Think about the longest thing a human institution has ever kept going. The Catholic Church, a few dynasties, the odd old university. Almost nothing runs to tens of thousands of years.

Now think about a problem that demands exactly that kind of timescale: spent nuclear fuel remains hazardous for extremely long periods, and its radioactive contents have to be isolated from people and the environment.

Finland is planning around a horizon of roughly 100,000 years, using a mixture of careful engineering and rock far older than complex animal life.

The site is called Onkalo, near the Olkiluoto nuclear power plant in Eurajoki, southwest Finland. It is intended to isolate spent nuclear fuel for about 100,000 years and is expected to become the world’s first operating deep geological repository for spent nuclear fuel.

It has not begun accepting radioactive waste yet. As of June 30, 2026, Finland’s Radiation and Nuclear Safety Authority, STUK, was still finalising its safety assessment of Posiva’s operating-licence application. The Finnish government cannot grant the licence unless STUK supports it.

What interests me here is less whether the concept works than the strangeness of the bet: that ordinary copper, swelling clay and very old rock can continue protecting people long after the society that built the repository has disappeared.

What Onkalo actually is

Onkalo is a network of tunnels blasted into hard crystalline rock. That rock is around 1.9 billion years old, part of the Fennoscandian Shield, an ancient section of continental crust beneath Finland and its neighbours.

The disposal tunnels will be located roughly hundreds of metres underground. During one visit, the elevator display read 433 metres. When fully extended, Posiva says, the underground complex will contain more than 60 kilometres of tunnels.

Excavation began in June 2004. At first, Onkalo was primarily an underground research facility, built to study the bedrock and confirm that the site was suitable. The Finnish government granted Posiva a construction licence for the encapsulation and final-disposal facility in 2015, after which the project moved towards becoming an operating repository.

The planned capacity is large. According to Posiva, Onkalo can accommodate approximately 6,500 tonnes of uranium in spent fuel, sealed inside about 3,250 disposal canisters. The canisters will be lowered underground and placed individually into holes drilled into the floors of the disposal tunnels.

As Lauri Parviainen, a chemist at Posiva, put it, “Basically, it needs to be safe forever.” That is a goal rather than a promise anyone can literally guarantee, and the bluntness of the phrasing is part of what makes it interesting.

Copper, clay and ancient stone

The design is based on the KBS-3 method developed by Sweden’s nuclear-waste company SKB. Its central principle is that no single part of the system should have to do all the work.

Spent fuel will be placed inside a strong insert made from spheroidal graphite cast iron. Around that sits a thick outer shell of copper, which provides resistance to corrosion. The lid is attached using friction-stir welding to create a tightly sealed container.

Each canister will then be surrounded by bentonite clay. When the clay absorbs groundwater, it swells, fills gaps and forms a dense buffer around the copper. The deposition tunnel itself will eventually be filled and sealed. Beyond those engineered barriers sits the surrounding bedrock, which slows the movement of groundwater and any radioactive material that might one day escape.

Emily Stein, a researcher in geological waste disposal at Sandia National Laboratories, described the logic plainly: “You’re never relying on a single barrier.” If one layer performs less effectively than expected, the others are meant to continue delaying or preventing the release of radionuclides.

The copper gets the most attention because it is expected to remain the main corrosion barrier for an almost unimaginable length of time.

Groundwater already exists in fractures at repository depth, so the system is not based on keeping every drop of water away from the canisters. Instead, the developers expect the deep groundwater to become chemically reducing and extremely low in dissolved oxygen. Oxygen trapped during construction should gradually be consumed by reactions in the rock, clay and groundwater, removing one of the main drivers of copper corrosion.

The copper-corrosion dispute

Not everyone accepts that account. The best-known critic is Peter Szakálos, a corrosion researcher associated with KTH Royal Institute of Technology in Stockholm.

In a 2007 study with Gunnar Hultquist and Gunnar Wikmark, Szakálos reported hydrogen production in experiments involving copper and deoxygenated pure water. The researchers interpreted the result as evidence that copper could react with water itself even in the absence of dissolved oxygen.

Szakálos has argued that this process, combined with other forms of degradation, could eventually damage the canisters. Speaking to Science, he offered a particularly blunt prediction: “It’s just a matter of time—between decades and centuries—before unalloyed copper canisters start to crack at Onkalo.”

That prediction should not be mistaken for the scientific consensus. It is a disputed minority interpretation.

SKB and Posiva argue that the experiments do not represent the chemical and physical conditions the copper will face inside saturated bentonite hundreds of metres underground. Follow-up experiments and reviews have challenged both the proposed corrosion mechanism and the amount of copper loss inferred from the original work.

In a 2019 review of canister-integrity research, SKB concluded that there was no reason to expect corrosion in pure, oxygen-free water to proceed beyond the small amounts predicted by established thermodynamic data. Other researchers and regulatory reviewers have continued to examine issues including sulphide corrosion, stress-corrosion cracking, radiation effects and hydrogen-related damage.

Whether 100,000 years of certainty is possible

Finland’s regulator has looked far beyond the period for which any engineering experiment could possibly run. Safety assessments consider possible changes in groundwater, climate, glaciers, earthquakes, corrosion and the movement of radioactive material through the bedrock and eventually into the surface environment.

Jarkko Kyllönen, a nuclear-safety expert at STUK who has worked on scenarios stretching as far as one million years into the future, has said that “the first 10,000 years are very important for keeping the capsules intact.”

That early period matters especially because the spent fuel is hotter and more intensely radioactive then. Its heat output and hazard decline over time, although the safety assessments continue much farther into the future.

Critics are less confident. Tapani Veistola of the Finnish Association for Nature Conservation has said, “No one can guarantee the safety of Onkalo for thousands of years.”

That sits against the more positive assessments produced by Posiva and examined by STUK. Both ideas can be true at once: the safety case can be extensive and carefully constructed, while an absolute guarantee over 100,000 years remains impossible.

Finland’s plan does not require the repository to be continually maintained for 100,000 years. Once the final tunnels have been filled and sealed, the safety concept is meant to become passive. It should not depend on future workers replacing equipment, pumping out water or repairing damaged containers.

What Finland has committed to is a piece of engineering intended to remain safe long after the institutions that designed it may have disappeared. No experiment can run for 100,000 years, so the case rests on geological evidence, laboratory studies, observations of natural processes, computer models and deliberately overlapping barriers.

The copper might not remain untouched. That is not what the design requires. The clay may change, groundwater may move and future ice sheets may press down on the bedrock again. The aim is that these changes happen slowly enough, and behind enough separate barriers, that any eventual release remains extraordinarily small.

What is being asked for is not absolute certainty. It is trust in metal, clay and stone to protect people for an amount of time human institutions have never had to comprehend.