Google Commits 13 Billion Euros to Finnish Electrons

The most consequential detail here is the Loviisa clause. Fortum's plant, per the announcement, would not have been able to continue operating beyond 2030 without the agreement Google signed ^[8], and Google's own estimate is that the deal keeps roughly 10% of Finland's electricity supply online while avoiding price increases for grid users ^[9]. So the buyer of the compute is also, in effect, underwriting the survival of a tenth of a national grid's generation. This looks less like a boring power purchase agreement and more like a hyperscaler taking a position on Finnish generation capacity, because the alternative was building into a market that had quietly lost a tenth of its supply.

Run the arithmetic on the headline number and it gets less impressive and more interesting. At least 13 billion euros over two years ^[2] is about 6.5 billion euros a year ^[11], spread across four locations ^[6] plus grid upgrades and renewables ^[7]. Against the pace of US hyperscaler capex that is a rounding line, which is exactly why the Loviisa piece matters more than the euro figure: the constraint being bought is electrons, not land. The money is doing something narrower than it looks: it is buying dispatchable Finnish generation, not a second Virginia-style cluster where interconnect queues are the binding item, and not capacity Google could flip to another use if European AI demand disappoints.

The concentration claim deserves scepticism in proportion to its arithmetic. Google's 13 billion euros, TikTok's separately announced 1 billion euros in Kouvola ^[4] and the 1.1 billion dollars it flagged in April ^[3] sit alongside a Microsoft position that is still, on the company's own account, 190 hectares of land under a preliminary agreement with the construction application not yet through ^[5]. Fortune totals the pledges at more than 30.2 billion dollars ^[1]; a large share of that is a two-year commitment from one buyer, and land options are not concrete. The more interesting version of the claim is simpler: Finland has won the energy argument, and the capital is following the energy, which is a different and more durable thing than winning on tax incentives.

Eighteen months from now, this could read three different ways. Google underdelivers against 13 billion euros because power interconnection slips, and the pledge becomes a headline that outlived its cash. Microsoft's 190 hectares never clears permitting, and Finland turns out to be one anchor tenant plus a video app. Or the Loviisa structure gets copied, at which point every Nordic utility with a life-extension decision pending has a hyperscaler in the room. The evidence in front of us supports the third more than the first two, because the counterparty being paid is a utility with a dated cliff and Google's stated interest in avoiding grid-user price increases ^[9] is the kind of self-interest that holds.

What the sources do not show is any figure for megawatts, any capacity number, or any disclosed price for the Loviisa extension, and without those the 13 billion euros is a commitment rather than a demonstrated cost per unit of compute. Prime Minister Petteri Orpo said attracting further investment remains a top priority and called the decision a clear testament to Finland's strengths ^[10], which is what a prime minister says. The test is whether the grid upgrade line item ever gets itemised.