Nuclear share of Japanese electricity generation (Ember, calendar year): 25.3 % in 2010, 14.8 % in 2011, 0.0 % in 2014, 9.1 % in 2025. That last one is 94 TWh.

Fleet today (net capacity, dataset of 2026-09-10, IAEA PRIS + Global Energy Monitor): 15 units operating, 13,946 MW. 18 units suspended, 17,733 MW. 22 units retired since 2011, 15,501 MW. 2 under construction plus 1 halted.

The part I did not expect. Median 2024 load factor of Japanese units: 85.0 % (n = 14). World median: 85.6 % (n = 401 operating units with a published figure). Two of the Japanese units restarted in the last weeks of 2024 (Onagawa 2 on 29 Oct, 568 GWh for the year; Shimane 2 on 7 Dec, 146 GWh), so their annual figures are partial-year artefacts. Excluding them, the twelve full-year units have a median of 85.7 %, above the world median. Japan ranks 6th of the 15 countries with 5+ units by median load factor, ahead of Russia, Canada, Spain and France.

Where the growth has to come from: 14 of the 18 suspended units are BWRs (14,659 MW). No Japanese BWR restarted at all until October 2024; three have now, most recently Kashiwazaki-Kariwa 6, back in commercial operation on 16 April 2026 after 14 years offline, the first TEPCO unit to generate since Fukushima. Of the 18 suspended units, 9 have a restart review filed with the NRA (8,664 MW), 8 have no application on record (7,961 MW), and Tsuruga 2 was ruled out in Nov 2024 over an active fault. Tomari 3 slipped after a fatal accident at its seawall works in August 2026.

Base note, because it bites: the restarted KK-6 is quoted as 1,356 MW gross, 1,315 MW net. The country page on the atlas shows 15 reactors and 13,946 MW because it counts only operating units; the map panel shows 33 and 31,679 MW because it adds the suspended ones, which are still standing there. Both are on the site, they answer different questions.

Source: IAEA PRIS, Global Energy Monitor and Ember, assembled in Reactor Atlas, a map I build. Data and method per reactor: https://reactoratlas.com/en/country/jp

Method question worth arguing about: is comparing a restarted fleet's load factor against a world median fair, when the world median includes units in their own partial years? I excluded Japan's two partial-year units but not everyone else's. Doing it symmetrically is more work than it sounds.

—
fedecaccia

2 comments

  2. It’s crazy to think how much more margin we’d have to address the climate crises both in political terms and raw physical facts without what has been essentially a freeze on the technological development and global buildout of nuclear power since the 60s-70s.

Comments are closed.