data. Sites that fail a gate are removed with the reason recorded.
- Score the shortlist. Each remaining site is scored on all nine factors, with the evidence for every score. We speak to people who know the market: utilities, brokers, operators, equipment suppliers and tenants.
- Test the leading site. We run the feasibility study on the leading site, or the top two if they are close, and build the base and downside cases.
- Present and stand behind it. A findings session with your team, and with your board, investment committee or lenders if needed. Follow-up questions are answered after delivery.
What you receive
| Deliverable | Format | Used for |
|---|---|---|
| Long-list screen | Short PDF | Removing sites that cannot pass a gate |
| Site scorecard | Excel and PDF | Comparing the shortlist on one scale |
| Feasibility report | PDF with charts and maps | Board, investment committee, lenders |
| Model | Excel with open formulas | Testing your own cases and phases |
| Downside case | Section of the report and model | Setting margin and mitigations |
| Findings session | Video call or on site | Questions from your team |
DC Exclusive: what does our model library add to site selection?
It adds a tested demand and supply baseline for every market on your list, so the study starts from numbers the market already uses. Our library holds 519 published market models, each with history from 2020, a 2025 base year and a forecast to 2035, in global, regional and country editions.
The library is organised in five layers, and each one feeds a different part of a site study:
| Layer | What it covers | How it feeds site selection |
|---|---|---|
| DC Core | Colocation, hyperscale, edge and modular facilities | Demand, supply and price by market |
| Upstream | Power, cooling, construction and equipment | Build cost, lead times and cooling design |
| Downstream | Cloud, AI compute, interconnection and hosting | Where tenant demand comes from |
| Ecosystem | DCIM, security, maintenance and finance | Operating cost assumptions |
| Extended | Adjacent markets in their data center form | Second-order demand near a site |
For example, our Iberia Data Center Market report values that market at USD 11.53 billion in 2025, which gives a developer comparing Madrid and Lisbon a demand baseline before any site is scored. Our UK Data Center Thermal Management Market report values UK cooling spend at USD 993.9 million in 2025, rising to USD 3,398.9 million by 2035, with London and Slough holding more than 65% of the market. That helps a developer judge cooling cost and design expectations in the UK’s busiest cluster.
Three proprietary datasets are in development and will feed future site studies: a Capacity Tracker (live and pipeline MW by country, city and operator), a Colocation Price Index (USD per kW per month by market) and a Deal Tracker (M&A, funding, land and power deals).
How do we make sure the numbers hold up?
We build every demand figure two ways and only use it when the two agree. The same method sits behind our published reports, so the baseline in your site study has already been tested.
- Bottom-up: we count capacity from the supply side: operating, under-construction and planned MW by operator, market by market, with absorption and pricing data.
- Top-down: we work down from the larger market: cloud and AI spending, electricity use and construction spend, and the share that data centers in each market take.
- Reconcile: where the two differ, we find the input that is wrong and fix it, rather than averaging the gap away.
- Time frame: history from 2020, a 2025 base year and a forecast to 2035, so you can see how a market has moved before you trust where it is going.
- Evidence trail: every score and every figure is traced to a document, a dataset, an interview or a stated assumption.
In numbers, a site study sets out for each market the demand in MW, the supply operating and under construction, the gap between them by year, and the share of that gap a new building could realistically win.
Who leads site selection work?
A named senior consultant leads every site study, and a separate editor checks every number before it leaves the firm.
- Priyanka Mor, Senior Consultant. Priyanka leads market entry feasibility, cross-border commercial due diligence and M&A screening engagements, and is our principal methodology reviewer. Her work on TAM, SAM and SOM sizing shapes how we turn market demand into the share a single site could win. She has more than 15 years in market research and consulting.
- Satyabrat Rajawat, Senior Research Analyst, Strategic Advisory. Satyabrat covers power, cooling and rack equipment, the upstream markets that decide build cost, design and lead times. He works from trade data, filings and equipment teardowns, and has contributed to more than 40 syndicated market studies.
- Deepti Agrawal, Senior Editor, Research. Deepti is the final quality gate for our reports and consulting work. She checks models, growth rates and scoring for internal consistency before anything is delivered.
Which DC Market Insights research supports site selection?
Every site study links back to the published reports for the markets on your list, so you can see the baseline before we tailor it. Useful starting points:
- Iberia Data Center Market for demand in Spain and Portugal
- UK Data Center Thermal Management Market for cooling cost and design in the UK
- OCP Rack Market, valued at USD 1.63 billion in 2025 and forecast to reach USD 13.34 billion by 2035, for hyperscale rack and density trends
- The full data center research library, organised in five layers
Related services: market sizing, policy and investment advisory and the full consulting overview.
Frequently asked questions
What does a data center site selection study cost?
Fees are quoted per engagement and depend on the number of candidate sites and markets, the depth of the feasibility study and your deadline; you receive a written quote after the scoping call.
What is the difference between site selection and a feasibility study?
Site selection compares several candidate locations on the same scale and produces a ranked shortlist. A feasibility study tests the leading site in detail: power timing, demand, price, design and supply, with a base and downside case. Most clients need both, and we run them as one engagement.
Do you replace our engineering and power advisers?
No. We cover the market, demand, price, competition and the commercial side of power timing. Engineering, grid studies and detailed design remain with your technical advisers, and we work alongside them so the commercial and technical views agree.
How do you judge whether a grid connection date is credible?
We look at the documents behind the date: the connection offer or agreement, the queue position, and the substation and transmission works it depends on. We check those against utility and regulator records and interviews, and give each step of the power curve a confidence grade.
Which regions can you cover?
Our published reports cover global, regional and country markets across the data center value chain, and our client base spans Europe (45%), the Americas (30%), Asia Pacific (13%) and the Middle East and Africa (12%).
How do we start?
Send us what you plan to build, the regions or sites in play and your deadline through the form below. A senior consultant will reply within one business day to arrange a scoping call.
Sources
- International Energy Agency: Energy and AI, executive summary (2025)
- JLL: Global data center sector to nearly double to 200GW amid AI infrastructure boom (6 January 2026)
- Lawrence Berkeley National Laboratory: Queued Up, 2026 Edition (June 2026)
- Data Center Dynamics: NESO reveals results of UK’s grid connection reforms (9 December 2025)
- CBRE: North American data center demand continues to outpace supply despite record construction activity (8 September 2026)
- US Department of Energy: DOE releases new report evaluating increase in electricity demand from data centers (LBNL) (20 December 2024)
- Central Statistics Office Ireland: Data Centres Metered Electricity Consumption 2025 (7 July 2026)
- CBRE: Global Data Center Trends 2026 (17 June 2026)
- Data Center Knowledge: AI drives data center uncertainty in Uptime’s 2026 survey (31 July 2026)