New nuclear power plants typically have high capital costs for building the first several plants, after which costs tend to fall for each additional plant built as the supply chains develop and the regulatory processes improve. Fuel, operational, and maintenance costs are relatively small components of the total cost. The long service life and high productivity of nuclear power plants allow sufficient funds for ultimate plant decommissioning and waste storage and management to be accumulated, with little impact on the price per unit of electricity generated. Additionally, measures to mitigate climate change such as a carbon tax or carbon emissions trading, would favor the economics of nuclear power over fossil fuel power.
Nuclear power construction costs have varied significantly across the world and in time. Large and rapid increases in cost occurred during the 1970s, especially in the United States. A single study has claimed that these cost increases are specific to the United States. However, the study has been criticized for the selective use of data and biased interpretation. There were no construction starts of nuclear power reactors between 1979 and 2012 in the United States, and since then more new reactor projects have gone into bankruptcy than have been completed. Recent cost trends in countries such as Japan and Korea have been very different, including periods of stability and decline in costs.
In more economically developed countries, a slowdown in electricity demand growth in recent years has made large-scale power infrastructure investments difficult. Very large upfront costs and long project cycles carry large risks, including political decision making and intervention such as regulatory ratcheting. In Eastern Europe, a number of long-established projects are struggling to find financing, notably Belene in Bulgaria and the additional reactors at Cernavoda in Romania, and some potential backers have pulled out. Where cheap gas is available and its future supply relatively secure, this also poses a major problem for clean energy projects. Former Exelon CEO John Rowe said in 2012 that new nuclear plants in the United States "don't make any sense right now" and would not be economic as long as gas prices remain low.
Current bids for new nuclear power plants in China have fallen below $2000/kW in 2016, as China continues accelerating its new build program after a pause following the Fukushima disaster. Therefore, comparison with other power generation methods is strongly dependent on assumptions about construction timescales and capital financing for nuclear plants.
Analysis of the economics of nuclear power must take into account who bears the risks of future uncertainties. To date all operating nuclear power plants were developed by state-owned or regulated utility monopolies where many of the risks associated with political change and regulatory ratcheting were borne by consumers rather than suppliers. Many countries have now liberalized the electricity market where these risks, and the risk of cheap competition from subsidised energy sources emerging before capital costs are recovered, are borne by plant suppliers and operators rather than consumers, which leads to a significantly different evaluation of the risk of investing in new nuclear power plants.
Two of the four EPRs under construction (the Olkiluoto Nuclear Power Plant in Finland and Flamanville in France), which are the latest new builds in Europe, are significantly behind schedule and substantially over cost. Following the 2011 Fukushima Daiichi nuclear disaster, costs are likely to go up for some types of currently operating and new nuclear power plants, due to new requirements for on-site spent fuel management and elevated design basis threats.
Although the price of new plants in China is falling rapidly, approaching $1500/kW (or about a fifth of the cost of some plants currently being built in Europe), John Quiggin, an economics professor, maintains that the main problem with the nuclear option is that it is not economically viable. Professor of science and technology Ian Lowe has also challenged the economics of nuclear power. However, nuclear supporters continue to point to the historical success of nuclear power across the world, and they call for new reactors in their own countries, including proposed new but largely uncommercialised designs, as a source of new power. Nuclear supporters point out that the IPCC climate panel endorses nuclear technology as a low carbon, mature energy source which should be nearly quadrupled to help address soaring greenhouse gas emissions.
Some independent reviews keep repeating that nuclear power plants are necessarily very expensive, and anti-nuclear groups frequently produce reports that say the costs of nuclear energy are prohibitively high.
In 2012 in Ontario, Canada, costs for nuclear generation stood at 5.9¢/kWh while hydroelectricity, at 4.3¢/kWh, cost 1.6¢ less than nuclear. By September 2015, the cost of solar in the United States dropped below nuclear generation costs, averaging 5¢/kWh. Solar costs continued to fall, and by February 2016, the City of Palo Alto, California, approved a power-purchase agreement (PPA) to purchase solar electricity for under 3.68¢/kWh, lower than even hydroelectricity. Utility-scale solar electricity generation newly contracted by Palo Alto in 2016 costs 2.22¢/kWh less than electricity from the already-completed Canadian nuclear plants, and the costs of solar energy generation continue to drop. However, solar power has very low capacity factors compared to nuclear, and solar power can only achieve so much market penetration before (expensive) energy storage and transmission become necessary.
Many countries, including Russia, South Korea, India, and China, have continued to pursue new builds. Globally, 71 nuclear power plants were under construction in 15 countries as of January 2015, according to the IAEA. China has 25 reactors under construction but, according to a government research unit, China must not build "too many nuclear power reactors too quickly", in order to avoid a shortfall of fuel, equipment and qualified plant workers. According to the World Nuclear Association, the global trend is for new nuclear power stations coming online to be balanced by the number of old plants being retired.
In the United States, nuclear power faces competition from the low natural gas prices in North America. Former Exelon CEO John Rowe said in 2012 that new nuclear plants in the United States "don’t make any sense right now" and won't be economic as long as the natural gas glut persists. In 2016, Governor of New York Andrew Cuomo directed the New York Public Service Commission to consider ratepayer-financed subsidies similar to those for renewable sources to keep nuclear power stations profitable in the competition against natural gas.
A 2019 study by the economic think tank DIW found that nuclear power nowhere in the world has been profitable.
"The usual rule of thumb for nuclear power is that about two thirds of the generation cost is accounted for by fixed costs, the main ones being the cost of paying interest on the loans and repaying the capital..." 
Capital cost, the building and financing of nuclear power plants, represents a large percentage of the cost of nuclear electricity. In 2014, the US Energy Information Administration estimated that for new nuclear plants going online in 2019, capital costs will make up 74% of the levelized cost of electricity; higher than the capital percentages for fossil-fuel power plants (63% for coal, 22% for natural gas), and lower than the capital percentages for some other nonfossil-fuel sources (80% for wind, 88% for solar PV).
Areva, the French nuclear plant operator, offers that 70% of the cost of a kWh of nuclear electricity is accounted for by the fixed costs from the construction process. Some analysts argue (for example Steve Thomas, Professor of Energy Studies at the University of Greenwich in the UK, quoted in the book The Doomsday Machine by Martin Cohen and Andrew McKillop) that what is often not appreciated in debates about the economics of nuclear power is that the cost of equity, that is companies using their own money to pay for new plants, is generally higher than the cost of debt. Another advantage of borrowing may be that "once large loans have been arranged at low interest rates – perhaps with government support – the money can then be lent out at higher rates of return".
"One of the big problems with nuclear power is the enormous upfront cost. These reactors are extremely expensive to build. While the returns may be very great, they're also very slow. It can sometimes take decades to recoup initial costs. Since many investors have a short attention span, they don't like to wait that long for their investment to pay off."
Because of the large capital costs for the initial nuclear power plants built as part of a sustained build program, and the relatively long construction period before revenue is returned, servicing the capital costs of first few nuclear power plants can be the most important factor determining the economic competitiveness of nuclear energy. The investment can contribute about 70% to 80% of the costs of electricity. Timothy Stone, businessman and nuclear expert, stated in 2017 "It has long been recognised that the only two numbers which matter in [new] nuclear power are the capital cost and the cost of capital." The discount rate chosen to cost a nuclear power plant's capital over its lifetime is arguably the most sensitive parameter to overall costs. Because of the long life of new nuclear power plants, most of the value of a new nuclear power plant is created for the benefit of future generations.
The recent liberalization of the electricity market in many countries has made the economics of nuclear power generation less attractive, and no new nuclear power plants have been built in a liberalized electricity market. Previously a monopolistic provider could guarantee output requirements decades into the future. Private generating companies now have to accept shorter output contracts and the risks of future lower-cost competition from subsidised energy sources, so they desire a shorter return on investment period. This favours generation plant types with lower capital costs or high subsidies, even if associated fuel costs are higher. A further difficulty is that due to the large sunk costs but unpredictable future income from the liberalized electricity market, private capital is unlikely to be available on favourable terms, which is particularly significant for nuclear as it is capital-intensive. Industry consensus is that a 5% discount rate is appropriate for plants operating in a regulated utility environment where revenues are guaranteed by captive markets, and 10% discount rate is appropriate for a competitive deregulated or merchant plant environment; however the independent MIT study (2003) which used a more sophisticated finance model distinguishing equity and debt capital had a higher 11.5% average discount rate.
As states are declining to finance nuclear power plants, the sector is now much more reliant on the commercial banking sector. According to research done by Dutch banking research group Profundo, commissioned by BankTrack, in 2008 private banks invested almost €176 billion in the nuclear sector. Champions were BNP Paribas, with more than €13,5 billion in nuclear investments and Citigroup and Barclays on par with both over €11,4 billion in investments. Profundo added up investments in eighty companies in over 800 financial relationships with 124 banks in the following sectors: construction, electricity, mining, the nuclear fuel cycle and "other".
Construction delays can add significantly to the cost of a plant. Because a power plant does not earn income and currencies can inflate during construction, longer construction times translate directly into higher finance charges. Modern nuclear power plants are planned for construction in five years or less (42 months for CANDU ACR-1000, 60 months from order to operation for an AP1000, 48 months from first concrete to operation for an EPR and 45 months for an ESBWR) as opposed to over a decade for some previous plants. However, despite Japanese success with ABWRs, two of the four EPRs under construction (in Finland and France) are significantly behind schedule.
In the United States many new regulations were put in place in the years before and again immediately after the Three Mile Island accident's partial meltdown, resulting in plant startup delays of many years. The NRC has new regulations in place now (see Combined Construction and Operating License), and the next plants will have NRC Final Design Approval before the customer buys them, and a Combined Construction and Operating License will be issued before construction starts, guaranteeing that if the plant is built as designed then it will be allowed to operate—thus avoiding lengthy hearings after completion.
In Japan and France, construction costs and delays are significantly diminished because of streamlined government licensing and certification procedures. In France, one model of reactor was type-certified, using a safety engineering process similar to the process used to certify aircraft models for safety. That is, rather than licensing individual reactors, the regulatory agency certified a particular design and its construction process to produce safe reactors. U.S. law permits type-licensing of reactors, a process which is being used on the AP1000 and the ESBWR.
In Canada, cost overruns for the Darlington Nuclear Generating Station, largely due to delays and policy changes, are often cited by opponents of new reactors. Construction started in 1981 at an estimated cost of $7.4 Billion 1993-adjusted CAD, and finished in 1993 at a cost of $14.5 billion. 70% of the price increase was due to interest charges incurred due to delays imposed to postpone units 3 and 4, 46% inflation over a 4-year period and other changes in financial policy. No new nuclear reactor has since been built in Canada, although a few have been and are undergoing refurbishment and environment assessment is complete for 4 new generation stations at Darlington with the Ontario government committed in keeping a nuclear base load of 50% or around 10GW.
In the United Kingdom and the United States cost overruns on nuclear plants contributed to the bankruptcies of several utility companies. In the United States these losses helped usher in energy deregulation in the mid-1990s that saw rising electricity rates and power blackouts in California. When the UK began privatizing utilities, its nuclear reactors "were so unprofitable they could not be sold." Eventually in 1996, the government gave them away. But the company that took them over, British Energy, had to be bailed out in 2004 to the extent of 3.4 billion pounds.
In general, coal and nuclear plants have the same types of operating costs (operations and maintenance plus fuel costs). However, nuclear has lower fuel costs but higher operating and maintenance costs.
Nuclear plants require fissile fuel. Generally, the fuel used is uranium, although other materials may be used (See MOX fuel). In 2005, prices on the world market for uranium averaged US$20/lb (US$44.09/kg). On 2007-04-19, prices reached US$113/lb (US$249.12/kg). On 2008-07-02, the price had dropped to $59/lb.
Fuel costs account for about 28% of a nuclear plant's operating expenses. As of 2013, half the cost of reactor fuel was taken up by enrichment and fabrication, so that the cost of the uranium concentrate raw material was 14 percent of operating costs. Doubling the price of uranium would add about 10% to the cost of electricity produced in existing nuclear plants, and about half that much to the cost of electricity in future power plants. The cost of raw uranium contributes about $0.0015/kWh to the cost of nuclear electricity, while in breeder reactors the uranium cost falls to $0.000015/kWh.
As of 2008, mining activity was growing rapidly, especially from smaller companies, but putting a uranium deposit into production takes 10 years or more. The world's present measured resources of uranium, economically recoverable at a price of US$130/kg according to the industry groups Organisation for Economic Co-operation and Development (OECD), Nuclear Energy Agency (NEA) and International Atomic Energy Agency (IAEA), are enough to last for "at least a century" at current consumption rates.
According to the World Nuclear Association, "the world's present measured resources of uranium (5.7 Mt) in the cost category less than three times present spot prices and used only in conventional reactors, are enough to last for about 90 years. This represents a higher level of assured resources than is normal for most minerals. Further exploration and higher prices will certainly, on the basis of present geological knowledge, yield further resources as present ones are used up." The amount of uranium present in all currently known conventional reserves alone (excluding the huge quantities of currently-uneconomical uranium present in "unconventional" reserves such as phosphate/phosphorite deposits, seawater, and other sources) is enough to last over 200 years at current consumption rates. Fuel efficiency in conventional reactors has increased over time. Additionally, since 2000, 12–15% of world uranium requirements have been met by the dilution of highly enriched weapons-grade uranium from the decommissioning of nuclear weapons and related military stockpiles with depleted uranium, natural uranium, or partially-enriched uranium sources to produce low-enriched uranium for use in commercial power reactors. Similar efforts have been utilizing weapons-grade plutonium to produce mixed oxide (MOX) fuel, which is also produced from reprocessing used fuel. Other components of used fuel are currently less commonly utilized, but have a substantial capacity for reuse, especially so in next-generation fast neutron reactors. Over 35 European reactors are licensed to use MOX fuel, as well as Russian and American nuclear plants. Reprocessing of used fuel increases utilization by approximately 30%, while the widespread use of fast breeder reactors would allow for an increase of "50-fold or more" in utilization.
All nuclear plants produce radioactive waste. To pay for the cost of storing, transporting and disposing these wastes in a permanent location, in the United States a surcharge of a tenth of a cent per kilowatt-hour is added to electricity bills. Roughly one percent of electrical utility bills in provinces using nuclear power are diverted to fund nuclear waste disposal in Canada.
In 2009, the Obama administration announced that the Yucca Mountain nuclear waste repository would no longer be considered the answer for U.S. civilian nuclear waste. Currently, there is no plan for disposing of the waste and plants will be required to keep the waste on the plant premises indefinitely.
The disposal of low level waste reportedly costs around £2,000/m³ in the UK. High level waste costs somewhere between £67,000/m³ and £201,000/m³. General division is 80%/20% of low level/high level waste, and one reactor produces roughly 12 m³ of high level waste annually.
In Canada, the NWMO was created in 2002 to oversee long term disposal of nuclear waste, and in 2007 adopted the Adapted Phased Management procedure. Long term management is subject to change based on technology and public opinion, but currently largely follows the recommendations for a centralized repository as first extensively outlined by AECL in 1988. It was determined after extensive review that following these recommendations would safely isolate the waste from the biosphere. The location has not yet been determined, and the project is expected to cost between $9 and $13 billion CAD for construction and operation for 60–90 years, employing roughly a thousand people for the duration. Funding is available and has been collected since 1978 under the Canadian Nuclear Fuel Waste Management Program. Very long term monitoring requires less staff since high-level waste is less toxic than naturally occurring uranium ore deposits within a few centuries.
The primary argument for pursuing IFR-style technology today is that it provides the best solution to the existing nuclear waste problem because fast reactors can be fueled from the waste products of existing reactors as well as from the plutonium used in weapons, as is the case of the discontinued EBR-II in Arco, Idaho, and in the operating, as of 2014, BN-800 reactor. Depleted uranium (DU) waste can also be used as fuel in fast reactors. Waste produced by a fast-neutron reactor and a pyroelectric refiner would consist only of fission products, which are produced at a rate of about one tonne per GWe-year. This is 5% as much as present reactors produce, and needs special custody for only 300 years instead of 300,000. Only 9.2% of fission products (strontium and caesium) contribute 99% of radiotoxicity; at some additional cost, these could be separated, reducing the disposal problem by a further factor of ten.
At the end of a nuclear plant's lifetime, the plant must be decommissioned. This entails either dismantling, safe storage or entombment. In the United States, the Nuclear Regulatory Commission (NRC) requires plants to finish the process within 60 years of closing. Since it may cost $500 million or more to shut down and decommission a plant, the NRC requires plant owners to set aside money when the plant is still operating to pay for the future shutdown costs.
Decommissioning a reactor that has undergone a meltdown is inevitably more difficult and expensive. Three Mile Island was decommissioned 14 years after its incident for $837 million. The cost of the Fukushima disaster cleanup is not yet known, but has been estimated to cost around $100 billion. Chernobyl is not yet decommissioned, different estimates put the end date between 2013 and 2020.
A 2011 report for the Union of Concerned Scientists stated that "the costs of preventing nuclear proliferation and terrorism should be recognized as negative externalities of civilian nuclear power, thoroughly evaluated, and integrated into economic assessments—just as global warming emissions are increasingly identified as a cost in the economics of coal-fired electricity".
"Construction of the ELWR was completed in 2013 and is optimized for civilian electricity production, but it has "dual-use" potential and can be modified to produce material for nuclear weapons."
Nuclear safety and security is a chief goal of the nuclear industry. Great care is taken so that accidents are avoided, and if unpreventable, have limited consequences. Accidents could stem from system failures related to faulty construction or pressure vessel embrittlement due to prolonged radiation exposure. As with any aging technology, risks of failure increase over time, and since many currently operating nuclear reactors were built in the mid 20th century, care must be taken to ensure proper operation. Many more recent reactor designs have been proposed, most of which include passive safety systems. These design considerations serve to significantly mitigate or totally prevent major accidents from occurring, even in the event of a system failure. Still, reactors must be designed, built, and operated properly to minimize accident risks. The Fukushima disaster represents one instance where these systems were not comprehensive enough, where the tsunami following the Tōhoku earthquake disabled the backup generators that were stabilizing the reactor. According to UBS AG, the Fukushima I nuclear accidents have cast doubt on whether even an advanced economy like Japan can master nuclear safety. Catastrophic scenarios involving terrorist attacks are also conceivable.
An interdisciplinary team from MIT estimated that given the expected growth of nuclear power from 2005 to 2055, at least four core damage incidents would be expected in that period (assuming only current designs were used – the number of incidents expected in that same time period with the use of advanced designs is only one). To date, there have been five core damage incidents in the world since 1970 (one at Three Mile Island in 1979; one at Chernobyl in 1986; and three at Fukushima-Daiichi in 2011), corresponding to the beginning of the operation of generation II reactors.
According to the Paul Scherrer Institute, the Chernobyl incident is the only incident ever to have caused any fatalities. The report that UNSCEAR presented to the UN General Assembly in 2011 states that 29 plant workers and emergency responders died from effects of radiation exposure, two died from causes related to the incident but unrelated to radiation, and one died from coronary thrombosis. It attributed fifteen cases of fatal thyroid cancer to the incident. It said there is no evidence the incident caused an ongoing increase in incidence of solid tumors or blood cancers in Eastern Europe. With 46 deaths in its entire six-decade worldwide history, nuclear power remains the safest-ever way to make electricity, by a very wide margin.
In terms of nuclear accidents, the Union of Concerned Scientists have claimed that "reactor owners ... have never been economically responsible for the full costs and risks of their operations. Instead, the public faces the prospect of severe losses in the event of any number of potential adverse scenarios, while private investors reap the rewards if nuclear plants are economically successful. For all practical purposes, nuclear power's economic gains are privatized, while its risks are socialized".
However, the problem of insurance costs for worst-case scenarios is not unique to nuclear power: hydroelectric power plants are similarly not fully insured against a catastrophic event such as the Banqiao Dam disaster, where 11 million people lost their homes and from 30,000 to 200,000 people died, or large dam failures in general. Private insurers base dam insurance premiums on worst-case scenarios, so insurance for major disasters in this sector is likewise provided by the state. In the US, insurance coverage for nuclear reactors is provided by the combination of operator-purchased private insurance and the primarily operator-funded Price Anderson Act.
Any effort to construct a new nuclear facility around the world, whether an existing design or an experimental future design, must deal with NIMBY or NIABY objections. Because of the high profiles of the Three Mile Island accident and Chernobyl disaster, relatively few municipalities welcome a new nuclear reactor, processing plant, transportation route, or deep geological repository within their borders, and some have issued local ordinances prohibiting the locating of such facilities there.
The proven dangers of nuclear power amplify the economic risks of expanding reliance on it. Indeed, the stronger regulation and improved safety features for nuclear reactors called for in the wake of the Japanese disaster will almost certainly require costly provisions that may price it out of the market.
The cascade of problems at Fukushima, from one reactor to another, and from reactors to fuel storage pools, will affect the design, layout and ultimately the cost of future nuclear plants.
In 1986, Pete Planchon conducted a demonstration of the inherent safety of the Integral Fast Reactor. Safety interlocks were turned off. Coolant circulation was turned off. Core temperature rose from the usual 1000 degrees Fahrenheit to 1430 degrees within 20 seconds. The boiling temperature of the sodium coolant is 1621 degrees. Within seven minutes the reactor had shut itself down without action from the operators, without valves, pumps, computers, auxiliary power, or any moving parts. The temperature was below the operating temperature. The reactor was not damaged. The operators were not injured. There was no release of radioactive material. The reactor was restarted with coolant circulation but the steam generator disconnected. The same scenario recurred. Three weeks later, the operators at Chernobyl repeated the latter experiment, ironically in a rush to complete a safety test, using a very different reactor, with tragic consequences. Safety of the Integral Fast Reactor depends on the composition and geometry of the core, not efforts by operators or computer algorithms.
Insurance available to the operators of nuclear power plants varies by nation. The worst case nuclear accident costs are so large that it would be difficult for the private insurance industry to carry the size of the risk, and the premium cost of full insurance would make nuclear energy uneconomic.
Nuclear power has largely worked under an insurance framework that limits or structures accident liabilities in accordance with the Paris convention on nuclear third-party liability, the Brussels supplementary convention, the Vienna convention on civil liability for nuclear damage, and in the United States the Price-Anderson Act. It is often argued that this potential shortfall in liability represents an external cost not included in the cost of nuclear electricity.
However, the problem of insurance costs for worst-case scenarios is not unique to nuclear power: hydroelectric power plants are similarly not fully insured against a catastrophic event such as the Banqiao Dam disaster, where 11 million people lost their homes and from 30,000 to 200,000 people died, or large dam failures in general. Private insurers base dam insurance premiums on worst-case scenarios, so insurance for major disasters in this sector is likewise provided by the state.
In Canada, the Canadian Nuclear Liability Act requires nuclear power plant operators to obtain $650 million (CAD) of liability insurance coverage per installation (regardless of the number of individual reactors present) starting in 2017 (up from the prior $75 million requirement established in 1976), increasing to $750 million in 2018, to $850 million in 2019, and finally to $1 billion in 2020. Claims beyond the insured amount would be assessed by a government appointed but independent tribunal, and paid by the federal government.
In the United States, the Price-Anderson Act has governed the insurance of the nuclear power industry since 1957. Owners of nuclear power plants are required to pay a premium each year for the maximum obtainable amount of private insurance ($450 million) for each licensed reactor unit. This primary or "first tier" insurance is supplemented by a second tier. In the event a nuclear accident incurs damages in excess of $450 million, each licensee would be assessed a prorated share of the excess up to $121,255,000. With 104 reactors currently licensed to operate, this secondary tier of funds contains about $12.61 billion. This results in a maximum combined primary+secondary coverage amount of up to $13.06 billion for a hypothetical single-reactor incident. If 15 percent of these funds are expended, prioritization of the remaining amount would be left to a federal district court. If the second tier is depleted, Congress is committed to determine whether additional disaster relief is required. In July 2005, Congress extended the Price-Anderson Act to newer facilities.
The Vienna Convention on Civil Liability for Nuclear Damage and the Paris Convention on Third Party Liability in the Field of Nuclear Energy put in place two similar international frameworks for nuclear liability. The limits for the conventions vary. The Vienna convention was adapted in 2004 to increase the operator liability to €700 million per incident, but this modification is not yet ratified.
The cost per unit of electricity produced (kWh) will vary according to country, depending on costs in the area, the regulatory regime and consequent financial and other risks, and the availability and cost of finance. Costs will also depend on geographic factors such as availability of cooling water, earthquake likelihood, and availability of suitable power grid connections. So it is not possible to accurately estimate costs on a global basis.
Commodity prices rose in 2008, and so all types of plants became more expensive than previously calculated. In June 2008 Moody's estimated that the cost of installing new nuclear capacity in the United States might possibly exceed $7,000/KWe in final cost. In comparison, the reactor units already under construction in China have been reported with substantially lower costs due to significantly lower labour rates.
A 2008 study concluded that if carbon capture and storage were required, then nuclear power would be the cheapest source of electricity even at $4,038/kW in overnight capital cost.
In 2009, MIT updated its 2003 study, concluding that inflation and rising construction costs had increased the overnight cost of nuclear power plants to about $4,000/kWe, and thus increased the power cost to $0.084/kWh. The 2003 study had estimated the cost as $0.067/kWh.
According to Benjamin K. Sovacool, the marginal levelized cost for "a 1,000-MWe facility built in 2009 would be 41.2 to 80.3 cents/kWh, presuming one actually takes into account construction, operation and fuel, reprocessing, waste storage, and decommissioning".[unreliable fringe source]
A 2013 study indicates that the cost competitiveness of nuclear power is "questionable" and that public support will be required if new power stations are to be built within liberalized electricity markets.
In 2014, the US Energy Information Administration estimated the levelized cost of electricity from new nuclear power plants going online in 2019 to be $0.096/kWh before government subsidies, comparable to the cost of electricity from a new coal-fired power plant without carbon capture, but higher than the cost from natural gas-fired plants.
In 2019 the US EIA revised the levelized cost of electricity from new advanced nuclear power plants going online in 2023 to be $0.0775/kWh before government subsidies.
Generally, a nuclear power plant is significantly more expensive to build than an equivalent coal-fueled or gas-fueled plant. If natural gas is plentiful and cheap operating costs of conventional power plants is less. Most forms of electricity generation produce some form of negative externality — costs imposed on third parties that are not directly paid by the producer — such as pollution which negatively affects the health of those near and downwind of the power plant, and generation costs often do not reflect these external costs.
A comparison of the "real" cost of various energy sources is complicated by a number of uncertainties:
A UK Royal Academy of Engineering report in 2004 looked at electricity generation costs from new plants in the UK. In particular it aimed to develop "a robust approach to compare directly the costs of intermittent generation with more dependable sources of generation". This meant adding the cost of standby capacity for wind, as well as carbon values up to £30 (€45.44) per tonne CO2 for coal and gas. Wind power was calculated to be more than twice as expensive as nuclear power. Without a carbon tax, the cost of production through coal, nuclear and gas ranged £0.022–0.026/kWh and coal gasification was £0.032/kWh. When carbon tax was added (up to £0.025) coal came close to onshore wind (including back-up power) at £0.054/kWh — offshore wind is £0.072/kWh — nuclear power remained at £0.023/kWh either way, as it produces negligible amounts of CO2. (Nuclear figures included estimated decommissioning costs.)
A May 2008 study by the Congressional Budget Office concludes that a carbon tax of $45 per tonne of carbon dioxide would probably make nuclear power cost competitive against conventional fossil fuel for electricity generation.
Estimates of total lifetime energy returned on energy invested vary greatly depending on the study. An overview can be found here (Table 2):
The effect of subsidies is difficult to gauge, as some are indirect (such as research and development). A May 12, 2008 editorial in The Wall Street Journal stated, "For electricity generation, the EIA (Energy Information Administration, an office of the Department of Energy) concludes that solar energy is subsidized to the tune of $24.34 per megawatt hour, wind $23.37 and 'clean coal' $29.81. By contrast, normal coal receives 44 cents, natural gas a mere quarter, hydroelectric about 67 cents and nuclear power $1.59."
Lazard's report on the estimated levelized cost of energy by source (10th edition) estimated unsubsidized prices of $97–$136/MWh for nuclear, $50–$60/MWh for solar PV, $32–$62/MWh for onshore wind, and $82–$155/MWh for offshore wind. However, their report did not take into account any of the significant integration costs associated with variable renewable resources, assumed an unrealistically short facility lifetime of just 40 years for the nuclear plant, and used global illustrative costs of capital instead of the significantly lower OECD or North American costs of capital.
However, the most important subsidies to the nuclear industry do not involve cash payments. Rather, they shift construction costs and operating risks from investors to taxpayers and ratepayers, burdening them with an array of risks including cost overruns, defaults to accidents, and nuclear waste management. This approach has remained remarkably consistent throughout the nuclear industry's history, and distorts market choices that would otherwise favor less risky energy investments.
In 2011, Benjamin K. Sovacool said that: "When the full nuclear fuel cycle is considered — not only reactors but also uranium mines and mills, enrichment facilities, spent fuel repositories, and decommissioning sites — nuclear power proves to be one of the costliest sources of energy".
In 2014, Brookings Institution published The Net Benefits of Low and No-Carbon Electricity Technologies which states, after performing an energy and emissions cost analysis, that "The net benefits of new nuclear, hydro, and natural gas combined cycle plants far outweigh the net benefits of new wind or solar plants", with the most cost effective low carbon power technology being determined to be nuclear power. Moreover, Paul Joskow of MIT maintains that the "Levelized cost of electricity" (LCOE) metric is a poor means of comparing electricity sources as it hides the extra costs, such as the need to frequently operate back up power stations, incurred due to the use of intermittent power sources such as wind energy, while the value of baseload power sources are underpresented.
An EU-funded research study known as ExternE, or Externalities of Energy, undertaken from 1995 to 2005, found that the cost of producing electricity from coal or oil would double, and the cost of electricity production from gas would increase by 30% if external costs such as damage to the environment and to human health, from the particulate matter, nitrogen oxides, chromium VI, river water alkalinity, mercury poisoning and arsenic emissions produced by these sources, were taken into account. It was estimated in the study that these external, downstream, fossil fuel costs amount up to 1–2% of the EU's Gross Domestic Product, and this was before the external cost of global warming from these sources was included. The study also found that the environmental and health costs of nuclear power, per unit of energy delivered, was lower than many renewable sources, including that caused by biomass and photovoltaic solar panels, but was higher than the external costs associated with wind power and alpine hydropower.
Kristin Shrader-Frechette analysed 30 papers on the economics of nuclear power for possible conflicts of interest. She found of the 30, 18 had been funded either by the nuclear industry or pro-nuclear governments and were pro-nuclear, 11 were funded by universities or non-profit non-government organisations and were anti-nuclear, the remaining 1 had unknown sponsors and took the pro-nuclear stance. The pro-nuclear studies were accused of using cost-trimming methods such as ignoring government subsidies and using industry projections above empirical evidence where ever possible. The situation was compared to medical research where 98% of industry sponsored studies return positive results.
Nuclear Power plants tend to be very competitive in areas where other fuel resources are not readily available — France, most notably, has almost no native supplies of fossil fuels. France's nuclear power experience has also been one of paradoxically increasing rather than decreasing costs over time.
A Council on Foreign Relations report on nuclear energy argues that a rapid expansion of nuclear power may create shortages in building materials such as reactor-quality concrete and steel, skilled workers and engineers, and safety controls by skilled inspectors. This would drive up current prices. It may be easier to rapidly expand, for example, the number of coal power plants, without this having a large effect on current prices.
Existing nuclear plants generally have a somewhat limited ability to significantly vary their output in order to match changing demand (a practice called load following). However, many BWRs, some PWRs (mainly in France), and certain CANDU reactors (primarily those at Bruce Nuclear Generating Station) have various levels of load-following capabilities (sometimes substantial), which allow them to fill more than just baseline generation needs. Several newer reactor designs also offer some form of enhanced load-following capability. For example, the Areva EPR can slew its electrical output power between 990 and 1,650 MW at 82.5 MW per minute.
The number of companies that manufacture certain parts for nuclear reactors is limited, particularly the large forgings used for reactor vessels and steam systems. Only four companies (Japan Steel Works, China First Heavy Industries, Russia's OMZ Izhora and Korea's Doosan Heavy Industries) currently manufacture pressure vessels for reactors of 1100 MWe or larger. Some have suggested that this poses a bottleneck that could hamper expansion of nuclear power internationally, however, some Western reactor designs require no steel pressure vessel such as CANDU derived reactors which rely on individual pressurized fuel channels. The large forgings for steam generators — although still very heavy — can be produced by a far larger number of suppliers.
For a country with both a nuclear power industry and a nuclear arms industry synergies between the two can favor a nuclear power plant with an otherwise uncertain economy. For example, in the United Kingdom researchers have informed MPs that the government was using the Hinkley Point C project to cross-subsidise the UK military's nuclear-related activity by maintaining nuclear skills. In support of that, the researchers from the University of Sussex, Prof. Andy Stirling and Dr. Phil Johnstone, stated that the costs of the Trident nuclear submarine programme would be prohibitive without “an effective subsidy from electricity consumers to military nuclear infrastructure”.
The nuclear power industry in Western nations has a history of construction delays, cost overruns, plant cancellations, and nuclear safety issues despite significant government subsidies and support. In December 2013, Forbes magazine reported that, in developed countries, "reactors are not a viable source of new power". Even in developed nations where they make economic sense, they are not feasible because nuclear's “enormous costs, political and popular opposition, and regulatory uncertainty”. This view echoes the statement of former Exelon CEO John Rowe, who said in 2012 that new nuclear plants “don’t make any sense right now” and won't be economically viable in the foreseeable future. John Quiggin, economics professor, also says the main problem with the nuclear option is that it is not economically-viable. Quiggin says that we need more efficient energy use and more renewable energy commercialization. Former NRC member Peter A. Bradford and Professor Ian Lowe have recently made similar statements. However, some "nuclear cheerleaders" and lobbyists in the West continue to champion reactors, often with proposed new but largely untested designs, as a source of new power.
Significant new build activity is occurring in developing countries like South Korea, India and China. China has 25 reactors under construction, However, according to a government research unit, China must not build "too many nuclear power reactors too quickly", in order to avoid a shortfall of fuel, equipment and qualified plant workers.
The 1.6 GWe EPR reactor is being built in Olkiluoto Nuclear Power Plant, Finland. A joint effort of French AREVA and German Siemens AG, it will be the largest pressurized water reactor (PWR) in the world. The Olkiluoto project has been claimed to have benefited from various forms of government support and subsidies, including liability limitations, preferential financing rates, and export credit agency subsidies, but the European Commission's investigation didn't find anything illegal in the proceedings. However, as of August 2009, the project is "more than three years behind schedule and at least 55% over budget, reaching a total cost estimate of €5 billion ($7 billion) or close to €3,100 ($4,400) per kilowatt". Finnish electricity consumers interest group ElFi OY evaluated in 2007 the effect of Olkiluoto-3 to be slightly over 6%, or €3/MWh, to the average market price of electricity within Nord Pool Spot. The delay is therefore costing the Nordic countries over 1.3 billion euros per year as the reactor would replace more expensive methods of production and lower the price of electricity.
Russia has launched the world's first floating nuclear power plant. The £100 million vessel, the Akademik Lomonosov, is the first of seven plants (70 MWe per ship) that Moscow says will bring vital energy resources to remote Russian regions. Startup of the first of the ships two reactors was announced in December 2018.
Following the Fukushima nuclear disaster in 2011, costs are likely to go up for currently operating and new nuclear power plants, due to increased requirements for on-site spent fuel management and elevated design basis threats. After Fukushima, the International Energy Agency halved its estimate of additional nuclear generating capacity built by 2035.
Many license applications filed with the U.S. Nuclear Regulatory Commission for proposed new reactors have been suspended or cancelled. As of October 2011, plans for about 30 new reactors in the United States have been reduced to 14. There are currently five new nuclear plants under construction in the United States (Watts Bar 2, Summer 2, Summer 3, Vogtle 3, Vogtle 4). Matthew Wald from The New York Times has reported that "the nuclear renaissance is looking small and slow".
In 2013, four aging, uncompetitive reactors were permanently closed in the US: San Onofre 2 and 3 in California, Crystal River 3 in Florida, and Kewaunee in Wisconsin. The Vermont Yankee plant closed in 2014. New York State is seeking to close Indian Point Nuclear Power Plant, in Buchanan, 30 miles from New York City. The additional cancellation of five large reactor uprates (Prairie Island, 1 reactor, LaSalle, 2 reactors, and Limerick, 2 reactors), four by the largest nuclear company in the United States, suggest that the nuclear industry faces "a broad range of operational and economic problems".
As of July 2013, economist Mark Cooper has identified some US nuclear power plants that face particularly significant challenges to their continued operation due to regulatory policies. These are Palisades, Fort Calhoun, Nine Mile Point, Fitzpatrick, Ginna, Oyster Creek, Vermont Yankee, Millstone, Clinton, Indian Point. Cooper said the lesson here for policy makers and economists is clear: "nuclear reactors are simply not competitive". In 2017 analysis by Bloomberg showed that over half of U.S. nuclear plants were running at a loss.
the plant had become unprofitable in recent years, a victim largely of lower energy prices resulting from a glut of natural gas used to fire electricity plants
Schucht says, in the region he is operating in, 42 per cent of the power supply (in output, not capacity), came from wind and solar – about the same as South Australia. Schucht believes that integration of 60 to 70 per cent variable renewable energy – just wind and solar – could be accommodated within the German market without the need for additional storage. Beyond that, storage will be needed.