Paraguay generates almost all of its electricity from water, and yet the question about small modular reactors (SMRs) keeps coming up in planning forums. There are reasons: demand keeps growing and the Paraná’s long droughts are fresh in everyone’s memory. The two charts above give the answer away: the first compares timelines to first revenue and the second lays out the region’s PESTEL grid.
The numbers and capital recovery
SMRs still don’t have a commercial track record to plan with their figures, but the most advanced case in the West already leaves an expensive lesson. The NuScale/UAMPS project in the United States starts from a target price of 58 dollars per MWh that by 2023 already stands at 89, and that with federal subsidies: without them it sits around 119 (NuScale/UAMPS and IEEFA, 2023). The estimated build goes from 5.3 to 9.3 billion dollars for 462 MW, about 20 million per MW, and since November 2023 the project stands canceled (NuScale/UAMPS, 2023). As a reference, Lazard (2024) puts new nuclear between 142 and 222 dollars per MWh, utility solar between 29 and 92, and solar with batteries between 60 and 210.
Capital recovery is where the math gets settled. A solar plant earns revenue after two years; an SMR after ten or more (the typical timelines in the first chart), and in Paraguay you have to add the regulatory framework. Every year of waiting is idle capital paying interest without producing one MWh, and at current financing rates that gap of 5 times or more in the time to first revenue dominates the math: the same capital put into solar with batteries, small hydro or biomass today is already working before the SMR gets a building permit. That only changes if the SMR’s cost per MWh drops a lot, with published evidence.
The gap
Paraguay has no nuclear energy law or regulator; that framework is built from scratch and takes longer than any of the local options. The detail that usually goes unnoticed: the 10-plus years in the first chart are counted in countries where the regulator already exists. Here that clock only starts after the framework is built.
The regional board in the second chart shows where that experience lives, and it lives close by. Argentina operates three reactors under its Law 24.804 and its ARN regulator, and has the region’s only home-designed SMR prototype, CAREM, with civil works well advanced and a contract signed with Nucleoeléctrica Argentina to complete the reactor building (CNEA and NA-SA, 2026). Brazil generates with Angra 1 and 2, with Angra 1’s license renewed to 2044 (CNEN, 2024), and since 2021 has a nuclear safety authority created by law (ANSN, Lei 14.222/2021). Mexico operates both units of Laguna Verde with licenses renewed to 2050 and 2055 (CNSNS, 2020 and 2025). For the monitor we propose to build in Paraguay that is first-hand information without putting up capital: every review can read how the neighbors are doing with their reactors and with CAREM, and that reading is worth more than any vendor brochure.
The decision today
Today an SMR is not an investment for Paraguay: against the local options it loses on cost, on timeline and on risk. Nuclear stays on watch, which is the cheap position. The signals we propose to watch are five: three or more Western SMRs operating with cost and schedule overruns published and contained, per-unit cost falling across repeated units of the same design, real stress on the local power system beyond what solar, small hydro and biomass cover, a Paraguayan nuclear law with its regulator under way, and Argentina’s CAREM connected to the grid and delivering power, the closest regional proof there is.
The data: the NuScale/UAMPS case and Lazard’s ranges, ANDE’s firm demand series and multi-year Paraná hydrology, refreshed at every review.
The technique: compared LCOE plus real options: the value of waiting is calculated, not guessed, and that is how we design the monitor: each review adjusts the probability that the signals to watch light up.
The decision: where to invest the same capital today (solar with batteries, small hydro or biomass) and what would change the answer: you commission the serious study when that monitor shows three of the five signals past their threshold.
How to measure it: tracking those five signals, reviewed every two or three years. Cheap and enough as long as none of them light up.
These five signals are published to be argued with: if you read them with a different threshold, or see a sixth one missing from the list, write to us and we’ll go over it with data. The same goes if you’re putting together a twenty-year energy plan, or deciding where to put that capital today, and the nuclear question has already landed on your table.