Large infrastructure announcements are exercises in the future tense, and the one Google made on 9 September is no exception. The company said it plans to invest at least 13 billion euros in Finland’s digital infrastructure during 2027 and 2028, including data centres in Hamina, Kajaani, Muhos and Vaala. The announcement is significant, and reading it accurately means separating the commitment from the projections attached to it.
What was actually announced
A forward commitment to spend at least a stated amount over two years, in named locations, with supporting energy arrangements. That is a concrete statement of intent by a company with the resources to carry it out. It is a forward commitment, and the money will be spent over time, subject to the ordinary risks of construction, permitting and market conditions.
Why Finland
Finland has several advantages for data centres: a cool climate that reduces cooling costs, a relatively low-carbon electricity supply, a stable regulatory environment and good connectivity. It has also become a hub for the sector over the past decade, with earlier Google facilities already operating there. The new investment extends an existing presence rather than starting from nothing.
Electricity is part of the plan
The announcement includes energy arrangements, among them a long-term agreement with Fortum connected to the Loviisa nuclear plant and a planned battery installation near Kajaani. This reflects the central constraint on data-centre expansion: power. AI workloads are electricity-intensive, and the availability of reliable, low-carbon supply is now a primary factor in where capacity is built.
Google’s estimates of employment and economic benefit are forecasts, not a count of benefits already delivered. Such figures are typically modelled using multipliers and assumptions, and they are contested in the economics literature. That does not make them worthless; it means they should be read as claims to be checked against what happens, not as established outcomes.
The local question
Data centres create construction jobs, a modest number of permanent roles, and local tax revenue. They also consume significant electricity and water, and they occupy land. The balance of costs and benefits varies by community, and it is often more complicated than a headline investment figure suggests. The communities named in the announcement will experience both sides.
The European context
The investment lands in the middle of a European debate about digital sovereignty and dependence on foreign cloud providers. Large commitments by US technology companies are welcomed for the investment and scrutinised for the concentration of control they imply. Both reactions are reasonable, and they coexist in most European capitals.
The evidence that matters will accumulate slowly: planning approvals, construction milestones, operating dates and the performance of the associated energy projects. An announcement sets an expectation; delivery is measured in years. Readers should treat the figure as the beginning of a story rather than its conclusion.
The energy economics
A data centre is, in energy terms, a large and constant load. That makes the price and carbon intensity of electricity central to its economics, and it explains why operators increasingly sign long-term power agreements and invest in generation or storage. The arrangements announced alongside Google’s Finnish investment reflect this logic: securing supply is part of the business case, not a side project. As AI workloads grow, competition for low-carbon power is likely to intensify.
Water and land use
Data centres use water for cooling, sometimes in significant quantities, and they occupy land. In a country with abundant water and space, these are manageable; in water-stressed regions, they are contested. Finland’s cool climate reduces cooling demand, which is one reason the country is attractive. Even so, the cumulative effect of several large facilities on local resources is a question that planning processes are expected to address.
European policymakers want both the investment that foreign technology companies bring and a degree of control over critical digital infrastructure. The tension is not easily resolved: the same facilities that create jobs and tax revenue also deepen dependence on a small number of non-European providers. Google’s Finnish investment will be welcomed and scrutinised at the same time, and that ambivalence is now the normal condition of European digital policy.
What could go wrong
Forward commitments can be revised. Companies adjust plans in response to demand, costs, regulation and technology. A two-year investment programme may be accelerated, delayed or reshaped, and the headline figure is a floor rather than a guarantee. That is normal in large capital projects, and it means the announcement should be tracked rather than banked. The local planning process, in particular, is where a proposal meets the practical constraints of land, power and consent.
The most useful way to read any large investment announcement is to note the date, the figure and the locations, then set a reminder to check what was actually built. Commitments of this size shape local economies for years, and the public record of whether they were honoured is what allows the next announcement to be judged against something real.
This report draws on Google Cloud.
Image: Rsparks3 · CC0 · via Wikimedia Commons.