The main source of income for a battery storage plant is the balancing market. It is where the transmission and distribution operators buy, a day ahead, “capacity that can be delivered at any moment”, so that gaps between supply and demand can be closed instantly. A 2 MW / 8 MWh-class plant bids into a frame of this market called primary offline. Since 1 September 2026 those bids have had a ceiling: a ¥10 cap (per kW per 30 minutes). That ceiling belongs to the market, and it sits at the same height for everyone.
This series is about the second ceiling below it. The government's Balancing Market Guideline (需給調整市場ガイドライン, revised 13 March 2026) says it is desirable to keep the bid price within “opportunity cost + fixed-cost adder + fee”. That is not an obligation. But the same revision added that unfairly earning profits through unreasonably high bids can be subject to a business improvement order.
We put only published figures into this formula and counted for a 2 MW / 8 MWh plant. Set in Tokyo, the adder comes to ¥1.4 in year 1 of operation, ¥0.96 in year 2 and ¥0.33 in year 18. That is an order of magnitude below the ¥10 cap. And this ceiling does not reach thermal or other resources — it reaches only batteries.
ΔkW (delta kilowatt) — the amount of capacity held open “so it can be delivered at any moment”. In this market you are paid for the promise to hold it open, separately from any electricity actually delivered. The unit is “¥/ΔkW·30min”.
Primary offline — a frame you can join without laying a dedicated line. It covers high-voltage plants of at least 1 MW and under 10 MW; a 2 MW / 8 MWh plant bids here.
Fixed-cost adder (一定額) — the amount the guideline accepts as “you may add up to this much to recover your fixed costs”. It is the subject of this series.
Opportunity cost (逸失利益) — what you fail to earn on the wholesale market (JEPX) because you hold capacity open for the balancing market.
Expected bid volume (想定応札量) — the volume you can bid in a year. A battery can bid into almost all of the 48 slots in a day, so this volume is large.
← Scroll sideways →
How the adder is set
Setting the adder takes only a subtraction and a division. From a year's fixed costs (depreciation, repairs and so on), take away the revenue available from other markets (capacity-market revenue and so on), and divide the rest by the volume you can bid in a year. The result is the adder you may put on top, per slot.
← Scroll sideways →
Two points matter here. First, even if you are not in the capacity market, the revenue you would have earned there is deducted (the only exceptions are “a year in which the plant was too new to bid” and “you bid but were not awarded”). Second, the divisor is not “the volume actually awarded” but “the volume you can bid”. A battery can bid into most slots of the day, so the divisor is large and the adder comes out small.
Why it reaches only batteries
Thermal and other resources have large fixed costs, and stop for inspections and the like, so the volume they can bid is small. Their adder was therefore large, and their ceiling sat far above market prices. In fact, of the 34 cases in which the adder was settled through consultation with the secretariat of the Electricity and Gas Market Surveillance Commission (EGC) in FY2024 and FY2025, 28 were resources other than batteries and battery VPPs, and the Commission judged the agreed amounts to be “at a high price level at which clearing is practically difficult” (50-5 Monitoring). They were too high to bite.
Batteries are the opposite. Their fixed costs are small and they can bid in almost every slot. Set from published figures, their adder is around ¥1. Meanwhile, actual primary offline prices in September 2026 were ¥2.74–3.69 in the six areas of Chubu, Hokuriku, Kansai, Chugoku, Shikoku and Kyushu, ¥9.51 in Tohoku and ¥8.58 in Tokyo (50-4 Market). The gap between the market price and the adder has to be explained by opportunity cost. It is batteries that are bidding close to this ceiling.
← Scroll sideways →
Why now
FY2026 is the first year in which operators calculate the adder themselves. Up to FY2025, large operators settled their adder through consultation with the EGC secretariat (the “Type B consultation”). That has ended, and operators now keep the basis of their own calculation. At the June 2026 expert committee, the secretariat asked operators that had won at high prices in April–June to revise their calculations. On top of that came the ¥10 cap on 1 September.
In September, with the cap lowered, primary offline prices split in two by area. Tohoku at ¥9.51 and Tokyo at ¥8.58 stayed just under the cap; the other seven areas are in the ¥1–3 range. The closer a bid stays to the cap, the more it will be asked whether the formula can explain it. Final figures for September are due by the end of October.
How to read this series
This overview carries only the conclusions. The evidence is split across four pages. The order is 50-2 → 50-5, but you can start from whichever question interests you.
- One line of formula. The adder is ¥0.33 or an amount calculated from fixed costs
- What goes into fixed costs, and what does not
- 8 points the text leaves open
- ¥1.411 in year 1, ¥0.959 in year 2 (Tokyo)
- 9 areas; changing construction cost or the share you can bid
- Blanks are for your own figures
- Primary offline nationally: ¥11.07 → ¥5.81
- Only Tohoku and Tokyo stayed under the cap
- Almost the same price by night and by day
- From report requests to consultation to ex-post monitoring
- June 2026: high-price calculations sent back
- How overseas rules differ
What you often hear, and what the documents say
- “The formula was created by this revision.” — The formula that takes other-market revenue away from fixed costs and divides was already in the 2021 original draft. What is new is that the divisor became “the volume you can bid”, and that the text now lists which items go into fixed costs and which do not. → 50-2 Rules
- “The adder runs to several yen.” — From published figures it is in the ¥1 range even in year 1, and below ¥1 from year 2. Items with no published unit price, such as personnel and aggregator costs, are not included; those are for you to add. → 50-3 Calculation
- “The aggregator's success fee counts as a fixed cost.” — The text says it is included “after the fixed costs have been identified”. You do not add the commission as it stands; you take the fixed costs, such as personnel and system costs, out of the breakdown and include those. → 50-2 Rules
- “The biggest drop comes when depreciation ends.” — The biggest step comes in the year deemed capacity-market revenue starts to be deducted (year 2 in our estimate). That is because what you would have earned in the capacity market is deducted even if you are not in it. → 50-3 Calculation
What is not settled
Even read to the end, the text leaves things open: how to count a battery's opportunity cost; whether a “resource” is the aggregator's bidding unit or a single plant; whether depreciation follows the accounts or the tax books; how to judge that “fixed costs have been recovered”; and so on. In the body we mark these “Unsettled” and list whom to ask (at the end of 50-2 Rules). How immediate depreciation is treated is one of them; the incentive itself is covered in COLUMN 45.
How the estimate is set up
The adder figures assume a high-voltage 2 MW / 8 MWh plant bidding into primary offline through an aggregator. We included only construction cost (depreciation), repairs, the generator-side charge, fixed asset tax and deemed capacity-market revenue. Every unit price is one published by the government, OCCTO or the transmission and distribution operators; none is our own.
Personnel costs, O&M outsourcing, insurance, land rent and the part of aggregator costs identifiable as fixed are left blank, because no unit price is published for them. Adding them raises the adder by that much — about ¥0.32 for every ¥10 million a year. So read these figures not as the height of the ceiling itself but as a guide to its lower end.
For owners, buyers and sellers
- Owners. Settling the amount through consultation ended with FY2025; from FY2026 you calculate it yourself and keep the basis. The breakdown of fixed costs, how you count the volume you can bid, and a record of whether you could bid into the capacity market: keep these three at hand, and you will not be stuck when asked to explain.
- Buyers. For the ΔkW price in your model, the safe choice is not the ¥10 cap but the lower of “the actual price in your area” and “the height the formula can explain”. Actual September prices ranged from the ¥1s to the ¥9s depending on the area.
- Sellers. When buyers build this ceiling into their models, they set the ΔkW price below the cap. A record of capacity-market bids and a breakdown of aggregator costs let you show the height of the ceiling with evidence.
Every number was counted by us from published materials of EPRX, the EGC, the Organization for Cross-regional Coordination of Transmission Operators (OCCTO) and others, and the calculation scripts are public too. Sources are listed at the foot of each page.