What Might a $ 10B Fusion R&D Initiative Look Like? -

What Might a $ 10B Fusion R&D Initiative Look Like?

Dr. Neil Mitchell and Glenn Weinreb



// php echo do_shortcode (‘[responsivevoice_button voice=”US English Male” buttontext=”Listen to Post”]’)?>

This article explores how a government or foundation could potentially resolve climate change with a $ 10B sized R&D program that accelerates the development of fusion power. We refer to it as “Commercial Fusion Now” (CFN), and the goal would be to produce commercial fusion this decade. We define “commercial fusion” as creating electricity at a cost comparable to electricity made with natural gas or coal. This would require a fusion reactor to produce more electricity than it consumes upon demand, reliably, and at low cost.

Currently, commercial fusion is not expected until the 2030’s or 2040’s. However, with more funding, it could perhaps appear sooner. Nations under great pressure sometimes spend money to bring a technology to market quickly. For example, the United States commercialized ‘space’ in the 2000s, and the Covid vaccine development in 2020. Global decarbonization will cost upwards of trillions of dollars. Therefore, it is reasonable to spend billions of dollars to accelerate commercial fusion.

Our first author is Dr. Mitchell, who has over 40 years’ of fusion experience. He was responsible for magnet development and manufacturing at the $ 25B ITER initiative in France, and is currently an advisor to the ITER Director General.

There are two types of nuclear power: fission and fusion. Traditional nuclear power plants generate electricity with uranium via fission. However, this is not popular due to meltdown risk, nuclear waste, proliferation risk (ie bomb), and cost. Fusion, on the other hand, does not have these issues; however, it is still in development. Fusion uses a small amount of radioactive tritium as a fuel. With more R&D, however, this could potentially be replaced with non – radioactive deuterium – deuterium, which is derived from sea water.

Figure 1: Fusion reactor ITER is expected to be operational in 2025. Notice person at bottom – center (Source: Wikipedia).

For an audio podcast interview with Dr. Mitchell, see “What would it take to get Commercial Fusion Running this Decade?”

Begin With a Grand Bargain

Approximately ten public and private fusion research organizations are currently in possession of talent, laboratories, and patents. They have resources and know – how, but in isolation they cannot produce commercial fusion this decade. Alternatively, they could pool their resources and advance more quickly. However, this would require a grand bargain that facilitates collaboration: CFN is that bargain. More specifically, each organization would be offered participation in the CFN program where they would receive:

  • Access to technology controlled by other participants (eg patents and design files).
  • Participation on panels that direct CFN’s $ 10B development budget.
  • A portion of revenue gained from the CFN program.
  • A portion of the $ 10B fund to do design, analysis, and experiments.

Each participant would incur two costs:

  • Other participants would gain access to their existing patents, at least until commercial fusion becomes a reality.
  • They would share revenue with others; as opposed to receiving more for themselves in the event they developed commercial fusion alone.

It is likely the ten organizations would agree in principle to the above terms; however, actual agreement would depend on details such as revenue sharing percentages. Subsequently, a negotiation would ensue between funding source (s) and participants. It is not clear how many would agree to the same set of details. However, $ 10B is a lot of motivation, and it is in their best financial interests to pool resources in order to get commercial fusion working before their patents expire.

Each participant has unique resources. For example, Commonwealth Fusion Systems and Tokamak Energy both have advanced magnets. And existing superconducting tokamaks in China, Korea, and Japan are capable of driving plasmas for long periods of time, which is vital when testing plasma performance, fueling, control, and heat removal.

Build Three Reactors Simultaneously

The initiative would build three reactors: CFN1, CFN2, and CFN3.

CFN1 would produce significant amounts of heat for short durations. It would be small, built quickly, and it would not produce electricity. CFN2 would generate electricity for long durations. However, CFN2 would not be commercially viable due to first – of – a – kind engineering costs. Finally, CFN3 would be designed to be manufactured and operated at a low cost for many decades. Industrial companies would improve upon CFN3, and produce their own versions, referred to as CFN4x.

CFN1, CFN2, and CNF3 would be completed by Dec 31, 2029.

Produce Lots of Heat with Reactor CFN1

CFN1 would maintain plasma for several seconds and produce a significant amount of heat (not electricity). CFN1 would demonstrate an energy gain (Qplasma = Heatoutput / Heatinput) sufficient for commercial fusion, which is on the order of Qplasma > 10.

CFN1 would validate reactor components and concepts, and be similar to SPARC, which is currently being built by CFS in the US Also, if collaboration is permitted, CFN1 could be SPARC. CFS hopes to complete SPARC by 2025, and CFS has raised $ 1.8B to work on both SPARC and its cousin ARC.

Continuously Make Electricity with Reactor CFN2

The CFN2 reactor would produce electricity, breed tritium fuel, operate continuously, and support the replacement of components damaged by neutrons. It would be made by hand, however, and be expensive to construct.

CFN2 might be based on ARC, which is currently being designed by CFS in the US CFS is not expected to complete ARC until the 2030’s. It could appear sooner if fusion institutes throughout the world pooled their expertise, technology, and facilities. This includes ITER, China, Korea, and Japan.

In the accelerated route to commercial fusion, CFN2 would be constructed after CFN1. Both will be designed, however, in parallel to accelerate development.

Dozens of CFN2 engineering challenges would be tackled immediately. For example, disassembly of CFN2 could potentially involve plug – in connectors with superconducting components. However, it is not clear these would survive over many decades, and engineers would need to better understand this before making fundamental design decisions. Additional funding and a grand bargain of patents would allow engineers to explore many different options, and ultimately select the best.

Low-cost Electricity with Reactor CFN3

CFN3 would be similar to CFN2 in performance. The major difference is it is intentionally designed to be manufactured / operated at low cost.

CFN3 engineers would start with a diagram of the electrical grid for the US, Europe, and China, and then consider how they would mass produce fusion reactors at nodes where large power transmission wires are located. Each node might need 1 to 10 GWe of power. CFN3 engineers would build one reactor and also explore how thousands of reactors might unfold; whereas CFN2 engineers would focus on building one reactor.

CFN3 would be similar to CFN2. Therefore, one could not construct CFN3 until CNF2 was fairly well understood. However, many things could be worked on early in parallel with CFN2 development. This includes site design, containment chamber design, exploring molded processes that reduce cost (as opposed to machining at high cost), exploring fabrication of large non – transportable components made on – site, development of redundant systems that reduce mean – time – between– failures, and exploring robotics that reduces manufacturing / operating costs.

CFN2 would be both over – instrumented and over – designed since engineers would not know the minimum requirements. In other words, CFN2 would contain more sensors than needed, be stronger than needed, and include more research equipment. CFN3 would reduce to save money.

The CFN steering committee might initially set the CFN3 budget to be relatively small since CFN3 is based on CFN2. However, it is possible this would not occur due to a long list of engineering challenges identified by industry executives as having long development times.

We would like the industry to mass – produce reactors in the 2030’s; as opposed to executing engineering in the 2030’s and beginning mass production in the 2040’s. CFN3 would help the industry get started quickly.

Empower Top People

There are two ways to manage a large development project. The traditional method is to develop a plan, get it funded, and implement it. Alternatively, one can set goals, assemble a top team, give them authority, provide funding, and get out of their way.

The traditional method works well if one knows what needs to be done. This is not the case, however, with commercial fusion. Subsequently, CFN would need the world’s top fusion scientists and engineers to be given authority and money, with little plan and open – eyed risk taking. Also, the funding source would reserve the right to replace steering committee members, and / or reduce funds if not satisfied.

If one needs to develop multiple items where each is based on the preceding item, one might be inclined to develop one after the other, in series.

Alternatively, one could develop in parallel with more waste and at greater cost. CFN would need to operate largely in parallel to produce results quickly.

The CFN steering committee would establish a list of steps required to produce a commercial reactor and set intermediate goals. They would also stimulate and fund proposals for basic research, component design and manufacturing, engineering integration, building of prototypes, and testing.

Transfer Technology to Industry

To expedite transferring technology to industry, CFN would provide transparency. In other words, designs, experiments, prototypes, presentations, and models would be made available as they are completed. This would better enable outside engineers to offer suggestions for improvement and to suggest engineering needed to be initiated before CFN3 is completed due to long development times.

The industry would be less likely to participate in CFN4x development if they considered fusion to be too technically and / or economically risky. Therefore, CFN would like to identify areas considered risky by industry executives and take steps to mitigate them. For example, if industry executives expressed concern over magnet longevity, CFN might do more magnet testing under duress, and for longer durations.

Nuclear fusion offers far less nuclear waste than fission. It does, however, involve radiation and radiation protection. CFN would need to engage with regulatory authorities at an early stage to establish a safety framework.


Recent advances in magnets and plasma confinement have led to an understanding of how commercial fusion is likely to unfold. Getting it working this decade requires collaboration by fusion research organizations, parallel development of multiple reactors, empowering the world’s top fusion scientists and engineers, and funding on the order of $ 10B over 8 years.

Climate change is likely to cost significantly more money. Therefore, governments and foundations should consider large initiatives to accelerate commercial fusion.

Leave a Comment