Skip navigation
Share via:


Simple negotiation of a global carbon price, enforced by import tariffs. Adding cheap molten salt reactors enables rapid decarbonization.



To prevent the construction of large numbers of coal plants worldwide, we need dispatchable energy which is cheaper than coal.

To rapidly decarbonize from current levels, we need early retirement of existing fossil plants. It's not enough to simply build clean energy when the old plants finally expire. To motivate early retirement, we need the total levelized cost of the clean energy plant to be significantly lower than the fuel cost of the fossil plant. We also need the construction time to be short, and the capital cost attainable. If we do this, then a plant owner can profit by halting fuel purchases and replacing them with loan payments on the new plant.

There are two ways to do this: invent new clean energy technology that's inherently cheap, or raise the cost of fuel with a carbon price. This proposal advocates both.

Global Carbon Price

For a carbon price, we advocate the recent "Climate Club" proposal by Nordhaus. Nations can individually enact revenue-neutral carbon fees, and disincentivize free-riding by means of import duties. Modeling done by Nordhaus shows that a regime like this leads to broad participation up to a fee of $50/ton CO2, leading to an emissions drop of up to 50% with current technology.

We can go much further if we develop cheaper non-carbon technology.


Solar and wind costs are dropping, and not that far from being competitive even without subsidies. Many people are especially optimistic about the progress of solar PV. 

These low costs depend on backup by fossil plants, especially natural gas. Nevertheless, with wind and solar currently providing only a small percentage of our power, we have room to expand them significantly.

Running on wind/solar alone would be much more difficult. A 2013 study using historical data from a portion of the U.S. grid found that the cheapest way to run on 99% wind/solar was to overproduce energy by a factor of three, and add substantial amounts of storage, more than tripling costs.

Molten Salt Reactors

Instead of trying to overextend solar, we replace natural gas backup with nuclear. This wouldn't be ideal as a backup to wind; nuclear is dominated by capital costs rather than fuel, so it's most economical to keep it running all the time. But it would work well with solar, since it produces energy during the day when demand is higher. The nuclear plants could be run as steady baseload, with solar adding supply when there's more demand. 

Fission has obvious concerns over safety, proliferation, and waste, but these problems are solvable in the near term.

In the past few years, there's been a lot of attention to liquid thorium breeder reactors. This approach still has serious technological challenges. However, molten salt designs without breeding, running at least partially on uranium, are much easier. These designs are extremely safe, and some are very proliferation-resistant and appropriate for developing countries.

The first commercial reactors should be available within a decade, and have the potential massive scalability and costs cheaper than coal.

The Timing Works

Solar's portion of electricity supply is very low at the moment. Over the next decade, its expansion won't won't be hindered by reliability concerns, since there's plenty of fossil-based power available for backup.

By the time solar is supplying most of our daytime peak power needs, the molten salt reactors will be ready for mass production.

If we tried to run on nuclear alone, it would run at lower capacity factors, since we'd need excess capacity for peak demand. Having solar handling the peaks means our reactors can run at high capacity factors, lowering their cost per kWh.

By combining a carbon price with cheap energy sources, we can make the fuel for fossil plants more expensive than the total cost per kWh of carbon-free energy. By making it cheaper to convert to fossil-free energy than to keep fueling fossil plants, we can motivate early plant shutdowns and achieve rapid carbon reduction.

Which plan do you select for China?

Value not set.

Which plan do you select for India?

Value not set.

Which plan do you select for the United States?

U.S. Molten Salt Reactors

Which plan do you select for Europe?

Moltex, a molten salt reactor for the U.K.

Which plan do you select for other developing countries?

Value not set.

Which plan do you select for other developed countries?

Value not set.

What additional cross-regional proposals are included in your plan, if any?

The carbon price mechanism is described in detail in

Using a carbon price instead of caps changes the incentives. "With individual emission caps, each country wants to have a high (weak) cap and wants all others have a low (strong) cap. This results in all countries advocating ambition for others, while doing little themselves....only a global commitment to a single price can crack the free-rider problem."

There are other benefits:

A single carbon price is simpler to negotiate than separate caps for every country.

A price helps with fairness concerns; it's unfair to put a low cap on a developing country, but a high cap is meaningless.

A cap that's too high, in an international trading system, can be a source of profit, letting a country sell its unused emission permits to other countries. This potential profit makes agreement even harder to reach.

The references section links more extensive arguments, by leading climate economists and Nobel winner Joseph Stiglitz.

In the U.S., advanced nuclear development and expansion can't happen without regulatory reform. A proposal addressing that is  

In addition to electricity, molten salt reactors can provide industrial process heat, since they operate at high temperatures. Decarbonizing transportation will take a little more work. Electrifying ground transportation is one approach; an especially efficient option is

To the extent we keep using liquid fuels, we can synthesize them fairly efficiently from CO2 in the oceans, using nuclear or solar energy and the chemical process in

How do the regional and cross-sectoral plans above fit together?

An achievable carbon price can prevent new fossil plant builds, but is unlikely to motivate early shutdown of existing plants. Advanced carbon-free energy sources can be competitive with new fossil plants, but are unlikely to be radically cheaper in the absence of a carbon price. By combining cheap sources with a carbon price, we have the potential for a step change in results: carbon-free energy with a total cost per kWh which is less than the variable cost of fossil plants, making early shutdown the most economical option.

Solar is immediately available and economical at low market penetration, but requires more-expensive overproduction and storage at high market penetration. We start with solar.

Inexpensive advanced nuclear won't be available for a decade, but once available, it can fill the baseload role where solar struggles.

Having solar providing additional power during the day means that nuclear plants can run at closer to full capacity. Since nuclear is dominated by fixed costs, it's more economical if we can keep it running, instead of reducing output at night. The combination of both technologies won't perfectly match demand, but it'll be better than either technology alone.

Research and Development

The U.S. has the potential to be a technology leader. Its strengths include expertise, venture capital, and testing facilities. U.S.-based molten salt reactor companies include Flibe, Transatomic, and ThorCon, and there are many other startup companies pursuing alternate designs. However, the U.S. is also hamstrung by inflexible regulations designed for large conventional reactors, using a process which is a strong disincentive to venture capital. The U.S. proposal advocates specific reforms. Until reforms can be achieved, U.S. companies may need to build prototypes in other countries with more appropriate regulations.

Canada has regulations much more suitable for innovative reactor development. Terrestrial Energy is a molten salt company in Canada which has very positive things to say about Canada's regulators. The company has a simple, proliferation-resistant design and is targeting production status by 2024.

France and the U.K. have a substantial amount of nuclear waste and plutonium it would like to dispose of. The U.K. has Moltex, an innovative molten salt reactor startup with a design that could efficiently burn that waste.

China has an aggressive program developing advanced reactor types, including molten salt reactors with both liquid and solid fuels, fast reactors, and supercritical water-cooled reactors, and high-temperature gas-cooled reactors. The U.S. reactor startup Terrapower recently signed a deal to build a prototype in China, and the DOE is providing assistance on molten salt reactors.


These reactors are designed for mass production in factories or shipyards, rather than painstaking on-site construction.

Terrestrial Energy uses reactors that are strongly proliferation-resistant, sealed units. These could be comfortably deployed in developing nations. Their smallest unit could be shipped by truck, providing a continuous 80MWe without refueling. After seven years, the unit is simply replaced.

Transatomic and Moltex use larger reactors, with simple designs, factory-produced modular parts, and simple assembly.

ThorCon's reactor is designed for shipyard production at massive scale. A single large shipyard could produce 100 GWe capacity per year. It's possible that similar construction principles could be developed for other reactor core designs, perhaps in collaboration with ThorCon.

With such rapid capacity expansion, we could replace most of the world's fossil-based electricity supply in two or three decades. Initial deployment would begin around 2025, with large-scale rollout beginning about five years later.

Global Fuel Cycle Eliminates Waste

The four reactors described here would play different roles in the global nuclear ecosystem. Even though some reactors produce long-term waste, others can use that as fuel, so we eliminate overall waste production. Reactors without strong proliferation resistance can be restricted to weapons-capable states.

Terrestrial Energy's IMSR uses low-enriched uranium. It's six times more fuel-efficient than conventional reactors, producing a sixth as much long-term waste, or none with added reprocessing. It's very proliferation-resistant.

Transatomic's design uses uranium with very low enrichment, at only about 2% U-235, making it very proliferation-resistant. Due to its unique core design it's also effective at burning up nuclear waste.

ThorCon's design would require 20% enrichment, so would be more appropriate for nuclear-capable countries. It doesn't eliminate long-term waste. However, it's a conservative approach designed for very rapid and scalable deployment, with a very short development timeframe and very low cost.

Moltex is a fast reactor, using a chloride salt instead of the fluoride salts used for thermal reactors. Being a fast reactor makes Moltex very good at eliminating long-term waste. On the other hand, fast reactors require higher fissile loads. Moltex would be best suited for use in nuclear weapons states, and for burning up long-term waste from other states.

These designs have overlapping capabilities, so our plan doesn't depend on all of them succeeding. It's possible that technology sharing could open up other possibilities, such as using ThorCon's shipyard construction techniques for other reactor core designs.

Carbon Price

As a carbon price mechanism, we suggest the Carbon Club approach recently described by Nordhaus: a single global price per ton of carbon, implemented by participating countries, with import duties against nonparticipating countries to incentivize participation. Since the duties are just an incentive, there's no need to estimate the carbon intensity of goods.

We also propose a simple mechanism to negotiate the global carbon price, similar to approval voting. Details are in the Climate Club proposal.

Explanation of the emissions scenario calculated in the Impact tab

Modeling by Nordhaus indicates that Climate Club would support broad participation for carbon fees up to $50/ton CO2. Agreement on this is unlikely in 2015 so we start with 2018. For reasons described elsewhere in this proposal, and extensively in the referenced works, a uniform price agreement would be easier to achieve than effective caps.

Molten salt reactors can be available by 2025, but won't be available at scale with economical cost right away. We used 2030 as the "energy breakthrough" date.

Estimated breakthrough is a 33% cost reduction. See the "references" sections for detailed justification. Since molten salt reactors lack some limitations of conventional nuclear, this was applied both to "nuclear" and "new technology." (Compared to conventional nuclear, MSRs are much better at load-following, and can be deployed much more quickly.)

A more modest cost reduction of 20% was applied to renewables, with a date of 2025, based on the likely learning curve from deployment and existing R&D.

The "accelerated retirement of coal" was set to a modest 3%, assuming that will be driven by pollution concerns, carbon pricing, and later availability of cheap mass-produced MSRs.

The efficiency increase for transportation is based on SkyTran, electric cars, and eventually on lower-carbon synthetic fuels, all mentioned in the "additional proposals" section.

There wasn't room in this proposal for other additional strategies such as land use changes, so those were left out of the EnROADS model. We certainly should do those things; the model simply estimates results of the efforts described in this proposal.

What are the plan’s key benefits?

A single global carbon price is simpler to negotiate than a separate cap for each country, especially with the simple approval-voting mechanism proposed here. Having a trade-based enforcement mechanism makes compliance likely.

A reliable power source which is abundant and affordable gives us a way to rapidly decarbonize with minimal hardship. Combined with an achievable carbon price, the incentive to do so will be strong.

Abundant nuclear energy can not only replace existing fossil plants, but add the new electrical output that we'll need as we electrify our transportation systems.

While the initial focus of molten salt reactors would be in replacing fossil-based power plants, their high operating temperature makes them good for industrial process heat as well.

To some extent, we will probably continue to use liquid fuels. Past 2050, molten salt reactors could be applied here as well, providing the energy to make liquid fuels from CO2 in the oceans or air. 

What are the plan’s costs?

As a rough estimate, $1 billion for development of each new reactor type. This is research cost, not reactor cost. Government funding would be nice, but we're already seeing private investment in advanced reactor development, which could expand if we eliminate the major uncertainties caused by unpredictable regulators. (Currently, investors get little feedback from the NRC until they've already spent several hundred million dollars for detailed design work. After all that, they can be left with little but a denied application.)

The first production reactors will likely be expensive, but costs should drop rapidly as mass production kicks in. Let's say $20 billion in excess costs during initial rollout. Later, the favorable economics of these reactors actually save us money, even compared to fossil.

Capital costs for conventional nuclear are high. They require extensive engineered safety mechanisms to overcome the inherent risks of solid-fueled water-cooled reactors. They are custom-built on-site, with construction techniques that are difficult to automate. They take a long time to build, increasing finance costs. Molten salt reactors fix all this, with good inherent safety and mass production in factories or shipyards.

Capital cost can be cheaper than coal, and fuel cost much lower than all fossil fuels. The company sites listed in the references have more detailed cost information, some with independent evaluations.

There will also be costs associated with the carbon price, as people pay to improve efficiencies and convert to lower-carbon but more expensive energy sources, prior to cheap reactors being widely available by 2030 or so. These would be similar to any plan with a similar carbon price.


What are the key challenges to enacting this plan?

Climate negotiations are difficult, no matter how simple we try to make them.

The molten salt reactors described here are designed for near-term development, with minimal requirement for technological advances, but they'll still require substantial and well-funded engineering efforts.

Heavy regulation of nuclear technology makes progress relatively slow. The U.S. in particular needs regulatory reform to make advanced reactor development feasible within its borders.

Public acceptance of nuclear fission is chancy, but a year after Fukushima, Gallop found that a majority of Americans still favored nuclear power. By our target date of 2030, the U.S. public will have experienced significant climate impacts, and be more anxious for low-carbon solutions. It should also be more familiar with the safety and other advantages of molten salt reactors.

Some other countries are more supportive already. Russia for example is quietly expanding its reactor exports as a way of extending geopolitical influence.

But the real leader is China, which is currently building 24 nuclear reactors. China plans to build seven reactors per year between now and 2030, and 400 new reactors by 2050. It's actively developing advanced reactors including molten salt reactors.

China has also shown itself to be open to collaboration with other countries. Terrapower recently signed an agreement to build its prototype fast reactor in China, and other reactor companies could follow their lead if U.S. regulators prove too difficult for now.


Carbon Price

Ideally, an adequate global carbon price with effective enforcement would be achieved during this year's climate negotiations, but this seems unlikely. We need a shift in strategy from the current caps approach, and linking climate to trade mechanisms may require additional work. This would be compensated by the simplicity of a global carbon price, but it's enough of a paradigm shift that it may not happen before the next round of climate talks. 

However, a uniform carbon price is strongly favored by many well-known economists. Surely we don't have to simply accept the status quo. Caps have been tried, and haven't worked well. We can and must do better. Perhaps we can get it done within the next three years, if not this time.


With a carbon price achieved, emissions should start to drop. Initially, efficiency improvements and renewables should be sufficient to make serious impacts. It's only when renewables start to supply a large portion of electricity that they may start to struggle.

So over the next 10 to 15 years, we focus on efficiency, some wind, and especially a major rollout of cheap solar PV.


Assuming good funding and rational regulation, commercial molten salt reactors should start to be available in a decade. Deployment over the next five years will probably be a bit cautious. However by 2030, about the time that renewables begin to struggle, MSRs should be ready for rapid deployment at massive scale.

Summing Up

2015-2030: Solar expansion to handle the extra daytime energy demand

2018 (approx): Global uniform carbon price

2025: Introduction of commercial molten salt reactors

2030: Begin major deployment of mass-produced molten salt reactors for baseload power


Carbon Price

Nordhaus, William. 2015. "Climate Clubs: Overcoming Free-Riding in International Climate Policy." American Economic Review, 105(4): 1339-70.

Further arguments for a carbon price rather than caps are at In particular, see the downloadable book Global Carbon Pricing (pdf), with contributions from Nordhaus, Joseph Stiglitz, and others.

Can Negotiating a Uniform Carbon Price Help to Internalize the Global Warming Externality? (pdf) by Harvard professor Martin L. Weitzman

Price Carbon: I Will If You Will, published in Nature, argues that "A global carbon price — so far excluded from consideration in international negotiations — would be the ideal basis for a common commitment in our view. A price is easy to agree and handle, relatively fair, less vulnerable to gaming than global cap-and-trade systems, and consistent with climate policies already in place." A BBC article has comments from one of its authors.

Carbonomics by climate economist Steven Stoft advocates a carbon price rather than cap, and James Hansen does the same in Storms of My Grandchildren.


A 2013 study on running the U.S. grid on 99% wind/solar found the cheapest option was to overproduce by a factor of three and add 9-72 hours of storage. Even with no storage, that triples our wind/solar cost.

For a sense of the scale of that much storage, see A Nation-Size Battery and Pump Up the Storage, by Berkeley physics professor Tom Murphy. These articles assume a need for seven days of storage; according to the above study we can reduce that by an order of magnitude, but that still leaves us at a level which doesn't look physically possible with current technologies.


Moltex Energy

Terrestrial Energy

ThorCon Power

Transatomic Power

Recent review of molten salt reactor companies from a U.K. perspective (pdf)

An introduction to China's massive nuclear expansion, with links to other articles, is at: China Shows How to Build Nuclear Reactors Fast and Cheap

Reactor Costs

Moltex has an independent cost estimate from nuclear engineering firm Atkins Ltd., giving a most-likely cost of £1414 per kW, with a possible range of  £909 to £2514 per kW. This compares favorably with the $3000/kW of a new coal plant, and the $6000/kW of an AP-1000 nuclear plant. Since nuclear is dominated by capital cost, while coal has significant fuel cost, this gives a realistic prospect for nuclear energy cheaper than coal, even without a carbon price.

Transatomic also has an independent estimate, from nuclear engineering firm Burns & Roe. Their white paper (pdf) summarizes this on page 36, giving $2 billion for a 520MWe output. (This puts it competitive with coal, for which a quarter to a third of the cost is fuel.)

ThorCon does not have an independent estimate but offers its own projections, based on material and construction costs. "ThorCon uses the same steam and electrical side as a standard 500 MWe supercritical coal plant. But gone are the massive coal handling systems, the 100 m high boiler, the flue gas treatment system, and the ash handling and storage system....The ThorCon nuclear island requires one-sixth as much steel and one-fourth as much concrete as the portion of the coal plant upstream from the turbine....ThorCon operating at near ambient pressure has a 2:1 advantage in steel and a 5:1 advantage in concrete over its nuclear competitors. Much more importantly, very little of ThorCon’s concrete is reinforced."

Altogether, ThorCon estimates a cost as low as 3.5 cents/kWh, but only with an assumption of very favorable regulation. Moltex and Transatomic's estimates are based on more conventional regulatory practices.

Molten salt reactors are inherently better at load-following than solid-fueled nuclear reactors, making a wind/solar/nuclear grid viable without fossil or storage backup. See the Transatomic whitepaper, p.44.