On Olkiluoto Island, off Finland’s west coast, a tunnel runs 430 metres down into bedrock that has sat largely undisturbed for about 1.9 billion years. The facility is called Onkalo. When it is full, the plan is to seal it, restore the surface, and leave nothing behind to indicate what is underneath. The reasoning, on the operator’s own account, is that any marker is more likely to attract attention than to deter it.

The clearest long-form treatment of the project we have come across is an essay published at Only Sky, which sets out both the engineering of Onkalo and the strange design problem that sits on top of it. The waste in question stays dangerous for roughly 100,000 years. Recorded human history is shorter than that by a factor of about twenty.

What is actually being built

Onkalo is run by Posiva, a company jointly owned by the two Finnish utilities that operate the country’s nuclear reactors. The facility is designed to hold spent fuel from those reactors. According to Posiva’s published materials, the spent fuel will be sealed inside copper canisters, the canisters will be set inside a bentonite clay buffer, and the buffer will be placed in deposition holes drilled into the granite at the base of the tunnel system. The granite, the clay and the copper are the three engineered barriers. The bedrock itself is the fourth.

The choice of site was driven by the geology. The Olkiluoto bedrock is part of the Fennoscandian shield, one of the older and more stable pieces of continental crust on the planet. It has not seen major tectonic activity for a very long time, and the groundwater chemistry at depth is reducing rather than oxidising, which slows copper corrosion. None of this is novel science. What is novel is the willingness to actually build the thing.

Finland began studying deep geological disposal in the 1980s. Site selection ran through the 1990s. Construction of the underground research and characterisation facility started in 2004. The operating licence for final disposal was granted by the Finnish government in 2015. First emplacement of spent fuel was originally planned for the early 2020s and has been pushed back. According to Posiva’s own statements, the timeline has slipped, and the company now points to the mid-2020s for the start of operational disposal. Industrial projects of this scale slip. This one has slipped less than most.

The marker question

The part of the Onkalo story that has drawn the most outside attention is not the engineering. It is the decision about what to put on top of the site once it is full. The dominant answer, in the documents Posiva and the Finnish regulator have published, is: nothing.

The reasoning runs roughly as follows. A 100,000-year horizon is long enough that no language, symbol system or institution can be trusted to carry meaning across it. Any structure built to warn future people will, on the available historical evidence, attract them. Pyramids were looted. Burial mounds were dug into. A monument that says “do not dig here” reads, across a sufficient gap in time, as a sign that something worth finding is buried at this spot. The safer bet, on the operator’s reading, is to restore the surface to ordinary forest and let the location fade from active memory.

This is not the only school of thought in the long-term waste storage literature. The United States Department of Energy spent significant time and money in the 1990s working on marker concepts for the Waste Isolation Pilot Plant in New Mexico, including proposals involving granite spikes, hostile architecture, and layered warning messages in multiple languages and pictograms. Sandia National Laboratories convened panels of linguists, anthropologists and semioticians for the work. The Finnish position is that those efforts illustrate the problem rather than solve it.

What “100,000 years” actually means

The figure is not arbitrary. It comes from the decay profile of the spent fuel and the regulatory standard set by the Finnish Radiation and Nuclear Safety Authority, known as STUK. The standard requires that the disposal system limit annual radiation doses to the surface environment below specified levels for a defined assessment period. The assessment period extends to the point at which the waste no longer poses a meaningfully elevated risk above natural background radiation. For the fuel in question, that point sits roughly 100,000 years out.

It is worth holding that number against the human reference frame. The oldest continuously occupied cities are, on the available archaeological evidence, somewhere between 5,000 and 10,000 years old. The earliest writing systems are roughly 5,000 years old. A 100,000-year design horizon is not a longer version of an ordinary engineering problem. It is a different kind of problem.

What the project is and what it is not

Onkalo is the first deep geological repository for spent nuclear fuel anywhere in the world to reach this stage of construction. Sweden’s KBS-3 design, which Posiva’s design closely follows, is at an earlier stage. France, Switzerland, Canada and the United States all have programmes at various stages of site selection or regulatory review. By the team’s read of the published reporting, none of them are as far along as Finland.

It is also worth being precise about what Onkalo does and does not solve. It addresses spent fuel from existing Finnish reactors. It does not address the larger global inventory of spent fuel sitting in cooling pools and dry casks at reactor sites around the world. It does not settle the question of whether deep geological disposal is the right answer in geologies less obliging than the Fennoscandian shield. It does not retire the political argument about nuclear power. It is a single facility, built carefully, in a country that decided to build it.

What to watch next

The next milestone is operational disposal: the first canisters of spent fuel actually emplaced in the deposition holes, the bentonite packed in, the tunnel sealed behind them. Posiva has said publicly that this work is imminent. Whether “imminent” means this year or the next is, on the available material, the kind of question a regulator answers and not a publication.

The marker question, in practical terms, is decades away. The repository is meant to operate for roughly a century before final sealing. The decision about what, if anything, to leave on the surface does not have to be made now.

It will, eventually, have to be made.