Japan’s Underground Lab Wants to Solve the World’s Nuclear Waste Problem

Reindeer pastures and dairy farms stretch across the flat, windswept plain of Japan’s northernmost prefecture, giving little hint of what lies roughly 300 meters below. Inside the Horonobe Underground Research Laboratory, teams of engineers are trying to answer a question no country has fully solved: how to keep highly radioactive waste isolated from the environment for periods measured not in decades, but in tens of thousands of years.

In May, I gained special access to the facility through the Japan Atomic Energy Agency (JAEA), touring the underground galleries and observing several of the experiments currently underway.

A Rare Laboratory in Sedimentary Rock

Horonobe is one of only two underground research laboratories Japan has built to study geological disposal of high-level radioactive waste, and the only one still operating. Its counterpart, the Mizunami Underground Research Laboratory in Gifu Prefecture, was built in crystalline rock and was decommissioned in 2022 after its land lease expired. Horonobe, by contrast, sits in Neogene-era sedimentary rock, giving the JAEA a very different and, officials argue, complementary geological environment to study.

The project traces back to November 2000, when the governor of Hokkaido, the mayor of Horonobe town, and the president of the JAEA signed an agreement to permit research on the deep geological environment at the site. Construction and excavation followed over the next two decades, and the laboratory now includes shafts and galleries used to test how engineered barriers, groundwater, and surrounding rock interact over time – the same combination of natural and artificial safeguards that would, in theory, protect a real repository.

Two of the central experiments underway involve what the JAEA calls the “full-scale engineered barrier system performance experiment” and a “solute transport experiment,” both designed to test, at real scale, how bentonite buffers, canisters, and cement-based materials behave once water and pressure are introduced over years rather than months. The JAEA has also been testing how the surrounding sedimentary rock buffers natural disturbances, earthquakes, and groundwater shifts that could compromise a repository’s isolation over the long term.

Don’t Call It a Dump

The JAEA is emphatic on one point, repeated in nearly every briefing given to visitors: Horonobe is a research facility, not a candidate site for Japan’s eventual high-level waste repository. Japanese law formally separates the two functions, and the agency overseeing the actual site-selection process, the Nuclear Waste Management Organization of Japan, operates independently of the JAEA’s research program.

That legal separation has not entirely dissolved local unease. Horonobe is a farming town of roughly 2,000 residents, many of whom have lived for two decades with an underground laboratory in their backyard and, despite official assurances, the lingering question of what happens to the facility once the research concludes.

Japan’s broader search for a permanent repository has proven difficult everywhere it has been raised. Only two municipalities, Suttsu and Kamoenai in Hokkaido, have so far agreed even to preliminary literature-based surveys, and neither has committed to hosting a final site.

An International Proving Ground

What distinguishes Horonobe today is its growing international role. In February 2023, the JAEA and the OECD’s Nuclear Energy Agency (NEA) launched the Horonobe International Project, bringing foreign research organizations into the underground laboratory to work on shared technical problems. The first phase ran through March 2025 with 11 participating organizations from countries including Australia, Bulgaria, Germany, South Korea, Romania, Taiwan, and the United Kingdom, alongside Japan. 

A second phase, running from April 2025 through March 2029, has continued with roughly 10 member organizations pursuing three joint technical tasks: modeling how dissolved substances move through fractured sedimentary rock, integrating different repository design options, and dismantling a full-scale engineered barrier system to see how it performed after years underground.

For countries such as Australia and South Korea – both still in the early stages of building their own waste-management institutions – Horonobe offers something they cannot yet replicate at home: a working underground facility where design assumptions can be tested in real-world conditions. South Korea’s state nuclear research institute has said it is benchmarking its own disposal-performance models against Horonobe data as it prepares to build an underground laboratory of its own in Taebaek. 

NEA officials have described the site as an important international platform for developing and demonstrating repository technology, with plans to eventually extend the laboratory to 500 meters below sea level.

The Finland Comparison

It is difficult to walk through Horonobe without thinking of Onkalo, the repository under construction on the island of Olkiluoto in western Finland, which is set to become the world’s first operational deep geological disposal site for spent nuclear fuel. Having visited Onkalo myself, the contrast with Horonobe is immediately clear. Onkalo is being excavated roughly 400 meters into Precambrian granite bedrock nearly 1.9 billion years old – rock that geologists consider to have remained structurally stable through repeated glacial cycles. Horonobe, by design, tests young, still-consolidating Neogene sedimentary rock, chosen precisely because it represents a geological setting Japan cannot avoid when it eventually has to site a real repository. Much of the country lacks Finland’s kind of ancient crystalline bedrock.

The difference between the two sides is as much political as geological. Posiva Oy, the utility consortium that owns and operates Onkalo, spent roughly two decades developing its technical solution and benefited from a Finnish system in which nuclear waste responsibility stays under public oversight even where the industry itself is privately run, with the national regulator STUK maintaining permanent inspectors on site. That framework was paired with sustained public consultation and a compensation system for the host municipality, Eurajoki, that turned three decades of engagement into local acceptance rather than opposition. Waste is due to begin arriving at Onkalo in 2026, sealed in copper-and-bentonite canisters designed to maintain their integrity for roughly 100,000 years.

Japan, by contrast, has yet to find a single municipality willing to host a final repository, and Horonobe’s own officials go out of their way to stress that the laboratory is not a step toward designating the town as one. 

Finland’s example has not gone unnoticed in Tokyo: Japanese institutions are already among the countries cooperating with Posiva on adapting Finnish containment technology to different geological and regulatory contexts, a channel that runs alongside, but separately from, the JAEA’s own international partnerships through the Horonobe International Project.

That gap helps explain why Horonobe matters beyond Japan’s borders. It is not a repository and, officially, will never become one. But it is one of a small number of places on Earth where engineers can watch, in real time and at full scale, whether the technology underpinning every country’s nuclear waste strategy actually behaves the way computer models predict over years of exposure to rock, water, and pressure.