Shota Furuya

Pieces日本語版

Isn't renewable energy too expensive up front? Can the investment be recovered?

“Renewables are expensive up front” was a fair criticism twenty years ago. In the middle of the 2020s, solar and onshore wind are already the cheapest new generation in the world, and the capital cost is firmly in territory that pays back.

This piece answers the question — can the investment be recovered? — through the latest trend in levelised cost of energy (LCOE), real payback periods in Japan, newer ways of financing such as power purchase agreements, and the comparison with fossil fuels including external costs and stranded-asset risk.

The short answer: (1) equipment costs have fallen by more than 70% in a decade, (2) residential solar in Japan commonly pays back in seven to ten years and commercial self-consumption in five to eight, (3) corporate and on-site PPAs let a user install with no up-front investment at all, and (4) it is the fossil side that has entered an era of fuel price shocks, carbon prices and stranded assets where the investment may not be recovered.

“Expensive up front” belongs to the past: what LCOE shows

The international yardstick for comparing generation is LCOE (levelised cost of energy) — capital, operation and maintenance, fuel and decommissioning divided by the electricity generated, in effect the lifetime cost per kWh.

IRENA’s latest edition, “Renewable Power Generation Costs in 2024” (published June 2025), puts the global weighted-average LCOE of renewables commissioned in 2024 at US$0.043/kWh (about ¥6.5/kWh) for utility-scale PV — 90% below the US$0.417/kWh (about ¥63/kWh) of 2010 — and lower still for onshore wind at US$0.034/kWh (about ¥5.1/kWh), the cheapest new generation of any kind. Total installed cost for PV fell 11% year on year to US$691/kW (about ¥104,000/kW).

By region the figures vary with supply chains, labour costs and regulation: US$0.033/kWh (about ¥5.0) in China, US$0.038/kWh (about ¥5.7) in India, US$0.070/kWh (about ¥10.5) in the United States.

Fossil thermal generation, the comparison, moves with fuel prices across roughly US$0.05–0.20/kWh (about ¥7.5–30/kWh). Conversions below use US$1 = ¥150 for convenience.

Two points stand out.

First, LCOE ticked up slightly in 2024 — 0.6% for PV, 3% for onshore wind. That was a year in which capacity factors, the regional mix and financing costs (higher interest rates) outweighed falling equipment prices; the long downward trend has not ended, and installed costs themselves hit a new low, down 11%.

Second, even so, utility-scale PV is on average 41% cheaper than the cheapest fossil generation, and onshore wind 53% cheaper. IRENA finds that 91% of utility-scale renewable capacity added in 2024 delivered electricity at a lower cost than the cheapest fossil option, and estimates that renewables avoided about US$467 billion (on the order of ¥70 trillion) in fuel costs in 2024.

New solar and onshore wind, in other words, have reached a level below the running cost of existing thermal plants. The intuition that renewables are expensive up front is a memory of module prices around US$2/W circa 2010. By 2024, crystalline silicon modules were trading below US$0.10/W on the spot market, and whole-system prices are a third to a fifth of what they were a decade ago.

How long payback takes in Japan

Land preparation, grid connection work and labour are expensive in Japan, so costs run above the global LCOE. Payback still works comfortably.

Residential solar (under 10 kW) costs about ¥250,000–280,000/kW according to the METI procurement price committee’s FY2024 data, so a typical 4 kW system costs ¥1.0–1.2 million to install. The return combines savings on self-consumed electricity with the FIT surplus tariff (¥16/kWh in FY2024, for ten years) — and with retail electricity above ¥30/kWh once fuel cost adjustments and the renewable levy are included, the self-consumed portion is worth a great deal. That puts payback at seven to ten years, against a panel life of 25–30 years, after which the generation is close to pure gain.

Commercial self-consumption (rooftop, 50–500 kW) benefits from high-voltage tariffs that rose to ¥20–25/kWh in 2023–24, which makes self-consumed electricity extremely valuable; payback of five to eight years is the usual range. Support such as SII subsidies under the revised Energy Conservation Act, and demand-response subsidies for co-located storage, can shorten it further.

Utility-scale projects (FIP or auction) worked to a FY2024 auction ceiling of ¥9.2/kWh for solar (onshore wind is set separately), designed around 20 years of operation and an IRR target of 5–7%. Project finance from banks now treats these as standard structures: they are no longer a special kind of investment.

The models that remove the up-front investment

Paying cash up front is indeed heavy. But models where the user does not carry the capital cost have spread widely, including to households and smaller companies.

One is the on-site PPA (third-party ownership model): the user lends the roof, and the PPA operator owns and runs the system. The user pays a contracted rate for the electricity — say ¥12–18/kWh — and gets clean power with no capital outlay and maintenance included, along with the renewable attributes (non-fossil certificates) that make it usable for RE100.

An off-site corporate PPA, by contrast, is a long-term contract with a plant somewhere else. Amazon, Toyota and AEON among others have signed large deals that are pulling the Japanese market along.

In housing, third-party ownership combined with leasing — “zero-yen solar” — is now established, so a household can install without paying cash. And local models are spreading too, through municipal utilities and community PPAs that keep money circulating within a region, supported by local governments and by ISEP.

What all of these share is that they remove the premise of a large up-front investment. For a user, adopting renewables has stopped being an investment decision and become a choice of electricity supplier.

Storage economics are improving fast too

A common rejoinder runs: renewables vary, so you need batteries, so it is expensive after all. But lithium-ion costs have fallen dramatically over the past decade and are still falling. BloombergNEF’s Lithium-Ion Battery Price Survey 2025 (December 2025) puts the weighted average pack price across all segments at US$108/kWh (about ¥16,000/kWh), down 8% year on year and a new low.

The striking figure is stationary storage packs at US$70/kWh (about ¥11,000/kWh), down 45% in a year — for the first time the cheapest segment of all, below EV packs. Turnkey BESS systems including power conversion and installation came in at US$117/kWh (about ¥18,000/kWh), down 31%. By chemistry, LFP is at US$81/kWh (about ¥12,000) and NMC at US$128/kWh (about ¥19,000), as the shift to low-cost LFP continues.

Global overcapacity in cell manufacturing, price competition and the move to LFP are the main causes. In Japan, prices have come down to roughly ¥100,000–150,000/kWh for commercial systems and around ¥100,000/kWh for households, and subsidies improve the payback case again.

With time-of-use tariffs, demand response and the balancing markets, business models in which storage pays for itself — grid-scale batteries — are multiplying in Japan too; the FY2024 auction for grid storage drew bids on the scale of several gigawatts.

The risks: curtailment and grid constraints

The things that can undermine payback deserve stating honestly. First, curtailment. Renewable output has been curtailed routinely in Kyushu since 2018 and in Tohoku and Chugoku since 2022, which hits the revenue of FIP projects directly. Second, the cost of grid connection works: in parts of Hokkaido and Tohoku, connection charges running to hundreds of millions of yen have undermined projects. Third, the time it takes — connection studies not uncommonly run one to two years.

None of these is a physical limit; they are problems of institutional design, and policy is steadily improving them through non-firm connections, full operation of redispatch, and reinforcement of interregional links (the Hokkaido–Honshu link and the eastern interconnection from 2027). Put another way, this is not a matter of investments that cannot be recovered, but of policy design that can make recovery certain.

ISEP’s analysis published in April 2026, Curtailment of renewables in Japan is a problem of rules and operation, takes the case of 27 April 2025, when 16.5 GW of renewable output was curtailed nationwide, and demonstrates that this stemmed not from physical grid limits but from operation and rules — pumped hydro (3,800 MW) and the downward range of LNG plants (1,248 MW) went unused. The report sets out nine measures on a timeline, which can be summarised as follows.

Immediately: make day-ahead plans subject to mandatory intraday optimisation, and add renewable priority clauses to long-term bilateral contracts. Within one to two years: ensure the batteries already cleared in the capacity market actually operate during FY2026 (estimated to cut monthly curtailment by 62%), and create compensation for curtailment (at ¥30/kWh or more). Over two to five years: formalise low-output operation of nuclear plants in spring (a further 19% reduction), strengthen the wide-area balancing market, and give the priority dispatch rules legal force. On the demand side, shifting electric water heaters to daytime operation and spreading smart control of household batteries should be combined with the rest.

What would reduce curtailment, in short, is already known; what remains is political decision and speed of implementation. For investors and developers, designing project finance with those developments priced in — hedging curtailment risk, co-locating storage, contracting directly with users through corporate PPAs — makes recovery more certain still.

The comparison that matters: can fossil investments be recovered?

Finally, the most important point. If we ask whether renewables pay back, fairness demands the same question of fossil generation.

First, fuel price volatility. After the invasion of Ukraine in 2022, LNG spot prices briefly passed US$70/MMBtu and fuel cost adjustments hit Japanese household bills directly. Fossil generation is structurally vulnerable to a single geopolitical event; renewables, with no fuel cost, are free of that risk.

Second, carbon pricing. Japan’s GX-ETS emissions trading system starts in earnest in FY2026, and a levy on fossil fuels follows in FY2028. The EU’s carbon border adjustment mechanism is already in force. These add directly to the LCOE of thermal plants.

Third, stranded-asset risk: in the IEA’s net-zero scenario, much new coal and gas capacity loses its economics before the investment is recovered.

Fourth, falling capacity factors. As renewables spread, existing thermal plants run less, which makes fixed costs harder to recover. Capacity markets compensate for that, but the money comes from consumers, and the arrangement worsens the overall cost to society.

Measured by the recovery of an up-front investment, in other words, it is now the fossil side that carries the greater risk.

Summary

  • On IRENA’s 2024 data, utility-scale PV is at US$0.043/kWh (about ¥6.5/kWh) and onshore wind at US$0.034/kWh (about ¥5.1/kWh) — 41–53% cheaper than fossil fuels, the cheapest generation there is.
  • In Japan, seven to ten years is standard payback for residential solar, five to eight for commercial self-consumption.
  • PPA models make adoption possible with no up-front investment.
  • Battery costs have collapsed to a level where grid storage stands as a business in its own right.
  • Curtailment and similar risks are solvable through institutional design.
  • It is thermal generation that is becoming hard to pay back, through fuel costs, carbon prices and stranded assets.

“Renewables are expensive” is received wisdom about ten years out of date. The question has moved on to how to accelerate deployment and keep the money circulating in local economies.

References and data sources

LCOE and the cost of renewable generation

Japanese tariffs and grid constraints

Battery costs

Fossil fuels, carbon pricing and stranded assets

← Back to list