Does building the same design again make it faster?
Of 43 reactor designs with at least 4 finished units, the later half of the series was built more than six months faster than the first two units for 16, and more than six months slower for 15. Repeating a design has not, on its own, made reactors quicker to build.
All designs together
| Position in the series | Units | Build time against the first two |
|---|---|---|
| 1st and 2nd | 86 | 1.00 |
| 3rd to 5th | 120 | 0.98 |
| 6th to 10th | 119 | 1.06 |
| 11th onward | 167 | 1.11 |
Why repetition is not the whole story
A long series often spans a change of era, and the era can matter more than the design. The Soviet VVER-1000 (V-320) units started between 1980 and 2010: the first two took 5.8 years, the later half 13.9, in a series whose building ran across the collapse of the Soviet Union. Designs built in one country by one programme give a cleaner read: 25 of the 43 were built in a single country, and the later half was faster for 9 of them.
Design by design
Units in the order construction started. "First two" is the median of the first two units; "later half" is the median of the second half of the series. Build time runs from construction start to grid connection.
| Design | Built in | Units | Starts | First two, years | Later half, years | Change |
|---|---|---|---|---|---|---|
| BWR-4 | 5 countries | 30 | 1967-1980 | 5.4 | 5.5 | +0.1 |
| Westinghouse 4-loop | United States | 30 | 1966-1977 | 11.6 | 12.0 | +0.4 |
| MAGNOX | United Kingdom, Italy, Japan | 28 | 1953-1963 | 3.3 | 5.5 | +2.2 |
| Westinghouse 3-loop | 4 countries | 25 | 1967-1980 | 4.5 | 8.6 | +4.2 |
| VVER V-320 | 4 countries | 25 | 1980-2010 | 5.8 | 13.9 | +8.1 |
| BWR-5 | United States, Japan, Mexico | 24 | 1972-2000 | 10.4 | 3.8 | -6.6 |
| VVER V-213 | 7 countries | 20 | 1971-1987 | 7.0 | 7.8 | +0.8 |
| P4 1300 MWe series | France | 20 | 1977-1984 | 6.8 | 7.6 | +0.8 |
| CP1 | France, South Africa | 20 | 1974-1979 | 5.6 | 5.3 | -0.3 |
| Indian PHWR | India | 19 | 1965-2011 | 10.0 | 8.0 | -2.0 |
| CPR-1000 | China | 18 | 2005-2012 | 4.7 | 5.5 | +0.8 |
| RBMK-1000 | Russia, Ukraine | 15 | 1970-1984 | 5.7 | 5.7 | +0.0 |
| AGR | United Kingdom | 15 | 1958-1980 | 10.9 | 10.6 | -0.3 |
| Westinghouse 2-loop | 6 countries | 13 | 1965-1975 | 3.7 | 5.5 | +1.7 |
| VVER V-230 | 4 countries | 12 | 1970-1973 | 4.3 | 7.3 | +3.0 |
| OPR-1000 | South Korea | 12 | 1989-2008 | 5.0 | 4.2 | -0.8 |
| CANDU 6 | 5 countries | 11 | 1974-1998 | 8.9 | 4.6 | -4.3 |
| CP2 | France | 10 | 1976-1981 | 4.8 | 5.8 | +1.0 |
| BWR-3 | United States, Spain, Japan | 9 | 1966-1970 | 4.4 | 3.7 | -0.7 |
| Mitsubishi 3-loop | Japan | 9 | 1970-2004 | 3.9 | 3.5 | -0.4 |
| HPR-1000 | China | 8 | 2015-2020 | 5.8 | 5.2 | -0.6 |
| APR-1400 | South Korea, United Arab Emirates | 8 | 2008-2015 | 8.5 | 8.5 | +0.1 |
| BWR-6 | 4 countries | 8 | 1974-1977 | 10.4 | 7.5 | -2.9 |
| Combustion Engineering 2-loop | United States | 8 | 1967-1977 | 5.7 | 6.0 | +0.3 |
| Mitsubishi 2-loop | Japan | 7 | 1968-1985 | 3.7 | 3.5 | -0.2 |
| B&W 177-FA | United States | 7 | 1967-1970 | 5.8 | 6.5 | +0.7 |
| CNP-600 | China | 6 | 1996-2010 | 6.3 | 5.5 | -0.8 |
| CNP-1000 | China | 6 | 2008-2016 | 6.6 | 4.7 | -1.9 |
| AP-1000 | China, United States | 6 | 2009-2013 | 9.1 | 10.1 | +1.0 |
| ASEA-Atom BWR | Sweden, Finland | 6 | 1973-1980 | 5.8 | 4.8 | -1.0 |
| CP0 | France | 6 | 1971-1974 | 5.6 | 5.1 | -0.6 |
| CNP-300 | China, Pakistan | 5 | 1985-2011 | 6.8 | 5.4 | -1.4 |
| Mitsubishi 4-loop | Japan | 5 | 1982-1992 | 3.6 | 4.3 | +0.7 |
| Westinghouse F | South Korea | 5 | 1977-1981 | 5.3 | 4.9 | -0.4 |
| VVER-1200/V-491 | Russia, Belarus | 4 | 2008-2014 | 9.9 | 8.0 | -1.9 |
| CANDU 750A | Canada | 4 | 1970-1972 | 5.7 | 5.9 | +0.2 |
| CANDU 750B | Canada | 4 | 1978-1979 | 6.5 | 7.2 | +0.7 |
| CANDU 850 | Canada | 4 | 1981-1985 | 8.5 | 8.0 | -0.5 |
| CANDU 500B | Canada | 4 | 1974-1976 | 8.1 | 9.1 | +0.9 |
| M310 | China | 4 | 1987-1997 | 6.0 | 4.8 | -1.2 |
| ACPR-1000 | China | 4 | 2013-2015 | 5.1 | 6.5 | +1.4 |
| N4 1450 MWe series | France | 4 | 1984-1991 | 12.0 | 9.0 | -3.0 |
| ABWR | Japan | 4 | 1992-2001 | 3.3 | 3.8 | +0.5 |
Only units with a design line on record are counted: 156 of the 657 finished units with both dates have none in the tracker and are left out. The record holds no construction costs, so this says nothing about whether repeat units were cheaper. See build times by decade for the trend across all units. Download: repeat-builds.csv.
Sources
- Global Nuclear Power Tracker, Global Energy Monitor, August 2026 release , CC BY 4.0 · retrieved 2026-09-06
How these sources are used, including the corrections applied on top of them and what is deliberately left blank · download the underlying data
Cite this page
NuclearChain. "Does building the same design again make it faster?." https://nuclearchain.net/learning-curves/. Accessed 2026. Built on the Global Nuclear Power Tracker, Global Energy Monitor, CC BY 4.0.
Free to reuse under CC BY 4.0 with that attribution. Download the data.