From the 14 March 2026 delivery day (13 March trading day), Japan's balancing market for ΔkW (EPRX) moved to day-ahead trading and 30-minute slots; the procurement volume for primary control reserve (primary) and secondary control reserve ① (secondary ①) went from "3σ-equivalent" to "1σ-equivalent" (the combined product was 1σ-equivalent from the start); and the price cap for primary, secondary ① and the combined product went from ¥19.51 to ¥15/ΔkW·30min. From the 1 September delivery, the cap is ¥10. Since then we often see the explanation that "3σ became 1σ, so procurement volume is one third." Divide out the numbers EPRX publishes, and it is not so.
This page deals only with the demand side: how procurement volume is set, what fell by how much, what did not fall, and what will move it from here. When and where batteries arrive is on 47-3, and where the price stops is on 47-4. No formulas. The order follows what the figures can show.
1. "1σ" is the width of how far you prepare
Balancing capacity (ΔkW) is the "spare" that transmission and distribution operators (TSOs) buy in advance to fill the fluctuation between electricity demand and supply. The fluctuation happens every day, but its size differs from day to day. Small fluctuations are common, large ones rare. The hill in Figure 1 is the distribution of that "common and rare," and the unit for the width of the hill is σ (sigma).
Preparing as far as 3σ means buying the volume that covers 99.9% of the fluctuation above the mean. Preparing as far as 1σ means stopping at the volume that covers 84%. The remaining 16% is handled by means outside the market. "3σ → 1σ" means that this width of preparation was narrowed.
What matters here is that cutting the width to one third does not cut the volume bought to one third. The reason is in the next figure.
2. The volume bought is built in 3 layers
The required volume for primary is the sum of two parts (System Design Working Group (ANRE), 110th meeting, Document 4; secondary ① follows the same thinking, with the contingency portion standing separately).
- Normal-condition portion: the width that prepares for day-to-day demand fluctuation. This is where σ multiplies.
- Contingency portion: the cover for the single largest power plant in that area stopping. It is set by the size of the plant, so moving σ does not change it.
By analogy, the normal-condition portion is an umbrella and the contingency portion is a spare tire. Changing the accuracy of the weather forecast does not change the number of spare tires.
Take the part secured outside the market out of that required volume, and what is left is the procurement volume put to the market. Outside the market means two things: the bilateral contract for pumped storage, under which a TSO contracts directly with a pumped storage plant, and the spare-capacity utilization contract with power sources awarded in the capacity market. In FY2025 what was procured was "the 3σ-equivalent volume less out-of-market balancing capacity," and the deadline for that deduction was set at March 2026 (109th meeting, Document 6).
So "required volume" and "procurement volume" are different numbers. In the next 2 sections we divide both of them out.
3. Required volume is 0.55 — it does not fall to one third
We averaged the required procurement volume tables EPRX publishes (the versions effective at the start of the fiscal year) over April–September and divided FY2026 by FY2025. Nationwide the figures are primary 0.55, secondary ① 0.56, combined 0.78, secondary ② 0.90, tertiary ① 0.77. Even for primary and secondary ①, the products the 1σ change applies to, it stops a little below half. That is because the spare tire — the contingency portion — remains.
By area, Hokkaido, Tohoku and Tokyo are 0.60–0.70, and Chubu and westward 0.44–0.50. In the 3 eastern areas, where the largest power plant is large relative to the size of the grid, the contingency portion is a bigger share, so narrowing the width cuts less — that is how we read it (the breakdown by area is not published, so this is a reading and not a measurement). Not one product × area combination fell to one third, and the lowest is Shikoku's secondary ① at 0.44 (the table by area is Reference table 1 at the end).
4. Procurement volume is 0.62, and the combined product 0.97
The volume the market actually put out fell even less than the required volume. Figure 3 averages the procurement volume in the clearing data (by TSO) over the same April–August window for FY2025 and FY2026.
Table 1 Ratio of procurement volume, FY2026 / FY2025 (April–August average, nationwide)
| Product | FY2025 MW | FY2026 MW | Ratio |
|---|---|---|---|
| Primary online | 2,105 | 1,497 | 0.71 |
| Primary offline | 1,099 | 498 | 0.45 |
| Primary, total | 3,203 | 1,995 | 0.62 |
| Secondary ① | 1,597 | 1,141 | 0.71 |
| Combined (primary–tertiary ①) | 3,830 | 3,700 | 0.97 |
| Tertiary ① | 2,970 | 2,857 | 0.96 |
| Secondary ② | 969 | 1,110 | 1.15 |
| Tertiary ② | 720 | 553 | 0.77 |
Source: our calculation from EPRX trading results (clearing data by area). FY2025 averages the procurement volume of the weekly product (3-hour blocks) and FY2026 that of the 30-minute slots
There are two ways to read it.
The battery seat halved. The primary offline frame is 0.45. The ceiling of this frame is defined as "the 1σ-equivalent value of the normal-condition portion" (110th meeting, Document 4), so narrowing σ came through almost one for one. This is the offline frame where batteries take 98–100% of the awarded volume (→ 47-3), about 500 MW per slot nationwide. Confined to "the battery seat," it is a fact that it fell by more than half.
The size of the market has not changed. The combined product is 0.97, tertiary ① 0.96 and secondary ② 1.15. The vessels of 3,700 MW for combined and 1,500 MW for primary online remain at last year's level. This is where the "one third" story goes most wrong.
By area the shape is the same (Figure 4). The combined product grew year on year in Hokkaido, Chubu and Kyushu, and even in the areas where it fell, Kansai's 0.72 is the floor. There is no one third in the combined product (the numbers by area are in Reference table 2).
← scroll horizontally →
5. Why the combined product did not grow — 266 × 10,000 kW of bilateral pumped storage contracts
In December 2025 the System Design Working Group showed that "unifying at 1σ-equivalent would leave the FY2026 procurement volume outlook 13% lower for primary and secondary ① and 50% higher for the combined product" (109th meeting, Document 6). The measured figures are 0.62 for primary and 0.97 for combined. What made the gap with the outlook is the hatched part of Figure 2 — the part secured outside the market. The FY2026 required-volume table has not been revised once during the year, and the deduction is being made on the procurement volume side rather than on the required volume. Taking "required volume minus procurement volume" area by area gives about 1,400 MW in Tokyo, about 980 MW in Chubu and about 860 MW in Kansai. That is more than the contracted capacity of the bilateral pumped storage contracts, so it reads as though other out-of-market balancing capacity, such as spare-capacity utilization contracts, is being deducted as well.
A footnote in the same document says that "from April 2026, depending on the situation, a further deduction could be considered if necessary." In practice 5 companies concluded bilateral contracts for pumped storage in FY2026, and EPRX gave notice of a revision to the procurement volume each time.
Table 2 Bilateral contracts for pumped storage in FY2026
| Area | Contracted capacity | Contract period | FY2025 actual unit price JPY/ΔkW·h | RC application unit price JPY/ΔkW·h |
|---|---|---|---|---|
| Tokyo | approx. 120 × 10,000 kW | 2026/5/22–2027/3/31 | 0.47 | 2.19 |
| Chubu | up to 61 × 10,000 kW | 2026/4/1–2027/3/31 | 0.67 | 2.25 |
| Kansai | 47 × 10,000 kW | 2026/4/4–2027/3/31 | 0.67 | 3.24 |
| Tohoku | up to 23 × 10,000 kW | 2026/4/1–2027/3/31 | 1.34 | 3.04 |
| Chugoku | approx. 15 × 10,000 kW (new) | 2026/6/8–2027/3/31 | — | 4.4 |
| Hokkaido | not conducted in FY2026 | — | — | — |
Source: Electricity and Gas Market Surveillance Commission (EGC), Expert Committee on System Design and Market Surveillance (制度設計・監視専門会合), 19th meeting (30 March 2026), Documents 5-1 to 5-4; 20th meeting (29 May 2026), Document 5; company announcements. Hokkaido's decision not to proceed is a secondary source
266 × 10,000 kW in total. Tokyo's 120 × 10,000 kW is double last year's 60 × 10,000 kW, which means it locked up more outside the market than the combined product procurement volume of the Tokyo area (1,061 MW) (Figure 5). FY2025 actual unit prices were 0.47–1.34 JPY/ΔkW·h, or 0.2–0.7 on a 30-minute basis, an order of magnitude away from the market cap of ¥10. The way the contract is built — ex-post settlement of the generator's fixed costs and opportunity cost — differs from a market bid, so a simple comparison is not possible, but it is certainly the cheap option for a TSO.
Chugoku Electric Power Transmission & Distribution moved to a new bilateral contract in June 2026 because "after the shift to day-ahead trading, the procurement unit price in the market rose to the same level as the RC application unit price (4.4 JPY/ΔkW·h)" (20th meeting, Document 5). The higher the market, the more procurement escapes outside it. That is why the combined product's procurement volume did not rise 50%.
There is movement the other way as well. Hokkaido is reported to have skipped the bilateral contract for FY2026 and consolidated on market procurement (the primary materials go only as far as "we will judge after seeing market results," at the March and May meetings). Hokkaido alone has "required volume minus procurement volume" of almost zero, which is consistent with there being no deduction. If bidding into the market thickens, bilateral contracts can move down as well. Procurement volume is not one-way.
Tertiary ② alone works on a different mechanism. Its procurement volume is set by multiplying the 1σ-equivalent table by a "procurement reduction coefficient" that reflects the unit price of out-of-market balancing capacity. That coefficient fell, and the effective procurement volume went from about 780 MW to about 590 MW, a factor of 0.76 (our calculation from the 1σ table EPRX publishes and the monthly coefficients).
6. The next step for the price cap, and the conditions for procurement volume to come back
The price cap has come down ¥19.51 → ¥15 (from 3/14 delivery) → ¥10 (from 9/1 delivery), and ¥7.21 is shown as the next step. The trigger is "if no improvement is seen in the state of competition in the market." The indicators are bid volume against procurement volume, the distribution of bid prices and the price level of spare capacity, and the timing is the record over "1 month, 2 months, 3 months, 6 months and so on after day-ahead trading starts" (110th meeting, Document 4). The document of the 4th Power Stable Supply Working Group (14 July), which decided the move to ¥10, gives as its reason that clearings above ¥14 were about 3% of volume but about 17% of procurement cost. The materials write "6 months and so on," a rough guide, but 6 months from the start of day-ahead trading falls in mid-September. The 5th WG on 7/29 and the 6th on 8/31 had no balancing-market item on the agenda.
The same set of documents contains a clause pointing the other way: "if it is confirmed that sufficient competition is working in the market, the procurement volume will be increased" (110th meeting, Document 4). The 1σ-equivalent level is provisional, and the procedure for increasing it after looking at the state of bidding is written down. This is not a design that only shrinks. Bidding growing until competition is confirmed is itself the condition for procurement volume to come back. We count the state of bidding on 47-3.
The portion above 1σ and up to 3σ is handled by spare-capacity utilization contracts with power sources awarded in the capacity market as having a balancing function. Those contracts settle on kWh only, with no payment for ΔkW (Transmission & Distribution Grid Council). The accurate way to put it is not that the demand disappeared, but that it moved onto a route that carries no ΔkW payment.
7. The 4 things that will move procurement volume from here
Demand growth. Data centers and semiconductor plants, counted individually, are 83 × 10,000 kW of national peak demand in FY2026 → 420 × 10,000 kW in FY2030 → 762 × 10,000 kW in FY2035, or 0.5% → 4.6% of the national total (OCCTO demand forecast 2026). The normal-condition portion is proportional to the size of demand fluctuation, so the effect on procurement volume stays in the order of a few percent. Where data centers bite is JEPX rather than the balancing market (→ 47-4).
Solar. Forecast error goes into the required volume for tertiary ②, and short-cycle fluctuation into the normal-condition portion of primary and secondary ①. In the Supply Plan, solar goes from 8,359 × 10,000 kW in FY2026 to 9,642 × 10,000 kW in FY2030, and this one pushes procurement volume up.
Nuclear. The contingency portion is set by "the largest power plant." Kashiwazaki-Kariwa Unit 6 (135.6 × 10,000 kW) resumed generation and transmission in February 2026 and entered commercial operation on April 16, which works to push Tokyo's primary required volume up. At the same time, when nuclear runs, thermal stops at night, which works to open up the online seats where thermal sits (→ 47-3). Nuclear itself does not bid into the balancing market.
Interconnectors. Tohoku–Tokyo goes from 573 to 1,028 × 10,000 kW in 2027/11 (in stages, depending on how generation develops); the frequency converter between Tokyo and Chubu reaches 300 × 10,000 kW by FY2028; and Kanmon reaches 600 × 10,000 kW in June 2030 (OCCTO). The balancing market already clears across areas, and where there is room on the interconnector a power source outside an area can clear against another area's procurement volume. Secondary ① alone starts in FY2027. When the lines are congested, the market splits (EPRX Trading Rules, Article 32(3)). An area's procurement volume is not a fixed vessel but one that gets tied together along the interconnector calendar.
Beyond that lies the co-optimized market (kWh and ΔkW cleared together, first half of the 2030s). The concepts of a price cap and a procurement volume are themselves under review, in the direction of being replaced by optimized clearing based on lost profit.
Summary
- Required volume 0.55, procurement volume 0.62, combined 0.97. "One third" fits no product.
- What fell is the battery seat (primary offline, 0.45, about 500 MW). The size of the market (combined 3,700 MW, primary online 1,500 MW) has not changed.
- What kept procurement volume from coming back is 266 × 10,000 kW of bilateral pumped storage contracts. The higher the market, the more procurement goes outside it; the thicker the bidding, the more it comes back.
- The condition for procurement volume to come back is confirmation of competition. More bidding is that condition itself.
When and where batteries arrive is on 47-3, and where the price stops is on 47-4.
Reference tables
Reference table 1 Ratio of required procurement volume, FY2026 / FY2025 (April–September average)
| Area | Primary | Secondary ① | Combined | Secondary ② | Tertiary ① |
|---|---|---|---|---|---|
| Hokkaido | 0.60 | 0.63 | 0.88 | 0.70 | 0.91 |
| Tohoku | 0.61 | 0.62 | 0.85 | 0.88 | 0.83 |
| Tokyo | 0.68 | 0.70 | 0.82 | 0.92 | 0.81 |
| Chubu | 0.50 | 0.48 | 0.67 | 0.86 | 0.67 |
| Hokuriku | 0.46 | 0.48 | 0.76 | 0.92 | 0.72 |
| Kansai | 0.48 | 0.49 | 0.72 | 0.91 | 0.70 |
| Chugoku | 0.47 | 0.47 | 0.81 | 0.98 | 0.78 |
| Shikoku | 0.46 | 0.44 | 0.63 | 0.86 | 0.62 |
| Kyushu | 0.48 | 0.47 | 0.82 | 0.98 | 0.87 |
| Nationwide | 0.55 | 0.56 | 0.78 | 0.90 | 0.77 |
Source: our calculation from EPRX required procurement volumes (the FY2025 and FY2026 tables effective at the start of the fiscal year)
Reference table 2 Ratio of procurement volume, FY2026 / FY2025 (April–August average, by area)
| Area | Primary on | Primary off | Secondary ① | Combined | Tertiary ① | Secondary ② |
|---|---|---|---|---|---|---|
| Hokkaido | 0.81 | 0.37 | 0.87 | 1.24 | 1.33 | 0.96 |
| Tohoku | 0.75 | 0.70 | 0.96 | 1.12 | 1.00 | 1.30 |
| Tokyo | 0.86 | 0.42 | 0.74 | 0.87 | 0.84 | 0.70 |
| Chubu | 0.93 | 0.47 | 0.76 | 1.33 | 1.43 | 2.58 |
| Hokuriku | 0.59 | 0.41 | 0.79 | 0.97 | 0.97 | 0.78 |
| Kansai | 0.51 | 0.43 | 0.48 | 0.72 | 0.69 | 0.82 |
| Chugoku | 0.55 | 0.43 | 0.55 | 0.94 | 0.98 | 1.00 |
| Shikoku | 0.66 | 0.63 | 0.65 | 0.73 | 0.74 | 0.90 |
| Kyushu | 0.66 | 0.43 | 0.81 | 1.20 | 1.52 | 1.86 |
| Nationwide | 0.71 | 0.45 | 0.71 | 0.97 | 0.96 | 1.15 |
Source: our calculation from EPRX trading results (clearing data by area)
Notes: Reference table 1 averages the months carried in the required-volume table effective at the start of the fiscal year (April–September); Table 1 and Reference table 2 average the procurement volume in the clearing data (by TSO) over April–August. August 2026 includes preliminary figures. The 84% and 99.9% in Figure 1 are one-sided values of the normal distribution; actual demand fluctuation is not normally distributed, so each transmission and distribution operator calculates it as an "equivalent volume."
Sources
- EPRX, required balancing capacity volumes (調整力必要量) https://www.eprx.or.jp/information/requirements_table.php
- EPRX, trading results (clearing data by area) (取引実績（エリア別約定データ）) https://www.eprx.or.jp/information/results.php
- EPRX, notices of procurement volume revisions on the conclusion of bilateral pumped storage contracts (揚水随意契約締結に伴う募集量見直し告知): Chubu and Tohoku (31 March 2026) https://www.eprx.or.jp/j_information/docs/news_202603311700.pdf / Kansai (1 April) https://www.eprx.or.jp/j_information/docs/news_202604011500.pdf / Tokyo (22 May) https://www.eprx.or.jp/j_information/docs/news_202605221500.pdf / Chugoku (8 June) https://www.eprx.or.jp/j_information/docs/news_202606081600.pdf
- EPRX, Balancing Market Trading Rules ver.18 (需給調整市場取引規程 ver.18), in force 1 July 2026 https://www.eprx.or.jp/outline/docs/kitei_ver.18_260701.pdf
- Agency for Natural Resources and Energy (ANRE), System Design Working Group, 109th meeting, Document 6, "On the balancing market" (需給調整市場について) (12 December 2025) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/jisedai_kiban/system_review/pdf/109_06_00.pdf
- Same, 110th meeting, Document 4, "On the balancing market" (需給調整市場について) (23 January 2026) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/jisedai_kiban/system_review/pdf/110_04_00.pdf
- ANRE, Power Stable Supply Working Group, 4th meeting, Document 6, "On the balancing market" (需給調整市場について) (14 July 2026) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/jisedai_kiban/stable_power_supply_wg/pdf/004_06_00.pdf / 5th meeting agenda (29 July) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/jisedai_kiban/stable_power_supply_wg/pdf/005_01_00.pdf / 6th meeting agenda (31 August) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/jisedai_kiban/stable_power_supply_wg/pdf/006_01_00.pdf
- Electricity and Gas Market Surveillance Commission (EGC), Expert Committee on System Design and Market Surveillance (制度設計・監視専門会合), 19th meeting (30 March 2026), Document 5-1 Chubu https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/019_05_01.pdf / Document 5-2 Tohoku https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/019_05_02.pdf / Document 5-3 Kansai https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/019_05_03.pdf / Document 5-4 Tokyo https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/019_05_04.pdf / Document 5-5 Hokkaido https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/019_05_05.pdf / 20th meeting (29 May), Document 5 Chugoku https://www.egc.meti.go.jp/activity/emsc_systemsurveillance/pdf/020_05_00.pdf
- Company announcements: TEPCO Power Grid (22 May 2026) https://www.tepco.co.jp/pg/company/press-information/information/2026/pdf/26x1301.pdf / Kansai Transmission and Distribution (1 April) https://www.kansai-td.co.jp/corporate/information/2026/pdf/20260401_2_01.pdf / Chubu Electric Power Grid (31 March) https://powergrid.chuden.co.jp/news/topics/1217576_3288.html / Tohoku Electric Power Network (31 March) https://nw.tohoku-epco.co.jp/information/1248501_2390.html / Chugoku Electric Power Transmission & Distribution (8 June) https://www.energia.co.jp/nw/information/news/16498.html
- Hokkaido Electric Power Network's decision not to conclude a bilateral contract in FY2026 (secondary source): Enehub (29 July 2026) https://enehub.jp/news/北電nw_2026年度_揚水随契を実施せず/
- Transmission & Distribution Grid Council, "Outline of and application for the spare-capacity utilization contract" (余力活用に関する契約の概要と申込) https://www.tdgc.jp/yoryoku/outline/
- Organization for Cross-regional Coordination of Transmission Operators (OCCTO), 117th Committee on Balancing Capacity and Supply-Demand Balance Evaluation, Document 4 (16 March 2026) https://www.occto.or.jp/assets/chousei_117_04.pdf
- OCCTO, "Interconnector operational capacity, FY2026–2035," Attachment 1 (2026〜2035年度の連系線の運用容量) (1 March 2026) https://www.occto.or.jp/assets/renkeisenriyou/oshirase/2025/files/oshirase_1_2026-2035_unyouyouryou.pdf
- OCCTO, "Demand forecast for the nation and for each supply area (FY2026)" (全国及び供給区域ごとの需要想定（2026年度）) (21 January 2026) https://www.occto.or.jp/assets/news/juyousoutei/260121_juyousoutei_r1.pdf
- OCCTO, "Compilation of the FY2026 Supply Plan" (2026年度供給計画の取りまとめ) (30 March 2026) https://www.occto.or.jp/assets/various/kyoukei/torimatome/260330_kyokyukeikaku_torimatome/260330_kyokei_torimatome.pdf
- ANRE, "Recent developments in nuclear policy" (原子力政策の最近の動向) (31 March 2026) https://www.meti.go.jp/shingikai/enecho/denryoku_gas/genshiryoku/pdf/048_01_00.pdf
- Study Group on the Design of a Co-optimized Market, Second Interim Report (同時市場の在り方等に関する検討会 第二次中間取りまとめ) (15 October 2025) https://www.meti.go.jp/shingikai/energy_environment/doji_shijo_kento/pdf/20251015_1.pdf
Energy Business Consultation
Project-specific information not covered in our articles
will be disclosed after NDA execution following your inquiry.