The image is a resonant one and rests indelibly in the mind. The plant sits, fuming restlessly, while all around an inept Soviet bureaucracy crumbles into the ensuing chaos. No one knows how many will have died as a result of the radioactive cloud expelled by Chernobyl's number 4 reactor on 26 April 1986; the fact that people are still debating it (see pages 982 and 993) says enough.

Skip forward 20 years, and nuclear power is edging back into vogue. It wasn't just Chernobyl that drove it out of favour, of course: the Three Mile Island incident in Pennsylvania in 1979, a catalogue of economic and technical setbacks in several nations, and the surprising resilience of fossil fuels as a cheap and available source of energy had already seen to that.

As memories of these mishaps recede, other factors have arisen to bring nuclear power back into play. Energy prices are high again, and governments are seeking to tackle climate change by limiting fossil-fuel emissions. Economists from California to Calcutta are looking at pie-charts of their future energy supply and saying that nuclear power needs to play a role. Will it?

The answer is a qualified yes — provided that governments absorb the true lesson of Chernobyl. This is not that nuclear power is unsafe, but that it is unsafe in the hands of a corrupt, unaccountable, irresponsible political system that fails to take reasonable measures to protect its citizens. The future of nuclear energy does not hinge primarily on the development of a safer reactor or a more geologically reliable waste repository, but on the ability of states to build public trust in their ability to safely implement and manage the technology.

Building trust

This trust can be achieved in different ways. In France, where the public appreciates the centralized technocracy that brought it high-speed trains, Concorde and an independent nuclear deterrent, nuclear power is ubiquitous and widely accepted. Scandinavian nations have a different political tradition, in which inclusive decision-making may soon open the way for the world's first permanent nuclear-waste repositories. Elsewhere, however, the future of nuclear energy is uncertain. The mechanisms that will assure its acceptance are not yet in place.

The key elements of this equation are the same the world over: nuclear power's real and perceived links with nuclear weapons; the available technology for power generation; its safety and economics; and options for clean-up and waste disposal.

Recent events in Iran serve as a painful reminder of the interaction between nuclear power and nuclear weapons. From the days of the ‘Atoms for Peace’ movement in the 1950s, advocates of the former have sought to separate the two, but they are inextricably linked. Public acceptance of nuclear power in Europe, Japan and the United States would benefit from a credible strategy to contain the proliferation of nuclear weapons. Expansion of nuclear power elsewhere, on the back of further proliferation of nuclear weapons, could have disastrous consequences.

The technology of nuclear power plants continues to improve (see Nature 429, 238–240; 2004). Chernobyl was a vastly archaic reactor whose safety systems, regulation and management were not close to acceptable standards. Modern working reactors are not susceptible to Chernobyl-style accidents, and some designs now under consideration could be safer still.

But safe reactor design, unfortunately, provides little protection against current fears — floated again this week in a sceptical report from a committee of British members of parliament — that nuclear plants may be vulnerable to terrorist attack.

Counting the cost

Restoring sites to a pristine state may be prohibitively expensive, as may the permanent disposal of waste and spent fuel.

The economics of such plants are subject to fierce debate (see page 984). The deregulated power-generation markets that have taken shape over the past two decades have little appetite for nuclear power's combination of high build costs, low running costs and uncertain future liabilities. Old debates about how many cents it costs to produce electricity have been superseded by a more subjective discussion about what goes in the bucket labelled ‘costs’. Restoring sites to a pristine state may be prohibitively expensive, as may the permanent disposal of waste and spent fuel. The construction of new plants will require either financial guarantees from the state (in Britain or the United States) or direct government involvement (in India or China).

Finally, nuclear-waste disposal remains the industry's Achilles' heel. For governments that advocate nuclear power to offer no solution — and leave spent fuel and other waste on the surface for future generations to deal with — is an abdication of responsibility. The threat of terrorist attack on nuclear power stations, as well as the risk that their spent fuel could be stolen and used for such an attack, renders the notion of long-term, localized waste storage at multiple sites even less tenable than it was before.

Different approaches are being taken to the management of waste disposal. Yucca Mountain in Nevada is in grave danger of becoming an expensive monument to failure (see page 987). The site was selected by default when Nevada was too weak to remove itself from the process — a hopeless, unscientific approach that may now reap what it sowed. Scandinavian states are doing a little better; Finland is winning support for a repository on the basis of a continuing nuclear energy programme, and Sweden is doing so on a promise to wrap the whole thing up. France has made some headway in site selection and can be relied on to address the issue with its customary determination. Britain has to start again from scratch, and is using its Committee on Radioactive Waste Management as an interesting, if not entirely convincing, experiment in public consultation.

So far, India and China, the biggest likely builders of nuclear power stations in the next 20 years, don't have much to say about waste disposal. Time will tell if either of them can handle the issue in an environmentally responsible way. However, if national pride in nuclear technology is a significant factor, the French example suggests that nuclear power has a solid future in Asia, with or without a waste repository.

In the West, however, the future options for nuclear power are far narrower than the heat of the current debate would suggest. Abandonment, as embraced fleetingly by the previous German government, isn't going to happen. The kind of major build-up envisaged before Three Mile Island and Chernobyl (see Nature 244, 392; 1973 and Nature 257, 346; 1975) isn't coming either.

Instead, nations are likely to tread a path somewhere between replacing some existing nuclear power capacity and its mild augmentation. Given global warming, high energy costs and doubts about the reliability of the oil supply, the latter approach has much to commend it, although it should not be pursued at the expense of renewable energy.

Nuclear energy's technical elegance has always appealed to the hearts and minds of scientists and engineers, who have been unusually prominent among its public advocates for half a century. Throughout, these advocates have promised to present to the public a clean and complete nuclear fuel cycle. Now it is time to stand and deliver.