Google to Invest €13 Billion in Finland AI Infrastructure, Signs First Nuclear Deal Outside US

Product / TechMacroCommodity Impact 4
โดย Reuters·FIUS·Read original
Summary · why it matters

Google will invest at least €13 billion, or about $15.1 billion, in AI infrastructure in Finland over the next two years, its biggest European investment, including three new data centers and a 22-year deal for up to 50% of the output of one of Finland's two nuclear power plants. The agreement with Finnish utility Fortum is Google's first nuclear energy deal outside the United States, and Fortum's stock jumped as much as 15.5% on the news, its biggest gain in Europe that day. The deal covers up to 50% of the electricity generated by Finland's Loviisa nuclear plant and gives Fortum greater economic certainty to extend and upgrade the plant through 2050. The Finnish project adds to Alphabet's already enormous AI spending program, with the company having raised its 2026 global capital investment target to $195 billion-$205 billion. Alphabet shares fell 2.3% after the Finland announcement, underlining market concern that rapidly rising AI investment could pressure returns before the resulting revenue growth fully materializes.

Impact on stocks 3

Artificial Intelligence · 2 stocks
Alphabet Inc Class C
GOOG
▼ NegativeCapitalrelevance

Alphabet's €13B Finland AI infrastructure investment and raised 2026 capex target ($195-205B) pressure returns, with shares falling 2.3%

Energy Transition & Power Demand · 1 stocks
Fortum Oyj
0HAH
▲ PositiveDemandrelevance

Fortum signs 22-year deal to supply up to 50% of Loviisa nuclear plant output to Google, giving economic certainty to extend and upgrade the plant

Theme Impact 7

Related news

2

BGRIM unveils plan to meet 8,800 MW data centre power demand, pushes Direct PPA and data centre COD this November

B.Grimm Power, or BGRIM, has announced a plan to serve massive electricity demand from future industries including data centres, artificial intelligence, electronics and the EV ecosystem. Noppadol Karansut, Chief Executive Officer for business in Thailand, Malaysia and industrial business solutions, told Thunhoon that the draft national power development plan, or PDP, estimates data centre electricity demand at 8,800 megawatts over the next 25 years, higher than Malaysia and Singapore, which each have roughly 2,000 to 4,000 megawatts of installed capacity and projects under development. Combined with its existing industrial customer base of 2,300 megawatts, the company sees an opportunity to develop 10,000 to 30,000 megawatts of power from gas and clean energy. The government is also allowing direct power sales to the private sector, or Direct PPA, and the company has secured high-voltage substation and transmission line sites covering 11 industrial estates to support the government's 2,000 megawatt Direct PPA pilot project. On the data centre business, the company has formed a joint venture with a Singaporean firm to develop a 96 megawatt Hyperscaler AI Data Center worth 26 billion baht in Chonburi province. Phase 1, at 48 megawatts, is scheduled for commercial operation, or COD, this November, while Phase 2, also 48 megawatts, is set for COD the following year. The company aims to reach 300 megawatts of total data centre capacity by 2030. As for rising gas prices, the company has completed negotiations with industrial customers on a new Gas Linking electricity pricing structure covering 400 megawatts out of a total of 800 megawatts, with another 400 megawatts still under negotiation. At the same time, the company is structuring its finances to keep its Net IBD/E ratio below 2.0 times, with a target of reducing it to 1.3 to 1.5 times.
ทันหุ้น·5hRead more →
impact 4

IEA Projects AI Data Center Power Demand to More Than Double by 2030

The International Energy Agency projects that electricity demand from AI data centers will more than double between 2024 and 2030, rising from 415 terawatt-hours to 945 terawatt-hours, with demand reaching 1,200 terawatt-hours by 2035, nearly triple the 2024 figure. The IEA noted that AI electricity demand in 2030 will exceed Japan's current power use, and data center electricity use has been growing at 12% per year since 2017, more than four times the overall rate of consumption growth. Among the companies positioned to benefit, Bloom Energy entered 2026 with a product backlog of $6 billion, up 140% over 2025's starting backlog, plus a service backlog of $14 billion. Constellation Energy, one of the largest operators of nuclear power plants in the United States, is already working with AI companies including Meta and Microsoft, while Cameco expects nuclear fuel demand to outstrip supply by the mid-2030s. NextEra Energy is doubling down on AI and data centers with its pending acquisition of Dominion Energy, which holds a monopoly in one of the world's largest data center markets.
The Motley Fool·6hRead more →

American Electric Power Files Updated Niagara Project Documents With FERC

American Electric Power Company has filed updated technical and digital documentation for the Niagara Project with the Federal Energy Regulatory Commission. The submission covers refreshed engineering details and digital system information tied to the Niagara Project's regulatory review, and supports ongoing federal oversight of the project as part of the broader approval and compliance process. The next thing to watch is how the Federal Energy Regulatory Commission responds to the updated Supporting Technical Information Document and Digital Project Archives, including any follow-up data requests or conditions that could affect project timing and, by extension, when American Electric Power Company can place Niagara-related investment into its regulated rate base. The filing reinforces the company's existing narrative that management is active on regulatory files and large grid projects, and supports the idea that regulatory work is a key catalyst for its US$78b capital plan and transmission build, which analysts link to future earnings and load growth. The update does not address the broader financial risk flagged in the data, where interest costs and dividend coverage already look tight, leaving investors without clarity on how new project spending will sit alongside those constraints.
Simply Wall St·8hRead more →