The Nordics: Cheap Clean Power at a Distance

The Nordic markets, meaning primarily Norway, Sweden, Finland, and to a lesser extent Denmark and Iceland, present the cleanest version of a single trade: abundant, low-carbon, low-cost electricity in exchange for distance from the demand centers that drive latency-sensitive workloads. Understanding this trade in detail is the template for reading every other region.

Start with the power. Norway generates the overwhelming majority of its electricity from hydropower, one of the highest such shares of any developed country. Sweden runs on a mix of hydro and nuclear with a large renewable component. Iceland runs almost entirely on hydro and geothermal. Finland has added nuclear capacity and imports Nordic hydro across a well-integrated regional grid. The practical consequence is that a data center in this region can plausibly claim near-total renewable or low-carbon power, often under long-term contracts at prices that have historically undercut continental Europe. For an operator under pressure to report low emissions and stable energy costs, this is a genuine and durable advantage, not a marketing gloss.

The cold climate compounds the benefit. For much of the year, a facility in northern Sweden or Norway can cool its servers with outside air rather than energy-hungry mechanical chillers, a technique called free cooling. Lower cooling load means a lower power usage effectiveness ratio, the standard measure of how much total power a facility draws for every unit that reaches the computing equipment. A better ratio means more of the purchased power does useful work, which lowers both cost and emissions per unit of compute. The Nordics can hit efficiency figures that warmer climates simply cannot match without expensive engineering.

So why is the entire industry not in the Arctic Circle? Distance. The Nordic population centers, and the interconnection hubs that serve real-time workloads, sit far to the south of the cheapest power. A user in London or Frankfurt requesting a response from a server in northern Norway pays a latency penalty measured in the added milliseconds of round-trip signal travel. For a training run that grinds for weeks, those milliseconds are irrelevant; the job does not care whether its answer returns instantly. For real-time inference, video conferencing, financial trading, or interactive gaming, the penalty is disqualifying. This is why the Nordics have won a disproportionate share of large training and batch workloads while the latency-sensitive tier stays close to FLAP-D.

There is a second, subtler constraint: grid geography within the Nordic countries themselves. The cheapest power sits in the north, but the transmission capacity to move it south, or to connect a large new load in the north, is finite. Norway and Sweden both divide into electricity price zones precisely because the grid cannot move unlimited power across the country. A campus sited in the far north may enjoy cheap generation nearby but still wait years for the transmission reinforcement that lets it draw at scale. And as more industry electrifies, local opposition to exporting cheap power, whether to other regions or to power-hungry data centers, has become a live political question in Norway in particular. The cheap power is real, but it is neither unlimited nor politically frictionless.

Here is Faraday's read on how to use the Nordics. Treat the region as the natural home for the workloads that are indifferent to latency and hungry for clean, cheap, stable power: model training, batch analytics, archival and backup, and scientific computing. Do not fight geography by forcing latency-sensitive workloads north to chase a low energy price; the round-trip penalty will erase the saving in user experience. And underwrite the grid connection with the same rigor you would anywhere else, because a Nordic connection agreement in a transmission-constrained zone can be as slow to energize as one in Amsterdam. The Nordic advantage is real and durable, but it is workload-specific, and a practitioner who places the wrong workload there has misread the single most important trade the region offers.