A Market Study Should Change a Decision

A market study should do more than describe a market. It should change a decision.

That sounds obvious, but a large share of infrastructure research still stops at inventory: project lists, capacity totals, growth rates and maps. Those outputs can be useful, but they become much less useful when the buyer has to decide what to do next.

At PowerlandMap, we have been developing a different format for data center market studies. Each study starts with a decision, not a table of contents. The question may be where capacity is genuinely available, how long power will take, which development projects are credible, or whether a facility can support a high-density AI workload. The research is then built backwards from that decision.

The current catalogue spans nine studies across capacity and leasing, AI infrastructure, power economics and development. They are separate from the PowerlandMap subscription and are designed as one-off research editions for investors, developers, operators, end users and advisers who need a deeper answer to a defined question.

Start with the decision, not the dataset

A study becomes more valuable when the client can state the decision it needs to support.

For a capacity buyer, the question may be: where can I actually secure space, and when? Our Global Data Center Capacity Availability study separates capacity marketed as available today from future capacity, offers where the quantity is not stated, and capacity that is already contracted.

For an investor or developer, the question may instead be: which markets can deliver power on a bankable timeline? The European Time-to-Power & Grid Readiness study compares requested, offered, accepted and energised capacity rather than treating a grid application as equivalent to delivered power.

The difference matters. In data centers, the same number can mean several different things. IT MW, electrical MW, MVA, campus design capacity and available capacity are related, but they are not interchangeable. A serious study must preserve those distinctions.

That principle is consistent with the evidence rules we already apply in PowerlandMap and explain in Can Data Center Market Intelligence Be Trusted?: every material claim should carry a source, a date, a capacity basis and a clear statement of what the evidence does — and does not — prove.

Capacity and leasing: availability is not a single number

The market can look liquid from a distance and become much less liquid once the evidence is tested.

Operators may advertise a campus with substantial capacity while disclosing little about what remains unreserved. A building can be technically complete but commercially committed. A future hall may have a target ready-for-service date without a fully delivered power path.

That is why the capacity study does not try to create one global “vacancy” number. It qualifies evidence market by market and operator by operator.

The objective is practical: help a buyer decide which operators to approach first, what questions to ask, and which apparent options should be removed before commercial discussions begin.

AI infrastructure: separate deployed capacity from the label

AI infrastructure has created a new layer of ambiguity.

“AI-ready”, “HPC-ready” and “high-density” are increasingly common labels, but they do not necessarily describe the same technical capability. The relevant question is not whether a facility uses the right terminology. It is whether the electrical, cooling and operational design can support the workload being considered.

The European AI & HPC Capacity Outlook 2026–2028 separates deployed AI capacity, AI-ready capacity and HPC capacity, with live and target dates kept apart.

The AI Liquid Cooling & High-Density Readiness study goes one level deeper. It is designed to qualify facilities and upgrade paths against a specified AI workload rather than accepting generic readiness claims.

We have also developed a dedicated UAE Data Center & AI Infrastructure Market Study 2026–2035. Its focus is broader: supply, sovereign and hyperscale demand, generation, transmission, electricity economics and project delivery risk across Abu Dhabi, Dubai and the Northern Emirates.

That distinction is important in markets where data-center announcements, AI programmes, grid infrastructure and generation projects are all expanding at the same time. The study keeps IT capacity, electrical capacity, substation ratings and generation figures on their stated basis instead of combining them into a single headline number.

Power economics: compare like with like

Power is usually the first constraint discussed in a data center project and one of the easiest areas to compare badly.

A wholesale electricity price is not the same as a delivered industrial electricity cost. A published tariff is not necessarily the price paid by a hyperscale customer. A PPA price is not directly comparable with a network-delivered cost unless taxes, grid charges, balancing, losses and other elements are treated consistently.

The Data Center Electricity Cost Benchmark is designed around that comparability problem.

The same discipline applies to construction cost. Our Data Center CAPEX & Construction Cost Benchmark separates conventional, modular and AI-ready development costs and focuses on the boundary behind each number.

A €/MW figure without a clear denominator can be misleading. Does it include land? Grid connection? Utility infrastructure? Shell and core? MEP? Commissioning? GPU infrastructure? The answer matters more than the apparent precision of the benchmark.

Development: distinguish pipeline from deliverability

Announced development pipelines are valuable signals, but they are not forecasts.

A project can have land without power, power without planning, planning without a customer, or a strong sponsor without a credible delivery sequence. That is why development research needs to test milestones rather than simply count announced MW.

The European Data Center Development Pipeline 2026–2030 is designed to identify which projects deserve commercial outreach or investment screening by looking at phase, owner, power milestone and delivery dependency.

The European Brownfield & Retrofit Opportunity Study addresses a different question: where existing industrial assets may offer a faster or more efficient route to data center development, and which conversion opportunities justify deeper due diligence.

Brownfield does not automatically mean faster. Existing buildings, electrical infrastructure and industrial zoning can create advantages, but they can also introduce structural, environmental, permitting and phasing constraints. The study is therefore a screening tool, not a substitute for technical due diligence.

The research product and the SaaS product do different jobs

PowerlandMap is first a market-intelligence platform. The platform is designed for continuous screening: projects, operators, power, markets, supply, demand, matching and evidence updates.

A market study is different.

It is a bounded research product built around a defined decision, geography, evidence date and scope. The output can go deeper on one question than a general-purpose dashboard because the research can be commissioned around the client’s exact decision.

The two products therefore complement each other.

A team can use PowerlandMap continuously to monitor markets and identify signals, then commission a study when a specific investment, procurement or development decision requires a deeper evidence pack. Equally, a study can become the starting point for a PowerlandMap subscription when the decision turns into an ongoing monitoring requirement.

What I want every PowerlandMap study to make explicit

There are five things I expect the research to state clearly.

First, what decision the study supports.

Second, what evidence date applies. Market intelligence becomes stale quickly, especially for capacity availability, grid milestones and active development pipelines.

Third, what basis each figure uses. IT MW should not silently become electrical MW, and an unknown quantity should never be converted into zero.

Fourth, what the study cannot prove. A nearby substation does not prove spare capacity. A marketed hall does not prove current vacancy. An announced RFS date is not a delivery guarantee.

Fifth, what the buyer should do next. Good research should narrow the next action, whether that is an operator shortlist, a grid discussion, a site due-diligence programme, an investment screen or a procurement decision.

That is the standard we are applying across the current PowerlandMap study catalogue.

The objective is not to produce more pages. It is to reduce the number of weak assumptions that survive into a real infrastructure decision.

You can review the current PowerlandMap Market Study catalogue, including scope, research briefs and decision frameworks for each edition. If you have a specific market, capacity, power or development question that is not covered by the current catalogue, request access or contact the analyst desk to define the scope.

*Matthieu Gallego*