When the Fukushima Daiichi accident struck in 2011, Japan had 48 other reactors. By 2013 every one of them had been taken offline, and the country replaced the lost output almost entirely with imported liquefied natural gas. Thirteen years later, the return of nuclear power has been far slower than either its supporters or its critics expected. Yet the effect on gas imports is already visible, and it is large enough to matter for an LNG market that is about to receive a wave of new supply.
A restart programme measured in single units
Japan allowed its first reactor to resume operation in 2015. Since then, according to the US Energy Information Administration, 12 units have restarted, bringing operating nuclear capacity to about 11 gigawatts. Ten more units are under review by the Nuclear Regulation Authority, and five have passed review but have not yet returned to service.
The most recent additions show how the process works. Kansai Electric restarted Takahama Unit 1 on 28 July 2023, more than 12 years after it entered an inspection outage in January 2011. The 780 megawatt pressurised water reactor first operated in 1974, and in 2016 the regulator approved it and its sister Unit 2 to run for up to 60 years, the first Japanese units granted an extension beyond 40 years under the post-Fukushima regulations. Unit 1 returned to full commercial operation at the end of August 2023, and Unit 2 followed later in the year. Together the two units added about 1.6 GW.
The timeline from application to restart for the Takahama pair was more than eight years. Kansai applied in March 2015, the regulator confirmed compliance in April 2016, and the restart was further delayed by additional fire protection work. That is typical. Each restart requires regulatory approval, local consent and often court challenges, and each proceeds at its own pace.
The gas effect
Small as the restart programme has been, its effect on gas demand is measurable. The EIA reports that Japan's LNG imports in 2022 were 15 per cent lower than in 2015, a fall of 1.7 billion cubic feet per day. When five reactors restarted in 2018, imports fell 7 per cent, or 0.7 Bcf/d, in that year alone. The agency expects imports to keep declining as more reactors return.
Fossil fuels still supplied 71 per cent of Japanese generation in 2022, with natural gas accounting for 35 per cent of total generation. So there is plenty of gas-fired output that nuclear restarts could displace. Each gigawatt of nuclear running at a high capacity factor replaces a significant volume of LNG over a year.
The contract problem
Japan's gas purchasing is built on long-term contracts, which supply up to 90 per cent of its LNG imports in a typical year. The remainder comes from short-term and spot purchases from as many as 20 countries, including re-exporters.
That structure creates a mismatch when demand falls. Japanese utilities hold contracts signed years ago for volumes that may now exceed what they need, particularly in years when several reactors restart. The options are to resell surplus cargoes, to divert them to other buyers in Asia, or to renegotiate. Japanese companies have become active LNG traders partly for this reason. Their portfolios now include cargoes that never reach Japan.
The supplier mix has also shifted. Australia has been Japan's largest supplier for 11 years, and its share of Japanese imports more than doubled from 18 per cent in 2012 to 42 per cent in 2023. Qatar, the second-largest supplier in 2012 with an 18 per cent share, provided only 4 per cent in 2023 and ranked seventh, partly because some long-term contracts expired. Malaysia remains a significant source.
The Qatari decline is instructive. When contracts expired, Japanese buyers chose not to renew at the same volumes, in part because they were uncertain about future demand given nuclear restarts. That uncertainty is a cost. It makes buyers reluctant to commit, and in a tight market it leaves them more exposed to spot prices.
Energy security cuts both ways
The case for nuclear restarts in Japan is often made in climate terms, but the stronger argument in Tokyo has been energy security. Japan has very few domestic energy resources and imports most of the fossil fuel it burns for power. Every unit of gas displaced by nuclear output is a unit not bought on world markets and not shipped through contested sea lanes.
The 2022 price spike made that argument vivid. When European buyers bid aggressively for LNG after Russia's invasion of Ukraine, Asian spot prices rose to record levels. Japanese utilities with surplus nuclear output were better protected than those without.
But there is a second security consideration. Utilities that reduce their long-term LNG commitments because they expect nuclear restarts are betting that those restarts will happen on schedule. If a court injunction or a seismic finding takes a reactor offline unexpectedly, the utility must buy replacement fuel at short notice. Long-term contracts provide insurance against that risk. Shedding too many of them in anticipation of reactors that do not arrive would leave Japan more exposed, not less.
What the next decade holds
The Japanese government's GX decarbonisation policy treats nuclear as a central part of the power mix, and recent legislation allows reactors to operate beyond 60 years by excluding time spent offline for safety reviews. If even half of the units still under review return to service by the end of the decade, Japan's LNG imports would fall further.
For the global LNG market, that matters at the margin. Japan has long been one of the largest LNG importers, and its declining demand helps explain why new supply from the United States, Qatar and elsewhere is increasingly being contracted by portfolio players and by buyers in China, South and Southeast Asia, and Europe rather than by Japanese utilities.
For Japanese planners, the policy question is how to combine restarts with a sensible contract book. The answer is probably to keep a core of long-term supply that covers demand even if restarts slip, and to rely on resale capacity to manage the surplus when they do not. That is more expensive than betting on reactors. It is cheaper than being caught short in a winter like 2022.