Google to Invest $15B in Finland’s AI Infrastructure

Google has announced plans to invest approximately $15.1 billion in Finland's artificial intelligence infrastructure over the course of 2027 and 2028. This massive project represents the tech giant's most significant financial commitment in Europe to date. The funding will build out data centers and supporting infrastructure across Hamina, Kajaani, Muhos, and Vaala, alongside clean energy and battery storage projects. These facilities will directly power Google services, including its Gemini AI models, Google Maps, and YouTube.

The development is projected to contribute about $3.6 billion to Finland's gross domestic product during its construction phase and will support roughly 7,000 jobs annually once the sites are fully operational. To power these energy-intensive facilities sustainably, Google has signed a 22-year power purchase agreement with Finnish energy operator Fortum. This contract, Google's first nuclear power deal outside the United States, grants the company up to 50 percent of the electricity generated by one of Finland's two nuclear plants.

Finland is rapidly becoming a key hub for European AI infrastructure due to its stable power grid, cold climate, and abundant clean energy resources. Other major tech players are also expanding in the region. For instance, Pure DC recently confirmed a $1.74 billion investment to construct a 110-megawatt campus in Finland that can scale beyond 550 megawatts, while Dutch neocloud provider Nebius plans to build a 310-megawatt AI factory in the country.

For AI practitioners and enterprise developers, Google's massive regional expansion promises localized, high-performance compute resources with lower latency for European deployments of Gemini and other cloud services. It also demonstrates a viable path forward for scaling computationally intensive AI workloads without violating strict European corporate sustainability mandates, leveraging Finland's robust nuclear and renewable energy grid to keep carbon footprints minimal.