How Grid-Scale Battery Storage Earns Revenue in Japan: JEPX, Capacity & Balancing Market Stacking

Overview

For an infrastructure fund, IPP, or developer evaluating a grid-scale battery in Japan, the first question is rarely "can it be built?" It is "where does the money come from, and how durable is it?" Japan does not have a single market that pays a battery a clean, predictable spread. Instead, a stand-alone storage asset earns from several distinct markets at once — a practice operators call revenue stacking. Done well, stacking is what turns a battery from a speculative trade into a financeable infrastructure asset. Done naively, it produces a business plan built on assumptions that the 2026 rule changes have already invalidated.

This article walks through the four revenue layers available to a grid-scale battery in Japan — wholesale (JEPX), the capacity market, the Long-Term Decarbonization Auction (LTDA), and the balancing market (需給調整市場) — how they interact, and what changed in 2026. It is written from an operator's seat, not a brochure's. We will be explicit about where revenue is being cut, because the most expensive mistakes in this market come from modeling yesterday's price caps.

The four revenue layers at a glance

A Japanese grid battery can, in principle, draw from all four of the following. The central design decision — made before financial close, not after — is which of them you treat as the anchor and which you treat as upside.

Revenue layer	What it pays for	Pricing unit	Revenue character	2026 status
Wholesale / JEPX	Buying low and selling high across the day (energy arbitrage)	JPY/kWh	Merchant, volatile	Active; new trading platform rollout in 2026
Capacity market (容量市場)	Being available years ahead to meet peak demand	JPY/kW/year	Fixed once cleared, ~4 years forward	Active; batteries cleared at scale
LTDA (長期脱炭素電源オークション)	Fixed-cost recovery for new decarbonized capacity	JPY/kW/year, 20 years	Long-term fixed, CPI-adjusted	Active; Round 3 results published May 2026
Balancing market (需給調整市場)	Holding capacity in reserve to balance the grid second-by-second	JPY/ΔkW per 30 min	Merchant, was high-margin	Price caps falling sharply in 2026

The honest summary up front: in 2024–2025 many battery business plans leaned heavily on the balancing market because its capacity prices were high. In 2026 that lean is being deliberately corrected by the regulator. Any model you review today should treat the balancing market as a thinner, more competitive layer than it was eighteen months ago — and should anchor the asset on something more durable.

Layer 1 — Wholesale energy arbitrage on JEPX

The Japan Electric Power Exchange (JEPX) runs the day-ahead spot market and an intraday market, with prices set for the whole system and for each of the nine regional zones (Hokkaido through Kyushu) in JPY/kWh (JEPX).

The battery logic here is the oldest one in storage: charge when the spot price is low, discharge when it is high, and capture the spread. In Japan that spread is driven heavily by solar. Midday spot prices in high-solar zones are frequently pushed down — sometimes to near zero on sunny shoulder-season days — while evening prices rise as solar drops off and demand peaks. A battery monetizes that daily shape.

What an operator actually watches:

The intraday price shape, not just the daily average. Arbitrage revenue is a function of the gap between the cheapest charging hours and the most expensive discharging hours, net of round-trip efficiency losses. A flat day earns nothing regardless of the average level.
Cycle budget. Every arbitrage cycle consumes warranty-bounded cycle life. The discipline is to dispatch only when the captured spread justifies the marginal degradation cost.
Platform change. JEPX is migrating its spot and intraday systems to a new API-based platform across 2026, which raises the bar on trading-system integration and automation (Volue).

Wholesale arbitrage is real revenue, but it is merchant revenue: volatile, weather-dependent, and impossible to bank a 15-year loan against on its own. It is upside, not anchor.

Layer 2 — The capacity market (容量市場)

The capacity market pays a resource simply for being available to deliver at system peak, several years in advance. Payment is in JPY/kW/year, fixed once the auction clears, and the delivery year sits roughly four years after the auction. For a battery, this is the first genuinely bankable layer: a known annual payment, independent of how volatile spot prices turn out to be.

Batteries have cleared at meaningful scale through Japan's long-term auction track: the FY2025 LTDA round awarded roughly 1.25 GW of battery storage across 19 projects (of which BESS-only cleared ≈575 MW), with a six-hour duration requirement (Energy Storage News; Shulman Advisory). Note that this ≈1.25 GW figure is the LTDA total, not the capacity market (容量市場) main auction — see the next section, which keeps the two distinct. The duration requirement matters: a forward capacity payment is only worth bidding for if the asset's energy capacity (MWh) is sized to satisfy the obligation window.

The capacity market and the LTDA below are related but distinct. The capacity market is the standard, repeating auction for availability. The LTDA is a separate, long-tenor instrument aimed specifically at new decarbonized capacity — and for batteries it is usually the bigger strategic decision.

Layer 3 — The LTDA: a 20-year fixed-cost anchor

The Long-Term Decarbonization Power Source Auction (LTDA) is the closest thing Japan offers to a contract-for-the-asset. A winning project receives a 20-year fixed capacity payment (JPY/kW/year), adjusted for the consumer price index, structured to recover the project's fixed costs (Nishimura & Asahi; White & Case). For a lender, a 20-year CPI-linked revenue line is exactly the kind of anchor that makes project finance possible.

There is a defining catch that every model must build in: the clawback. A winning bidder must return roughly 90% of net profits earned in other markets — wholesale trading, balancing, environmental attributes — to OCCTO (National Law Review). The trade is explicit:

You gain: 20 years of fixed, financeable, inflation-protected revenue, and the elimination of merchant risk.
You give up: ~90% of the merchant upside from the other layers. Your JEPX arbitrage and balancing earnings are largely passed through to OCCTO; you keep about a tenth.

This single rule reframes the whole stacking question. An LTDA-backed battery is not optimizing for merchant upside — it is optimizing for fixed-cost recovery and operational reliability. A battery not in the LTDA keeps 100% of its merchant earnings but carries 100% of the merchant risk. These are two genuinely different business models, and an investor must decide which asset they are underwriting before debating dispatch strategy.

Recent results show the program maturing and tightening. Round 3 (results published May 13, 2026) introduced a six-hour minimum storage duration, removing the earlier 3–6 hour bracket and raising the capital bar for qualifying projects. The decarbonization category was undersubscribed for the first time (~4,261 MW awarded, ~85% of target), while storage categories were oversubscribed against their target, and BESS bids cleared at a notably higher success rate (~46%) than the ~20–24% of earlier rounds (National Law Review; Renewable Energy Institute).

Layer 4 — The balancing market (需給調整市場), and why 2026 changes everything

The balancing market pays resources to hold capacity in reserve so the grid operator can call on it to balance supply and demand in near-real time. Products range from the fastest primary response (一次調整力) through secondary (二次) and tertiary (三次) products, priced as a capacity payment in JPY per ΔkW per 30-minute block. Fast-responding batteries are technically well suited to the premium high-speed products, and through 2024–2025 these products carried high enough capacity prices that they became the headline revenue line in many battery business cases.

That is precisely what is being unwound in 2026. The regulator has concluded that competition in these products was too thin and prices too high, and is bringing the caps down — in stages, with monitoring at the 1-, 2-, 3-, and 6-month marks (e-energy):

Item	Before (through early 2026)	From 13 March 2026	Conditional further path
Price cap	19.51 JPY/ΔkW·30 min	15 JPY/ΔkW·30 min	Staged down to 10, then 7.21 JPY if competition does not improve
Procurement volume	~3σ equivalent	~1σ equivalent	—
Trading timing	—	Day-ahead (前日) trading begins for primary / secondary-1 / tertiary-1 as a combined product	—

Two things are happening at once, and both cut revenue. The price cap on the premium products drops from 19.51 to 15 JPY per ΔkW per 30 minutes on 13 March 2026 (for 14 March delivery), with an explicit path to 10 and then 7.21 if competition does not improve. Separately, the procurement volume the grid operator buys shrinks from a ~3σ reserve to a ~1σ reserve — a smaller pie, not just a lower price per slice. In parallel, a related FY2026 (April) proposal would pull premium products further toward the 7.21 JPY floor that secondary-2 and tertiary-1 products already sit at (Greenberg Traurig).

The operator's takeaway is blunt: any battery model that still relies on 2024–2025 balancing-market capacity prices is overstating revenue. The balancing market remains a legitimate stacking layer — and the move to day-ahead trading rewards operators who can forecast and bid well — but it is now thinner, more competitive, and trending lower. It belongs in the model as managed upside, not as the foundation.

Putting the stack together: anchor plus upside

The practical architecture that survives 2026 looks like this:

Pick the anchor. Either a long-term fixed payment (LTDA, 20-year, CPI-linked, with the 90% merchant clawback) or a capacity-market payment (JPY/kW/year, fixed once cleared). This is the layer your debt is sized against.
Treat JEPX arbitrage and the balancing market as upside — sized conservatively at current and expected (lower) balancing caps, and adjusted for the LTDA clawback if applicable.
Optimize dispatch across layers continuously. At any given 30-minute block, the asset must decide whether its MWh are worth more sitting in balancing reserve, arbitraging on JEPX, or meeting a capacity obligation. That decision is a live optimization problem, not a fixed schedule — and it is where day-to-day operating skill creates or destroys margin.

A note on real assets: stacking is no longer theoretical in Japan. Capacity-market batteries have been commissioned and grid-scale projects are operating commercially across JEPX trading and balancing-market participation — including our own 2 MW / 8.1 MWh commercial battery storage facility in Saitama, which participates in JEPX trading and the balancing market. The lesson from running real assets is consistent with the analysis above: the revenue is real, but it is multi-layered, rule-sensitive, and unforgiving of stale assumptions.

What this means for your investment case

Do not underwrite a single market. A model anchored entirely on the balancing market is an out-of-date model in 2026. A model anchored entirely on JEPX arbitrage is a merchant bet, not infrastructure.
Decide the LTDA question explicitly. Fixed 20-year recovery minus ~90% of merchant upside is a fundamentally different asset from a fully merchant battery. Underwrite one or the other on purpose.
Size MWh to the obligation. Six-hour duration requirements in the capacity market and LTDA Round 3 mean energy capacity, not just power rating, determines eligibility.
Build the rule changes into the base case. 19.51 → 15 → potentially 7.21 JPY, and 3σ → 1σ procurement, are known, dated, public facts. They should already be in your downside.

Where operations meet returns

Every layer above ultimately turns on one capability: dispatching the asset, hour by hour, to the highest-value use across four interacting markets — under rules that are still moving. That is an operating function, not a one-time financial structuring exercise. Forecasting JEPX shape, bidding day-ahead balancing under falling caps, honoring capacity and LTDA obligations, and respecting the battery's cycle budget all have to happen simultaneously, every day, for twenty years.

This is exactly where an operations-and-trading agency (運用代行) earns its place: handling market registration, day-ahead and intraday bidding, balancing-market participation, and cross-market dispatch optimization on the owner's behalf — so the asset's revenue stack is actively managed rather than left to a static plan. If you are evaluating a grid-scale battery in Japan and want the revenue model pressure-tested against the current 2026 rules — or want a partner to run the asset once it is built — that conversation is the natural next step.

Figures in this article are drawn from public regulatory and market sources cited inline. Internal operating data from specific assets is not disclosed.

Sources: