Research and version note This is a version 0.1 research draft in The Future Has No Representative. “One hundred thousand years” indicates the very long safety horizon of high-level radioactive waste; it is not a single duration common to every waste, radionuclide, or regulatory regime. The article makes no safety finding for a particular site. Technical and project information was reviewed to August 2026.
The Future Has No Representative · Article 6
Quick read
The nuclear-waste problem is often presented as a question of how to make a sign saying “danger—do not enter” intelligible in a hundred thousand years. That is a real question, but it can obscure a more important engineering principle. The long-term safety of geological disposal must not depend on a government, language, or warning remaining continuous for the entire period. International Atomic Energy Agency standards use stable geology, the waste form, containers, and engineered barriers for containment and isolation. Safety after closure is assessed without assuming that institutional control will remain effective indefinitely.
Records, knowledge, and memory are still important. They reduce the chance of inadvertent drilling or excavation and allow later societies to understand where waste was placed, why the decision was made, and whether to monitor further. The OECD Nuclear Energy Agency’s intergenerational RK&M project concluded that no marker or archive can span every timescale on its own. Records, land registers, regulatory practices, education, site markers, and widely distributed key information should operate together.
Finland’s ONKALO is among the geological repositories closest to operation. As of August 2026, Posiva was still testing and preparing for all required permissions; the project should not be described as having completed the demonstration of long-term safety. Its design places encapsulated spent fuel in bedrock at roughly 430 metres. The safety case rests on multiple barriers rather than confidence that every future generation will preserve present institutions.
No existing organisation can promise its own survival for a hundred thousand years. What institutions can do is layered. They can maintain funding, oversight, knowledge, and accountability over decades and centuries; design passive safety that does not require that oversight forever; and send essential information through multiple media as far as possible. Later generations should be able to update the record and reconsider action.
Human beings lack an institution that remains unchanged over the whole period, and should not pretend otherwise. We can build an arrangement whose safety is not staked on institutional immortality and improve the transfer of resources and knowledge while institutions do exist. A warning cannot be guaranteed to retain meaning, but the society that produced the waste still owes one. Silence would expose the future to avoidable ignorance.
Why a hundred thousand years exceeds ordinary political time
One hundred thousand years ago there were no modern states, surviving religious institutions, or languages used today. Writing has existed for only several millennia. A digital format can lose software support in decades. A demand that one marker remain correctly understood at the far horizon goes beyond anything that history allows us to verify directly.
High-level waste and spent fuel contain radionuclides whose activity, heat, and hazard change on different paths. A hundred-thousand-year horizon commonly communicates that safety assessment must cover an exceptionally long period. It does not mean that danger suddenly ends in the following year or that all radioactive waste requires identical isolation. The relevant period, dose criteria, and scenarios depend on waste, disposal design, and regulation.
A decision is nevertheless unavoidable. Continued surface or near-surface storage needs security, power, maintenance, and stable organisations. Geological disposal requires a site, geological evidence, engineered barriers, and eventual closure. Refusing a permanent choice does not remove the waste; it postpones active management obligations.
The philosophical difficulty follows from the horizon. Decision-makers will never observe the final outcome. They use experiments, geological history, material models, and conservative scenarios to construct a safety argument. Other people inherit the consequence and cannot return to the time before the waste was generated. Because verification cannot cover the whole period, institutions must distinguish observed evidence, modelled inference, and the normative safety standard.
Why disposal safety cannot depend on permanent guards
IAEA SSG-14, Geological Disposal Facilities for Radioactive Waste, describes geological disposal as isolation of solid radioactive waste in a facility underground in stable formations, providing long-term containment and separation from the accessible biosphere. Its distinctive long-run mission is to be achieved through passive safety. Assessment cannot assume that operational controls and institutional measures remain effective for thousands of years and longer.
The principle acknowledges institutional mortality. A country can split, a regulator be reorganised, and finance or expertise deteriorate. A repository that remains safe only if every generation pumps, repairs, and guards indefinitely has converted weaknesses in technical design into an unending political duty.
Passive safety generally employs several barriers. Waste form, metal container, buffer, backfill, and surrounding rock delay water access, release, and movement by different mechanisms. No component is simply declared immortal. Overall performance relies on different barriers across time. A safety case assembles site data, experiments, models, scenarios, and uncertainty to show why regulatory standards are satisfied.
This is not a direct experiment over a hundred thousand years. A model can omit a process, while material and groundwater behaviour remain uncertain. Independent regulation, peer scrutiny, staged licensing, and opportunities to stop before closure matter for that reason. Passive safety does not mean burying and forgetting. It means refusing to make future life depend on an organisation that cannot be guaranteed to survive.
Why warn if the engineering must not depend on memory?
Geological disposal is designed chiefly around natural evolution, but later people may drill for resources, sink a water bore, or excavate another underground project. A society unaware of a repository could penetrate its barriers deliberately but unknowingly. Records and markers can reduce this inadvertent intrusion.
They also preserve options. A later society may have better means of monitoring, remediation, or retrieval. To decide, it needs the inventory, canister positions, and original safety reasoning. Complete forgetting has sometimes been imagined as safest because no one would interfere, but it also removes the possibility of informed later judgment.
The OECD Nuclear Energy Agency’s Records, Knowledge and Memory Across Generations initiative distinguishes mediated transmission, in which one generation hands information to the next, from non-mediated transmission intended to reach a distant receiver in its original form. Its central finding is not a perfect sign. No single mechanism covers every timescale, so technical, administrative, and social methods require a systemic strategy.
The two aims can conflict. A prominent warning may deter intrusion but advertise something exceptional underground and invite curiosity or recovery. A technical archive contains more information but depends on language, expertise, and preservation. A simple skull may not have the same cultural meaning in the future. Redundancy and diversity can reduce common failure; no design can guarantee that meaning remains unchanged.
What should be left behind?
The first layer is exact information: waste inventory and location, canister and geological data, construction changes, monitoring, incidents, and licensing decisions. Future specialists should be able to reconstruct the disposal system rather than receive a bare warning. Uncertainties and unresolved issues belong in the record so a safety conclusion does not persist as authority after losing its reasons.
A second layer provides key information for wider circulation. The NEA project developed concepts including a Key Information File and a Set of Essential Records, selecting indispensable content from a large specialist archive. Copies can be distributed across national archives, local land records, international institutions, and several media, reducing a single point of loss.
Land-use records, markers, local memory, education, and continuing monitoring provide another layer. While institutions are stable, information can be translated and renewed as languages and technologies change. If mediation eventually fails, durable material and dispersed copies may still leave clues.
Selection and preservation are not purely engineering choices. Some details may be security-sensitive. Excessive secrecy prevents scrutiny; unrestricted detail may facilitate malicious access. Local communities live with the spatial consequence and should participate in determining how knowledge is maintained. On Indigenous land, sovereignty, culture, and intergenerational relations to country cannot be cancelled by invoking a global future interest.
What ONKALO demonstrates—and what it does not
Posiva is developing the ONKALO disposal system at Olkiluoto, Finland. Spent fuel is intended to be placed in a copper-and-cast-iron canister, deposited in holes in bedrock at about 430 metres, and surrounded by bentonite and other barriers. Decades of research, site investigation, and licensing have made it an important case in moving geological disposal toward implementation.
Its current status needs precision. By August 2026, Posiva had completed or advanced full-process trials using non-radioactive material and reported underground installation milestones in June. The company’s objective was to begin actual disposal during 2026 once all licences and permissions were in place, but its current project update still described the work as progress toward operation. This article does not say that spent-fuel disposal was already under way.
Posiva explains long-term safety through a safety case: scientific information, analysis, observations, trials, and other evidence support the assessment. The case cannot directly prove a hundred thousand years in human time, but it makes assumptions and evidence available for regulatory examination.
ONKALO also shows that a far-future design is implemented by a present community in a particular place. Engineering maturity does not conclude the ethical questions. Whether producers have funded the cost, whether local acceptance was just, how closure and retrievability are balanced, and who preserves records all require institutional answers. A strong barrier system cannot by itself solve representation and distribution.
How long can funding last?
Research, operation, monitoring, and closure need long-term finance. Establishing funds from waste producers while revenue exists usually fits responsibility better than asking taxpayers centuries later to improvise. Governance of those funds, investment risk, cost estimation, and protection against diversion remain significant.
Even a well-designed fund cannot promise a hundred thousand years. Money, legal persons, and financial markets will not retain their current form indefinitely. Finance is best suited to the governable stages: investigation, construction, operation, closure, and the post-closure period required by regulation. Beyond them, safety should primarily be passive.
Responsibility is layered accordingly. Near-term institutions should be strong because they can be inspected, corrected, and resourced. Engineering should anticipate that remote institutions weaken. Information systems try to carry intelligibility beyond both. Placing every hope in the fund, geology, or marker misunderstands the different temporal capacity of each mechanism.
Underfunding creates moral hazard. Present electricity users receive the benefit of nuclear generation, while an underestimated disposal bill can be left to a later public. A decision about nuclear energy should include waste management and decommissioning rather than assigning zero to a cost merely because it is distant and uncertain.
May the future change our solution?
Permanent disposal may appear to let the present make an irrevocable choice. Continuing surface storage leaves later people a continuing security and management burden. Both constrain the future in different ways.
Staged disposal preserves some capacity to reconsider. Investigation, excavation, trial, emplacement, and closure receive decisions at different points, allowing new evidence to halt progress. Some regimes preserve retrievability for a period, although retrieval consumes resources and may increase operational risk. Reversibility is not a cost-free ethical advantage.
Future freedom cannot mean leaving every option open forever. Any safety measure restricts land uses and technological paths. A more credible principle avoids unnecessary irreversibility for present convenience. Where closure has a strong safety basis, decision-makers should explain why it is preferable to indefinite active management and preserve information that supports later understanding and review.
Nor should policy assume that future technology will solve the waste. Leaving spent fuel in storage because successors will be cleverer builds current responsibility on optimism to which they did not consent. Claiming that present disposal eliminates every future issue would exceed the evidence in the opposite direction. A candid safety case includes model uncertainty and institutional failure.
Provisional judgment: admitting institutional disappearance is the beginning of responsibility
Human beings have no government, company, language, or archive guaranteed to operate continuously for a hundred thousand years. If institutional capacity means one organisation remaining identical and understanding unchanged, we do not possess it. Nuclear-waste safety cannot rely on it.
What we have is a more limited combinatorial capacity. Present societies can construct passive isolation, establish independent regulation and staged permission, fund the governable period, and preserve records in multiple places, media, and levels of explanation. While mediation works, each generation can update language, verify condition, and hand knowledge to the next. No layer is permanent; together they reduce common failure.
This does not make the present an authorised representative of the future. We do not know which technologies will exist or how later societies will assess nuclear power. We do know who produced the hazard and who controls disposal resources now. Causal power and control establish a present responsibility without waiting for a future request.
A warning may cease to mean what we intended. An archive may be scattered and a state disappear. Those possibilities are not reasons to leave no information. They are reasons not to use information as the sole barrier, while still doing everything feasible to prevent later people from encountering the underground inheritance in complete ignorance. A mature long-term institution does not promise immortality. It incorporates its own eventual interruption into the safety structure.
Primary sources and further reading
- International Atomic Energy Agency, Disposal of Radioactive Waste, SSR-5.
- International Atomic Energy Agency, Geological Disposal Facilities for Radioactive Waste, SSG-14.
- OECD Nuclear Energy Agency, Preservation of Records, Knowledge and Memory Across Generations and its final report.
- Posiva, Final disposal and 2026 testing update.
- US Environmental Protection Agency, EPA’s Role at the Waste Isolation Pilot Plant, an example of a 10,000-year regulatory protection period.
Series navigation: The Future Has No Representative — series overview
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.