I recently met some senior officers of a private firm aiming to develop a nuclear fusion reactor. Fusion sticks together light atoms like hydrogen; today’s fission reactors split heavy atoms like uranium. Fusion powers the sun and hydrogen bombs. The U.S. has spent tens of billions of tax dollars on fusion R&D—currently $0.4 billion a year—and now startups are emerging too, some with nine-figure budgets.
As a recovering physicist, I thought the proposed technology was intriguing. Fusion can also be made safer and cleaner than fission. But I fear the investors are likely to be disappointed. The firm’s officers had never before heard that their business case was fatally flawed because they’d assumed the wrong competitor.
Advocates of fission reactors (which have their own startups), coal-fired, and even gas-fired power plants often make the same mistake. So to help capital markets understand the competitive landscape, I’ll summarize here a brief tutorial adapted from a comment I posted 22 August 2014 on a Bulletin of the Atomic Scientists feature article by Jason Parisi. He urged a rapid transition from fission to fusion energy to help meet the 5–7-fold increase in global electricity demand he expects in this century while reducing the risk of spreading nuclear weapons. Unfortunately, his argument, like all others I’ve seen for fusion power, falls at the first hurdle—basic economics.
Mr. Parisi cited a 2006 summary paper from an elaborate multi-decade fusion design study as evidence for his paraphrased claim that “first-generation fusion reactors would be competitive with renewables and fossil fuel, even without a carbon tax.” But the paper he cites doesn’t say that. It claims competitiveness only with “current energy sources,” not mentioning renewables. Modern renewables now make the claim untrue.
Mr. Parisi’s cited source summarized the results of a rather old (≤2005) calculation of fusion power sent out from a billion-watt reactor for 4.8¢/kWh in 1992 $, equivalent to 7.2¢/kWh in 2011 $. Scaling up to an even more speculative 1.5-billion-watt design would supposedly drop this price to ~6.2¢/kWh. But if that calculation for an extraordinarily complex machine not yet built were accurate—for the first time in the history of fission or fusion technology—it’d still only match today’s best unsubsidized levelized market price of utility-scale solar power in the sunbelt. It’d also be two-thirds higher than today’s best unsubsidized levelized market price of windpower in the midwestern windbelt. [1] Those prices, too are rapidly dropping with no end in sight—utility-scale photovoltaic power prices just halved 2006–11 and halved again 2011–13—while the prices of central-thermal-plant technologies are generally rising. Nor would a fusion reactor’s steady operation add significant value. The costs of reliably integrating variable renewables (photovoltaics and windpower) into the grid, even at high fractions of total generation, are typically just a few tenths of a cent per kWh. The costs of reliably integrating central thermal stations, like modern billion-watt coal and nuclear power plants, are usually omitted but may be larger [2].
Fusion’s economic challenge is fundamental. Let me explain it in three steps:
1) Modern fission plants are grossly uncompetitive. That’s why none of the 60+ units officially under construction worldwide was fairly competed or compared with available alternatives—all were bought by central-planning processes; why no new nuclear unit has been bid into a competitive power auction (three old units just failed to compete in a midwestern auction based on operating cost alone—a growing generic problem); and why even ~100+% construction subsidies offered to new U.S. reactors since 2005, plus operating subsidies slightly greater than those formerly given to windpower, have failed to attract private risk capital, e.g. to build a merchant plant that must make its way in wholesale power markets.
2) Roughly 35% of a new fission plant’s prohibitive capital cost is for its “nuclear island”—the nuclear steam supply system, generously construed. The rest is for the non-nuclear components, such as turbine, alternator, switchyard, heat sink, general controls and buildings, site, etc.
3) New fission reactors are so costly—the subsidized Hinkley Point C market price is at least 3–4 times the unsubsidized market price of midwestern U.S. windpower—that even if the nuclear island (the source of the steam) were free, the rest of the plant would still be grossly uncompetitive. Whether the steam is made from burning fossil fuel or from fission or fusion or from burning energy studies, any sort of big new thermal power station is now fundamentally uneconomic. That’s why the world has instead invested over a quarter-trillion dollars in non-hydro renewables, adding >80 billion watts, in each of the past three years; why the majority of investment and capacity additions in the power sector is now renewable (in 2013, 53% in the world, 68% in China, 72% in Europe); and why orders for thermal stations are fading: they simply cost too much and present too much financial risk.
Potential pricing of U.S. carbon emissions is of course irrelevant to competition between fusion and renewables (let alone end-use efficiency), since those too are carbon-free. Competition with fossil-fueled generation is generally irrelevant because those are out of the money too. And renewables cannot spread nuclear weapons—unlike fission and even fusion (which, for the orthodox design Mr. Parisi describes, is far from benign because its copious flux of fast neutrons could be inconspicuously used to breed bomb materials like plutonium).
Some believe the economics may become better for small modular fusion reactors. I await a convincing analysis. But since fusion, like fission, is just a way to raise steam to turn a turbine to drive an electric generator, the analogy of small modular fission reactors is instructive. As a matter of physics, fission reactors don’t scale down well; that’s why we build them big. The notion of overcoming small modular fission reactors’ higher capital cost per kilowatt by achieving economies of mass production then runs into another kind of SMR—small modular renewables. These do scale down very well into mass-producible modules, and are already decades ahead in exploiting their powerful economies of scale. Fission SMRs are far too late to have a hope of catching up. This is not encouraging for fusion SMRs, unless perhaps new physical principles emerge in practical forms that can bypass the relatively high capital cost of steam cycles. But they would need to be very cheap indeed to beat the modern renewables and efficiency that are now taking over the market.
Ah, yes, efficiency: did I mention that three-fourths of U.S. electricity can be saved at an average technology cost around 0.7¢/kWh? To be sure, U.S. utilities often pay several times that much because their programs have transaction costs and may be suboptimally designed, but they’re getting better, and regardless, their $7 billion of annual investment is far cheaper than building and running any nonrenewable power generator. The case for efficiency is even stronger in developing countries because they're less efficient to start with, can less afford waste, and are building more of their infrastructure, so they can more easily build things right than fix them later.
In short, the most promising fusion reactor is already in place, free, well engineered, and appropriately sited 93 million miles away. There's no business case for reproducing it in miniature form on earth—if you understand who your competitors are. Caveat investor.
[1] For readers who want to reconcile with published Power Purchase Agreement prices, whose levelized versions appear in Lawrence Berkeley National Laboratory’s annual market reports, the best utility-scale photovoltaics (the 2013 report is to be published imminently) are around 4.7¢/kWh net of 30% Federal tax credit and no state subsidy. Windpower is as low as 1.8¢/kWh net of a Federal tax credit with a levelized value around 1.8¢/kWh—consistent with Bloomberg New Energy Finance’s April 2014 Summit datum of 3.7¢/kWh unsubsidized.
[2] For readers who think big thermal plants operating relatively steadily—in this case, 85% of the time—cannot be validly compared in levelized cost (or in functionality for ensuring reliable power supply) with variable PVs or windpower, whose best modern U.S. capacity factors are respectively >30% and >50%, please see here, here, and here with its citations.
