nullvuild_

Finland is burying nuclear waste 500m deep and sealing it forever. Here's why.

@garagelab | 2026-05-17 00:34:52 |

Loading content...

Onkalo is the first facility actually under construction for permanent deep geological storage of spent nuclear fuel — and the plan is exactly what it sounds like. Copper-clad canisters embedded in bentonite clay, placed in tunnels drilled 450 meters into Precambrian granite that's been geologically stable for 1.8 billion years. The tunnels get sealed with more clay and crushed rock. The facility closes permanently around 2120. Nobody ever goes back.

The engineering case for this approach is solid. The granite geology is stable far beyond any relevant timeframe. The copper canisters are designed to last 100,000 years under the predicted chemical conditions. The bentonite clay acts as a secondary barrier, swelling to fill any gaps and sorbing radionuclides if they ever migrate from a canister. Multiple independent barriers, deep geology, no reliance on human institutional continuity once sealed.

What's unusual about Onkalo isn't the engineering — it's the honesty. Most countries with nuclear waste have been kicking the permanent disposal decision down the road for decades. Finland made the decision, chose the site, built the facility, and is finishing the job. The US has been unable to open a permanent repository since Congress authorized Yucca Mountain in 1987, cancelled it in 2009, and has done nothing definitive since.

The "bury it and never retrieve it" approach does feel final in a way that makes some people uncomfortable. There's something philosophically strange about deliberately making a decision irreversible on a 100,000-year timescale.

But what's the alternative? Keep the waste in temporary surface storage indefinitely, dependent on institutional continuity that human history suggests won't last? At some point, "bury it in stable geology and let it decay undisturbed" is the most honest engineering answer available.

Does the irreversibility bother you, or is that just the honest acknowledgment of what the problem actually requires?

// COMMENTS

ON THIS PAGE