Google plans to invest at least €13 billion in Finnish digital infrastructure over 2027 and 2028, a scale it calls its largest single investment in Europe. The September 9 announcement covers data centers and supporting infrastructure in Hamina, Kajaani, Muhos and Vaala, but the most interesting element is the attached energy and community architecture. Google is not presenting compute as an isolated building project. It is offering a package of grid-location planning, long-term power arrangements, wind, battery storage, heat recovery, workforce initiatives and local funds. That approach acknowledges the central AI-infrastructure constraint: a new data center is also a large, persistent electricity customer.
The energy commitments are more specific than generic clean-power language. Google describes a 22-year power-purchase agreement with Fortum designed to support the Loviisa nuclear plant’s life-extension program. It also signed a memorandum of understanding with Fortum to explore flexibility capacity, new generation and business models for possible new nuclear reactors. Separately, it says additional wind agreements bring the new-to-grid capacity it supports in Finland to 629 MW, and a contracted 94-MW battery system near Kajaani is expected to become operational in late 2027. Each of these components serves a different system function; none is a substitute for the others.
A long-term nuclear agreement can provide revenue certainty for firm low-carbon generation. Wind contracts can add new supply but vary with weather. Batteries can help balance shorter-term variation and support grid services, yet their duration and charging source determine what they can contribute during sustained scarcity. Location matters too. Google says it will work with Fingrid and Business Finland to identify grid-friendly sites in northern Finland near existing capacity and carbon-free resources. This is a more sophisticated argument than simply claiming that data-center load will be “green.” The difficult work lies in matching the timing, location and reliability of load and supply.
The announcement also frames infrastructure as local economic policy. Google projects an average annual contribution of €3.6 billion to Finnish GDP and more than 37,000 jobs during the 2027–28 construction period, including about 16,000 construction jobs. It projects 7,000 annual jobs once the facilities operate. These are company estimates, not independently audited outcomes, and should be treated as such. Data-center construction can generate significant activity, but the number, quality and persistence of local jobs depend on procurement, contractor practices, skills pipelines, tax arrangements, power costs and the degree to which spending remains in the region.
Google is attempting to address that concern with €31 million over four years for community partnerships, including €10 million for research and innovation, as well as AI education and technical training. The company also describes investments in municipal energy projects, heat-pump and solar installations for low-income households, backup power for essential local services and biodiversity restoration. These elements can improve community acceptance and generate public value. They should not be treated as a waiver of scrutiny. Residents and regulators will still need transparent data on electricity use, water use, local prices, grid upgrades, construction impacts and the delivery status of promised benefits.
AI infrastructure is becoming a form of industrial policy because model training and inference require energy, transmission, chips, real estate and skilled labor at unusual scale. Finland’s combination of low-carbon power, industrial capability and available grid capacity creates an opportunity, but it does not remove trade-offs. Electricity must remain reliable and affordable for households and existing industry; projects must respect environmental constraints; and energy investment must proceed fast enough to avoid turning future capacity promises into bottlenecks. The most credible data-center plans will be those that make these trade-offs measurable rather than merely aspirational.
Google’s proposal is therefore worth watching as a model of how AI capacity can be linked to the energy transition. The company has put real financial commitments and named partners behind the message. The test comes in execution: whether new facilities, renewable additions, battery storage, nuclear life extension, heat recovery and local programs arrive on schedule and interact as planned. If they do, Finland could show that compute growth can be paired with system planning. If they do not, the expansion will underline that pledging clean infrastructure is easier than operating it through the constraints of a real power grid.