3 countries, 3 different ways to structure a data center project

Over the past ten years, I have worked on data center projects across Europe and the Middle East, from early site qualification to design, construction, commissioning and commercial structuring.

One of the most important lessons I have learned is that a data center project cannot simply be copied from one country to another.

The fundamental requirements remain the same. Every project needs land, power, connectivity, permits, capital and customers. However, the order in which these elements must be secured can change significantly depending on the country.

In France, the challenge is often to manage several formal procedures that must progress together. In Poland, the challenge can be to convert a large long-term opportunity into an investable first phase. In the UAE, the administrative and electrical timelines can sometimes move faster, but the project must correspond to real and immediate customer demand.

These observations are based on projects I have developed, delivered or reviewed. I have kept the projects anonymous and some figures deliberately broad in order to protect ongoing discussions.

France requires patience, documentation and parallel workstreams

France is one of the markets I know best. It offers strong demand, a reliable electricity system, experienced engineering companies and a mature investment environment. It also requires a highly structured development approach.

A French project rarely depends on one single approval. The land, grid connection, planning application, environmental authorisations, technical studies and commercial strategy must usually progress in parallel.

A site may be compatible with the local planning framework but still face environmental constraints. A power connection may be technically feasible but require several years of network works. A municipality may strongly support the project while the permitting process remains incomplete. A test fit may demonstrate that the site can accommodate a certain IT capacity, but the final development density may still depend on acoustic, flood, fire safety or industrial risk constraints.

I have worked on French projects where the local political support was exceptionally strong. In one case, the project received unanimous support at the municipal level. That was an important signal for investors and operators, especially at a time when large data center developments can face increasing public scrutiny.

However, I was careful not to present that political support as equivalent to a fully authorised project. A favourable local vote does not replace the planning process. It does not remove the need for environmental studies. It does not guarantee that the grid capacity will be delivered on time. It also does not prevent future questions concerning water, noise, energy consumption or the visual impact of the development.

This distinction is essential when presenting a French opportunity to an international investor. The correct message is not that the project is risk-free because the municipality supports it. The correct message is that one important development risk has been reduced, while the remaining risks are clearly identified and managed.

Power phasing is another recurring subject in France. On several projects, I have had to consider a smaller initial connection through the distribution network before a much larger transmission connection becomes available. The first capacity can help create an earlier commercial phase, while the main connection supports the long-term campus.

This only works if the masterplan is designed correctly from the beginning. The initial phase must not block the location of the future substation, buildings, access roads or utility corridors. It must also make commercial sense on its own. A smaller bridge phase is not useful if it is too small for the target customer or if the cost of delivering it cannot be recovered.

France has taught me to manage development as a collection of interconnected workstreams. The project programme must show when the land becomes secure, when the grid position becomes sufficiently reliable, when the permit can be submitted, when environmental studies can be completed and when the first commercially useful capacity can be delivered.

The strength of a French project does not come from one spectacular announcement. It comes from the quality and consistency of its documentation.

Poland requires a clear distinction between short-term capacity and long-term ambition

Poland presents a different development environment. The country can offer larger land opportunities, lower entry costs than some Western European markets and significant potential for new digital infrastructure. It can also provide a route into markets that are increasingly important for cloud, sovereign infrastructure and AI.

However, large land availability can create the illusion that the project is more mature than it really is. I recently reviewed a large campus opportunity in Poland. The overall transaction perimeter was close to one hundred hectares, but the area immediately identified for data center development was much smaller.

That difference immediately became one of my main due diligence questions. What part of the site was actually owned by the seller? Which parcels had the appropriate zoning? Which areas were affected by access requirements, environmental constraints or future energy infrastructure? Could the relevant land rights be transferred to a project company? Was the entire site necessary for the first phase?

A headline land area is useful for describing long-term potential. It should not automatically be used in the initial valuation or development model.

The electrical strategy created a second important distinction. The opportunity included a relatively small initial power solution, followed by a possible medium-scale phase and a much larger long-term connection through the transmission network.

At first sight, this could be presented as one campus with several hundred megawatts of potential.

In practice, I saw three different investment projects.

The first project was an early deployment based on limited bridge power. It could potentially be delivered faster, but it needed a customer whose requirements matched that capacity. The second project was a larger phase that would require additional grid work and probably an anchor tenant. Its success depended on the relationship between customer commitment, utility investment and development financing. The third project was the long-term hyperscale and AI campus. This required major transmission infrastructure, a longer programme and a completely different level of capital.

The mistake would have been to value the land as if the full long-term electrical capacity were already secured.

My approach was to separate confirmed capacity from conditional capacity and strategic potential. Each category needed its own programme, investment requirements and evidence. This also changed the way the project could be presented to investors.

A powered-land investor may be interested in acquiring and progressively de-risking the site. A hyperscale operator may focus on the date at which the first significant phase can be delivered. A long-term infrastructure investor may be attracted by the full campus potential but unwilling to take early land and permitting risk.

My experience in Poland has taught me that phasing is not simply a technical exercise. It is a way of making a large and uncertain opportunity investable.

The objective is to identify the smallest first phase that can create genuine value without compromising the long-term campus.

The UAE can move quickly, but speed does not replace customer validation

The UAE offers another development logic ! Discussions with landowners, local partners and infrastructure providers can move quickly. Strategic projects can receive strong institutional attention. The country also has major ambitions in artificial intelligence, cloud infrastructure and digital sovereignty.

This can create a very attractive environment for new data center platforms. However, a fast-moving environment can also encourage developers to progress too many assumptions at the same time.

On one UAE project, I received documentation indicating that several tens of megawatts could be supplied to a compact development site. The proposed grid programme was significantly faster than many European connection timelines.

The opportunity initially appeared capable of moving rapidly towards an LOI and a first modular deployment. When I began building a credible Ready for Service programme, I asked for the complete critical path behind the announced date.

I needed to understand the formal power position, connection design, cable works, transformers, remaining approvals, construction scope, site conditions, procurement programme, testing process and final energisation sequence.

I also discovered that the site had originally been considered for a different type of computing infrastructure. Converting the concept into an AI data center required more than replacing the description in the presentation.

The electrical architecture had to be reviewed. The cooling strategy had to reflect high ambient temperatures and higher rack densities. The site layout had to support appropriate redundancy, maintenance and future expansion. The authorised use also needed to match the proposed data center development.

An initial operator reviewed the opportunity but was unable to identify a suitable customer. The feedback was that the proposed first phase was too small for that operator’s deployment strategy.

This was a useful reminder that the same capacity can be attractive to one category of customer and irrelevant to another. At that point, I decided that producing more legal and technical documentation would not solve the real issue. Before progressing to a formal project commitment, the priority became validating demand with several potential tenants and another operator.

This changed the project sequence. Instead of securing the site, completing the design and then looking for a customer, we started testing the proposed location, capacity and delivery model directly with the market.

If the tenant discussions confirm demand, the project can move forward with a clearer design basis and a more credible business plan. If the feedback is negative, the project can be resized or repositioned before significant development capital is spent.

The UAE has taught me that speed is valuable only when the project is moving in the right direction.

A short permitting timeline and available power can create a strong opportunity. They do not compensate for a product that does not match the requirements of operators and end users.

The critical path is different in every country These three countries illustrate why there is no universal development sequence.

In France, I generally focus on coordinating a dense set of formal procedures. The challenge is to keep the land, power, planning, environmental, technical and commercial workstreams aligned over a relatively long programme. In Poland, I focus on separating what can be delivered in the short term from the much larger long-term vision. The challenge is to create an investable first phase without overvaluing future power or undeveloped land. In the UAE, I focus on validating the relationship between speed, technical assumptions and customer demand. The challenge is to make sure that a fast-moving project corresponds to a real deployment requirement.

My first questions therefore change depending on the country. It is not about applying one model across several countries.

It is about understanding each market well enough to build the right model for each project.