Back to InsightsAnalysis
March 31, 2026
A Japan powered by 90% clean energy is achievable only with a flexible system
New analysis shows how interconnectors, flexible nuclear and long-duration storage could help reduce wasted renewable energy in Japan’s power system.
Using interconnectors more flexibly can keep the curtailment well below 1% when renewables supply half of Japan’s electricity. However, as clean energy share rises to 90%, nuclear flexibility and long-duration storage show great promise.
Summary
In the first blog of this series, modelling by TransitionZero and ClientEarth showed that Japan could achieve 50% renewable electricity by 2040 with minimal curtailment. In this second blog, we explore a key follow-up question: can Japan reduce renewable curtailment even further through operating more flexibly several asset classes – namely pumped hydro, interconnectors and nuclear power?
Our modelling shows that Interconnector and nuclear flexibility could nearly completely eliminate curtailment at 50% renewable penetration. But in the 90% Clean scenario — which includes 74% renewable penetration and the rest from nuclear — curtailment rate rises to a concerningly high level of nearly 12%. Nuclear flexibility alone can lower curtailment to 7% — which is still economically unviable.
Further, given the tense politics around nuclear power in Japan, it is crucial to unlock other sources of flexibility. Interconnectors should be built as planned and fully utilised. Looking further ahead, as renewable penetration rises beyond 50%, authorities could then start supporting long-duration energy storage.
Recap: Curtailment could fall well below official estimate
TransitionZero has partnered with ClientEarth to explore how the 2040 Japanese power system might evolve under the 2025 7th Strategic Energy Plan (SEP7). The first phase of the study replicates Japan's national grid coordinator, Organisation for Cross-regional Coordination of Transmission Operators (OCCTO)’s 2021 modelling of Japan’s 2040 power system, updates it to reflect SEP7 targets, providing revised estimates of renewable curtailment. It explores three possible visions for the Japanese power system in 2040: a 50% renewables case (High-RES SEP7), a delayed-nuclear case compensated by thermal generation (Nuclear-delayed SEP7), and a more ambitious decarbonisation pathway featuring 74% renewables and 16% nuclear (90% Clean).
Our results suggest that despite the higher renewable target, the curtailment rate could be lower than OCCTO’s 2021 estimate. For the full breakdown, read our first blog in this series.
Building on those findings, this second phase explores how to eliminate ‘sticky’ curtailment that optimising the thermal capacity mix and distribution alone cannot fix. Specifically, we test the impact of adding operational flexibility to assets that OCCTO treated as relatively inflexible — namely pumped hydro, interconnectors, and nuclear power.
Modelling flexibility: What and how?
Our focus on these three asset classes is guided by Japan's 'priority dispatch rules', which dictate the order of generation reduction when supply exceeds demand.
We explored whether operations ahead of renewable curtailment in this queue could be managed more dynamically. For example, adapting Japan's long-standing practice of running nuclear plants in flat baseload mode to mirror the European practice of flexible adjustment.
Pumped hydro flexibility
In our main scenarios, the model is calibrated to replicate the daily and weekly operating behaviour of pumped hydro as represented in the 2040 OCCTO 45GW scenario (OCCTO’s scenario produced in 2021 exploring the deployment of 45 GW of offshore wind). Our analysis reveals that both the 2024 and 2040 OCCTO scenarios exhibit relatively flat charging profiles throughout the day, with only a modest peak in discharging during the evening hours. However, the 2040 scenario approximately doubles the nationwide average daily peak operating activity, with maximum peak charging and discharging reaching around 800 MW, compared to roughly 400 MW in 2024.
In contrast, the operating profile in the 2050 OCCTO MidFlex scenario (OCCTO’s scenario produced in 2023 exploring mid demand flexibility) evolves into a more pronounced "duck curve". Increased solar generation drives midday charging to a -2,200 MW minimum, followed by a rapid ramp-up to a generation peak exceeding 1,200 MW around hour 18. These variations show that, even with the same capacity, pumped hydro can be operated in very different ways, highlighting the potential for more flexible dispatch.
We define flexibility here as removing these inherited operational patterns, allowing pumped hydro to respond purely to the model's internal marginal price signals. We, however, do not model seasonal operating patterns, which could meaningfully influence dispatch patterns and curtailment outcomes.
Interconnector flexibility
As with pumped hydro, interconnectors in the model inherit flow patterns from 2040 OCCTO 45GW scenario, represented through simplified limits on how much the interconnectors can operate, approximating real-world congestion and operational constraints.
Compared with 2024, nationwide interconnector yearly utilisation rates increase in both the 2040 and 2050 scenarios, by 8 and 16 percentage points, rising from 22% to 38% and 30% respectively. These increases are largely driven by interconnectors transmitting surplus renewable generation from the northern and southern parts of the country to the three major demand centres (Tokyo, Chubu, and Kansai).
Despite the higher utilisation rates in OCCTO’s projections, there remains significant scope for increased regional flows. In our flexibility sensitivity, we therefore relax these constraints and allow interconnectors to respond more freely to regional supply–demand conditions.
Nuclear flexibility
In both the 2040 and 2050 OCCTO scenarios, nuclear power is assumed to operate as a flat baseload throughout all 8,760 hours of the year, with no ramping allowed, at capacity factors of 59% and 70% respectively.
To assess the potential role of nuclear flexibility, we introduce a modest level of operational flexibility, drawing on practices observed in France. While acknowledging that not all French reactors exhibit the same flexibility, for this exploratory study we assume that all Japanese reactors will operate similarly. We implemented this French-inspired nuclear flexibility through three key modelling constraints: nuclear output may ramp up or down by no more than 60% of nameplate capacity per hour; daily average ramping events should not exceed 2 throughout the year; and the annual average utilisation rate is capped at 70%.
Does flexibility actually reduce waste?
Interconnectors work, until the grid saturates
Freeing up interconnectors sharply reduces curtailment at 50% renewable penetration, cutting the curtailment rate from 1.5% in both the High-RES SEP7 and Nuclear-delayed SEP7 scenarios to 0.8% and 0.5% respectively. This is unsurprising as surplus generation in one region can flow freely to others. However, in the 90% Clean scenario with renewable penetration rising to 74%, Japan becomes saturated with renewables. With renewable generation highly correlated across all zones, surplus events increasingly coincide, leaving nowhere for excess power to flow. Consequently, curtailment rates remain virtually unchanged between the 90% Clean scenario and its corresponding sensitivity.
Nuclear flexibility is a heavy hitter
At 50% renewable penetration, flexing nuclear nearly eliminates curtailment entirely, dropping the rate from 1.5% to just 0.4%. Crucially, even in the saturated 90% Clean scenario, nuclear flexibility remains impactful, dragging the curtailment rate down from 11.7% to 7.1%.
By dropping output during solar peaks or shutting down over low-demand weekends, flexible nuclear directly addresses the temporal mismatch between renewable generation and demand, allowing surplus renewables to be absorbed locally rather than curtailed. Importantly, this is achieved while maintaining capacity factors largely consistent with the main scenarios.
Batteries beat pumped hydro at 50% renewables
Last but not least, pumped hydro flexibility, unlike the other two flexible options, reduces curtailment very marginally across the main scenarios. At 50% renewable penetration, pumped hydro (assumed 70% round-trip efficiency) is largely outperformed by smaller, more efficient batteries (assumed 90% round-trip efficiency), which already eliminates most economically recoverable curtailment.
In the 90% Clean scenario with 74% renewable penetration, pumped hydro however sees increased utilisation thanks to its larger energy capacity, pointing to the rising need for long-duration energy storage.
Looking Ahead
How Japan could adapt
Our study highlights three key insights for managing curtailment in high-renewables Japan:
- Making full use of the interconnectors halves curtailment but effectiveness vanishes almost entirely as more renewables penetrate the system.
- Running Japanese nuclear reactors more flexibly like French ones eliminates nearly all curtailment at 50% renewable penetration and remains effective even when the system is saturated by renewables.
- Operating pumped hydro more flexibly is useful only at 90% clean energy, when their storage size outcompetes batteries’s efficiency.
Looking toward 2040 and given the present tense politics around nuclear power in Japan, the most actionable lever is ensuring that interconnectors get built as envisioned by OCCTO, and that they are used flexibly. With a view towards the longer term or when renewable penetration exceeds 50%, the Japanese authorities should rethink the regime around nuclear reactor operations, and stimulate long-duration energy storage.
Better modelling of flexibility
Across all scenarios and sensitivities modelled, Hokkaido, Tohoku, and Kyushu remain the regions most exposed to persistent curtailment, likely reflecting structural imbalances between local supply and demand.
We therefore wonder if a shift toward a bottom-up modelling approach that evolves from today’s grid through to 2040, allowing economic signals and investment returns to determine what gets built, could better identify the regionally targeted generation, storage, and flexibility investments needed to alleviate structural curtailment.
This blog is the second in a two-part series based on research by Alex Luta and Joel Yap for ClientEarth. Read part one here.
