A €13 billion signal for Europe’s AI map
Google has put Finland at the center of its European AI infrastructure strategy with a plan to invest at least €13 billion in digital infrastructure, clean energy projects and local partnerships over 2027 and 2028 . The company says the package covers data center and supporting infrastructure investments in Hamina, Kajaani, Muhos and Vaala, four municipalities that now form the map of its Finnish expansion . Google has described the move as its largest single investment in Europe, and the scale makes the announcement more than a conventional data center build-out: it is a strategic commitment to where AI computing capacity, energy supply and cloud connectivity will be located in the next phase of European digital infrastructure .
The investment is designed to support demand for Google services including Gemini, Search, Maps and YouTube, while strengthening the infrastructure used by Finnish and European organizations building AI-powered services . In practical terms, the announcement means new and expanded facilities, grid-oriented siting decisions, power purchase agreements, workforce programs and community funds are being presented as one integrated package rather than as separate corporate social responsibility add-ons . That bundling matters because data centers are no longer judged only by capacity and latency; they are increasingly judged by how they affect national electricity systems, local labor markets and regional planning.
The Finnish locations are also important. Hamina is already part of Google’s history in the country: the company bought a former paper mill there in 2009 and converted it into a data center, later developing cooling and heat-recovery approaches that it now uses as evidence for responsible data center development . Kajaani, Muhos and Vaala shift the growth story further toward northern and central Finland, where Google says it worked with Fingrid, Business Finland and municipalities to identify sites near existing grid infrastructure and carbon-free energy resources . The company’s message is clear: Finland is not simply a cold place to put servers; it is being pitched as a test case for connecting AI infrastructure to a decarbonized power system.
Jobs, GDP and local commitments
Google estimates that the initial 2027-2028 construction phase will support more than 37,000 jobs nationwide and contribute an annual average of €3.6 billion to Finland’s GDP . About 16,000 of those jobs are expected to be in construction, with the construction sector receiving an estimated €911 million in average annual labor income during the build phase . Once the facilities are operational, Google projects that the infrastructure will support around 7,000 jobs annually across on-site technical and facility roles, upstream suppliers, and local businesses serving workers and their families .
These figures should be read as company projections, not realized outcomes. Still, the numbers explain why the announcement has been welcomed at the national level. Finnish Prime Minister Petteri Orpo said the decision shows Finland’s strengths and argued that the data economy’s value goes beyond direct investment by stimulating innovation, research and development . The political framing is that AI infrastructure is not just a technology sector issue but a national economic development tool.
Google is also attaching community spending to the investment. The company says it will invest €31 million in Hamina, Kajaani, Muhos and Vaala over four years for education and economic growth, including AI upskilling programs for more than 4,400 workers and new training opportunities for 100 Finnish students preparing for data center careers . It also plans nature and community initiatives, including restoration of former forestry lands near data center sites and support for local energy-affordability projects , . These details are not peripheral: for municipalities hosting large energy-intensive facilities, visible local benefits can determine whether expansion is seen as an opportunity or an imposition.
Why energy is the core of the deal
The heart of the Finnish announcement is power. AI infrastructure requires huge quantities of reliable electricity, and Google’s Finnish plan is built around a portfolio of nuclear, wind, battery and grid measures intended to match new demand with additional low-carbon supply . The most consequential piece is a 22-year power purchase agreement with Fortum tied to the life extension of the Loviisa nuclear power plant . Fortum says the agreement covers the lifetime-extension period and will begin in 2028 with a smaller contracted capacity, reaching 50% of Loviisa’s capacity during 2030-2049 .
Fortum’s statement makes the stakes unusually explicit. The Loviisa plant employs about 580 people and provides roughly 10% of Finland’s electricity, and Fortum says that without the lifetime-extension investment program the plant could not continue operating beyond 2030 . The utility has an approximately €1 billion investment program to extend Loviisa’s lifetime to 2050, with about €700 million of capital expenditure still pending investment decisions at the time of the announcement . Fortum says the Google PPA provides revenue certainty for those investments and is expected to enable an additional 10 MW power increase on top of a planned 38 MW uprate expected to be live in 2028 .
Google is also presenting the Loviisa agreement as part of a broader model for integrating AI infrastructure into power systems. The company says the deal is its first agreement for a nuclear plant life extension and uprate, and that it will work with Fortum to identify potential opportunities for new nuclear reactors at Loviisa . Fortum and Google have also signed a memorandum of understanding covering new nuclear, renewable capacity, flexibility solutions and energy portfolio management services . In other words, the data center investment is being tied to long-term electricity market design, not merely to a short-term procurement contract.
Wind, batteries and grid-friendly siting
Nuclear power is only one part of the package. Google says additional onshore wind power purchase agreements with Valorem and Suomen Hyötytuuli will bring the total new-to-the-grid onshore wind capacity supported by the company in Finland to 629 MW . It is also contracting for a 94 MW battery system near its Kajaani site, expected to be operational in late 2027, to help balance the grid and reduce price volatility during periods such as cold, low-wind weather . Google says it will also explore demand response, including the possibility of temporarily reducing data center power demand during grid stress .
The siting strategy is central to Google’s argument that the investment can be compatible with affordability. The company says it worked with Fingrid, Business Finland and municipalities to place new infrastructure in northern and central Finland, closer to available grid capacity and carbon-free generation . It cites a commissioned study modeling a hypothetical 1 GW of new energy demand in the Oulu-Kajaani region compared with southern Finland, saying such siting could produce an estimated €520 million in savings for Finnish energy consumers over 20 years . Because this is a modeled and commissioned estimate, it should be treated as a scenario rather than a guarantee. But it shows the logic behind the location choices: large AI loads are being positioned as flexible industrial demand that can support, rather than strain, the grid if placed correctly.
Reuters reported that Google President and Chief Investment Officer Ruth Porat described the approach as “bring your own power,” and said it was Google’s first nuclear energy deal outside the United States . That phrase captures the political challenge facing hyperscale AI infrastructure in Europe. Companies cannot simply request grid capacity and assume public acceptance. They increasingly need to show that new computing demand will be accompanied by new generation, storage, balancing tools and local benefits.
What this means for Finland
For Finland, the investment strengthens a position the country has been building for years: a cold-climate, high-connectivity, low-carbon platform for data centers and AI services. The combination of nuclear power, wind resources, district heating potential and strong grid institutions gives Finland a pitch that is broader than cheap land or low temperatures. Google’s announcement turns that pitch into a high-profile industrial policy moment, because it ties AI infrastructure directly to national energy planning and regional development.
The opportunity is substantial. If the jobs, supplier and wage projections materialize, the investment could create a large construction wave followed by a permanent ecosystem of technical, facility, maintenance, security, catering, supplier and local service roles . The company also says its Hamina presence supported more than 600 Finnish suppliers across construction, operations and network infrastructure from 2023 to 2025, which it presents as evidence that data centers can generate local multiplier effects .
The risks are just as clear. AI infrastructure can concentrate electricity demand, intensify debates over land use and biodiversity, and test whether promised community benefits arrive at the same pace as construction. Google is trying to pre-empt those concerns by combining the data center announcement with nuclear life extension, wind PPAs, battery storage, grid-friendly siting, workforce training and restoration projects. The credibility of the plan will depend on execution: whether the sites are built on schedule, whether the battery and energy contracts operate as described, whether electricity prices remain manageable, and whether local communities see durable benefits.
As of the latest announcements within the past 72 hours, the core story is therefore not only that Google is investing €13 billion in Finland. It is that the company is attempting to define what a politically viable AI infrastructure build-out in Europe looks like: large enough to serve Gemini and cloud demand, local enough to create jobs and training pathways, and energy-aware enough to answer the question that now shadows every AI data center project — where will the power come from, and who benefits from the system built to supply it?