Spain's Data Centre Market Needs Two Shortlists

Spain rewards two different commercial approaches. A customer seeking a small block of capacity next year needs evidence of available halls and a credible handover. A developer preparing a large campus for the end of the decade needs a staged power, permitting and financing plan. Putting both searches into one national megawatt total hides the distinction that matters most.

A PowerlandMap database snapshot taken on 30 October 2026 makes that distinction visible. The Spanish inventory contains 235 published, non-demo public-supply records. Of these, 218 carry the verified-public flag. Within that bounded subset, 131 are classified as operating and 87 are at other stages. Only 28 have a populated, positive IT-capacity field.

These are counts of database records, not a census of unique physical campuses or a measure of vacant supply. Their value is that they make it possible to organise a search, identify the missing evidence and choose the right next conversation. The underlying coverage methodology matters as much as the headline number.

What the Spanish records actually show

The current stage distribution is useful before any capacity aggregation.

The last group contains five pre-development records, three announced, three approved, three in planning, one in site preparation and one on hold. It is deliberately not described as committed future supply.

The 28 populated IT values represent 12.8% of the 218 records. The remaining 190 do not have a positive value in that field. That is a disclosure and reconciliation gap within this extraction, not evidence that those facilities have no capacity. Nor does a populated field establish current availability. Some values describe a planned full build, while others describe an operating facility or a phase.

My interpretation is that Spain should be screened through separate delivery horizons before anyone ranks sites by size. This avoids two expensive mistakes. The first is overlooking an operating location because its public capacity is missing. The second is treating a well-documented future campus as an answer to an immediate customer requirement.

Geography changes the shortlist

The inventory also shows why city names need careful handling. Among the verified records, 28 use the exact city label Madrid and 20 use Barcelona. Alcobendas has another six, Alcalá de Henares five and Cerdanyola del Vallès six. These are separate locality labels, not complete metropolitan totals.

A search restricted to the literal word Madrid would therefore omit nearby records that could matter commercially. Conversely, counting every record in a wider area as interchangeable would ignore latency requirements, fibre routes, operating preferences and the actual connection point.

PowerlandMap's market-screening approach is most useful when these layers are combined around a defined requirement. A map can identify candidates and reveal gaps. Route diversity, available space and contracted power still require their own evidence. Geographic proximity alone does not establish any of them.

Two Spanish projects illustrate different delivery models

MERLIN's Zaragoza-WIND announcement, published on 23 July 2026, describes a planned 144 MW IT development in Botorrita, with operations targeted for the second half of 2029. The company presents renewable self-consumption as a central part of the design and says battery integration is being studied. Those are development commitments and targets. They do not establish commissioned capacity or space available today.

For a buyer planning a large future deployment, that record raises useful questions about phase delivery, the energy design and the evidence supporting the schedule. For a requirement that must be operating in 2027, the announced timetable places it in a different commercial conversation.

The hscale Madrid project page describes another approach. It lists a Q3 2028 target, 12 MW phases and subsequent delivery intervals of three to six months. It also describes power applications as underway. The phase schedule and power status should therefore be read together.

Neither example establishes that one location is preferable. They show why a campus headline and a customer's first required block must be compared on the same time basis. A developer's published programme is an input to diligence, not a reservation offer.

Grid access provides a separate national signal

In its release dated 1 October 2026, Red Eléctrica reported 6,329 MW of granted transmission access for data-processing centres, with none of that reported allocation commissioned at the time of the snapshot. The release explicitly excludes distribution-network applications.

This is not a claim that Spain has no operating data centres. It describes a specific transmission-access dataset. It also cannot be added to IT-load figures or used as the denominator for PowerlandMap's record coverage.

The commercial implication is to maintain three separate questions. What is the project's stated IT design? What electrical access has been granted, on what terms? What can actually be handed over by the customer's date? Our analysis of grid MW and live IT capacity explains why each answer needs its own evidence.

Build two commercial lists

Consider two hypothetical requirements. They illustrate a screening method and are not client mandates or offers of available capacity.

An 8 MW IT requirement for 2027 should start with operating sites and construction phases that have dated delivery evidence. The first enquiries should establish available halls, power allocation, cooling compatibility, commissioning milestones and commercial terms. A facility being operational does not mean it has 8 MW to sell.

A 48 MW IT requirement starting in 2029, with room to expand, can justify a wider development search. The relevant questions become first-phase size, subsequent delivery dates, land control, grid conditions, permit milestones and financing. A large ultimate campus envelope is useful only if the initial phase can stand on its own.

The same distinction helps advisers and engineering firms target work. Operating assets may create opportunities around upgrades and expansion. Earlier projects create different needs around feasibility, connection strategy and phased delivery. These are commercial hypotheses to qualify against each record, not revenue estimates.

Turn the data into a decision

For a Spanish market review, I would ask for a shortlist with a defined IT requirement, service date, geography and cooling specification. Each candidate should then carry a source, a verification date, its current stage, the basis of every capacity number and a clear list of unresolved questions.

That produces a practical output for the next call. It tells the buyer which operators to approach, the developer which milestones to substantiate and the adviser where additional work can change a decision.

Explore the relevant PowerlandMap use cases, or request a Spain-focused screening discussion. Bring the required capacity and delivery date. The purpose is to identify plausible candidates and the evidence needed to progress them.

Sources and method

The database extraction uses published, non-demo Spanish records classified as public supply. The stage and IT-field analysis further restricts the population to the 218 records carrying the verified-public flag. That flag describes evidence qualification in the database. It is not an independent technical certification of every field or a guarantee of current project status.

Counts are not deduplicated physical-campus market totals. No estimated IT-capacity fields are used. No confidential project information is included. No capacity sum, occupancy estimate or national market share is inferred. The 12.8% figure is 28 divided by 218, rounded to one decimal place.

External checks were completed on 30 October 2026 against the linked Red Eléctrica release, MERLIN announcement and hscale project page. Their publication dates and development qualifiers are preserved. The PowerlandMap methodology explains the broader distinction between recorded evidence and assumptions.