Nuclear represents high risks at multiple areas of operation:
- The fuel, waste, and plants are all radioactive, in some cases for geological time periods, presenting unprecedented disposal issues. Perhaps not insurmountable, but very, very significant.
- Scale of operations of nuclear plants means that the plants themselves, their operations, and their revenues are immense. A large nuclear plant generates over $2 billion worth of electricity (retail) annually. As we've learned with other large financial systems, profit maximization can skew risk management objectives.
- Risks of scale include management, financial, terrorist, natural disaster, labor action, and other factors. Any of these could put, at the least, power reliability at risk, and at worst, health and safety for millions of people. One consequence of the 2011 Tōhoku earthquake and tsunami in Japan was TEPCO's loss of 27% of its generating capacity (http://www.pcworld.idg.com.au/article/379610/japan_facing_ma...). Exacerbated by Japan's dual-cycle grids (eastern and western Japan operate on 50 and 60 Hz respectively: http://www.wired.com/gadgetlab/2011/03/tech-legacy-tokyo/).
- The timescale and emergent phenomena associated with nuclear power mean that risks aren't known, appreciated, and/or understood at construction time. Chernobyl, TMI, and Fukushima all exhibited fundamental engineering and/or operational flaws. Siting of the Fukushima plant turned out to be problematic. Several California plants are now known to be on or near active seismic zones. We still don't have a permanent waste disposal solution in the US (and much of the rest of the world).
Most alternate forms of renewable / sustainable power are much more diverse and distributed. Though this means higher capital and maintenance costs per MWh of generation, it also means vastly smaller risks of any individual plant failure, and far lower likelihood of cascade failures.
Interesting times.
But my major problem with most forms of nuclear is that from a geopolitical standpoint, if you are proposing a technology for clean global power, it has to by definition be something you would be able to share freely with a potential enemy. Most current nuclear power technologies, including the pebble bed, conspicuously fail at this.
No, it's mostly just fear.
Personally, I consider the current generation of fission reactors technologically unsound: Neither the problem of radiactive waste nor shutdown in case of catastrophic failures are adequately solved.
Research in Thorium reactors is worthwhile considering the growing energy needs of China and India, but smarter transnational grids with more efficient load balancing can probably make renewable energy production viable in Europe.
As far as new breeder designs, it's a bit too late. Nuclear promised us "energy too cheap to meter" and construction and maintenance costs have kept going down. Meanwhile solar and wind have been regularly going down in price with each doubling of production.
The case against nuclear isn't environmental; it's economic.