The highest-earning area is
the one most easily cannibalized

A grid battery's revenue is decided by "which contract you run it under" — that is what we have written up to the previous column. There is a second variable of equal weight: which area you place it in.

Let us put the answer first. Built on the realized figures of the last 12 months, this reference asset earns roughly ¥290-300M/yr (median) in the five areas where the balancing-market door stayed open all year — Kyushu, Tokyo, Chubu, Hokuriku, Kansai, with Kyushu leading at about ¥295M. The breakdown is about 80% balancing market, with JEPX and the capacity market together about 20% — an owner's P&L begins with balancing, so this piece begins there too. And Hokkaido and Tohoku, No. 1 and No. 2 in JEPX, cannot join those five. What was blocking the door was pumped-hydro bilateral contracts — the arbitrage map and the ΔkW map do not overlap. We fix the three markets in order of revenue size, and show buyers "which area to buy" and owners "how to run it," in real numbers, from both the seasonal and the regulatory angle.

01 — Premises and how to read them

The JEPX measurements use a single calculation spec. One cycle per day, perfect foresight: discharge uses each day's eight highest-price slots (2 MW x 0.5 h x 8 = 8,000 kWh), charging draws 9,412 kWh from the lowest-price slots up (8,000 / 0.85; the nine lowest slots plus a pro-rated tenth). Daily gross margin = discharge revenue minus charging cost. Because foresight is perfect, this value is a theoretical upper bound — we call it "the ceiling". In real operation it is eroded by forecast error, availability and aggregator fees, so we always show the ceiling and realized revenue separately (a level of 60-80% of the ceiling is often cited, but it is not an established figure in the measured data, so we do not assert it).

The balancing market has a break partway through these 12 months. Procurement of primary through tertiary-① moved from weekly, 3-hour blocks to a day-ahead, 30-minute product, and at the same time the ΔkW price cap was cut from ¥19.51 to ¥15.00 per ΔkW per slot — both from deliveries on 14 March 2026 (confirmed in the rule text the balancing platform published on 5 February 2026). This piece tallies the old-regime period (1 Jul 2025 - 13 Mar 2026, 256 days) and the new-regime period (14 Mar - 30 Jun 2026, 109 days) separately, and constructs no average that straddles the break.

02 — The balancing market: the revenue body sits here

In order of revenue size, we begin with the biggest earner. First, the post-break terms. From deliveries on 14 March 2026, the cap for composite, primary and secondary-① is ¥15.00; for secondary-② and tertiary-① it is ¥7.21 (unchanged); and tertiary-② has no cap. The trading fee doubled from ¥0.03 to ¥0.06 per ΔkW per slot for actual delivery from 1 April 2026 (what changed on 14 March was the products, caps and procurement volumes; 1 April was the fee only), and the procurement volume for primary and secondary-① was cut from a 3σ to a 1σ equivalent (about 13% down; the band from just above 1σ to 3σ is covered by the reserve-utilization contract in the capacity market). The new regime begins with all three levers — price, volume and cost — tightened at once.

Unit: ¥ per ΔkW per 30-min slot. Shortfall rate = the share of procurement volume left unmet (buyer-side unmet, not the sellers' loss rate). Sources: balancing platform, FY2025 trading results (18 Jun 2026); ANRE Stable Electricity Supply WG, 1st meeting, Doc 8 (13 May 2026, the 28 days after the shift to day-ahead trading). Old and new cannot be cross-averaged because their calculation periods and definitions differ. Full-period confirmed figures for the new regime are unpublished ❓.

Still, applying that average price straight to a battery misreads the market, because of the market's tilt. Primary, secondary-① and composite are markets in procurement shortfall — the shortfall is buyer-side non-fulfilment, and in principle every bid meeting the requirements clears. So a battery's real price is not the market average (which includes thermal and pumped hydro) but the battery's realized winning price: a monthly ¥8.8-13.5 per ΔkW per slot (balancing-platform data by generation type; primary and composite at the same level). Building 12 months on this, a JEPX-combined pattern (bidding 30 slots, excluding the 18 charge/discharge slots) gives about ¥97,000-148,000 per kW per year — which for a 2 MW unit is about ¥190-300M/yr (median ~¥240M). Going all-in on balancing (48 slots) gives about ¥155,000-237,000 per kW per year = ¥310-470M/yr. For reference, building the same 30 slots at the composite market average of ¥2.83 yields about ¥62M/yr — this roughly fourfold gap is the distance between "the regulator's statistics" and "the operator's lived experience."

We disclose the deduction side up front too. What can be pinned down is small — the balancing platform's fee of ¥0.06 per ΔkW per slot (which the guidelines permit passing through into bid prices), the JEPX spot fee of ¥0.03 per kWh, and wheeling charged only on storage losses under the battery exemption. In annual terms none of these move the displayed unit. What does move it is the one undisclosed sheet, the aggregator fee. The only published rate is the 5% for RE100 power; applying the 10-30% that is said to prevail (secondary information ❓) to the median ~¥240M/yr means ¥24-73M/yr goes to the delegatee. Before you debate the digits of the price cap, this fee negotiation is what moves the revenue.

It squares with an external check, too. Yano Research Institute puts the storage-business market size (on an operator-revenue basis) at ¥45 billion for FY2024; dividing by the roughly 250 MW of grid-connected capacity at the same time gives about ¥180,000 per kW per year — the equivalent of ¥360M/yr on a 2 MW basis. It lands in the same order of magnitude as the all-in realized range (¥310-470M/yr) (a rough consistency check that carries a timing mismatch of year and capacity ⚠️). What explains a rush of connection applications to about 3.9x in a year, to 24.31 GW, is this figure — not the ¥62M of the composite average.

How to read the area differences. Area-by-product clearing prices exist as numbers in the daily confirmed data, but no annual table lets you survey a full year, and even in aggregated annual results and council documents the per-area prices stop at graphs (see Section 13 ❓). Since the area differences in price cannot be pulled from a table, what matters in practice is whether the bid clears — whether the door is open. Lining up the primary-reserve shortfall rate across all nine areas for the 28 days after the shift: Hokuriku 75.8%, Tokyo 70.3%, Chubu 66.1%, Kansai 49.0%, Kyushu 42.4%, Shikoku 24.3%, Chugoku 22.8%, and Tohoku 0.1%, Hokkaido 0.0%. The two zeros have a mechanism behind them. In both areas, pumped-hydro bilateral contracts (out-of-market procurement) filled the requirement, effectively closing the market's door (the procurement-volume-deduction structure is spelled out in the 109th Institutional Design Subcommittee, Doc 6; the deduction ran to March 2026). Those bilateral contracts ended in Hokkaido and Tokyo at end-March 2026 — Hokkaido's ΔkW door opens from this fiscal year, and Tokyo's already-large opening widened further. Only in Tohoku do the bilateral contracts continue, and the door stays shut. Meanwhile the old regime's annual confirmed results noted procurement shortfalls of 40-90% or more in all areas except Hokkaido, Chugoku and Shikoku, so old-regime Tohoku was on the shortfall (clearing) side. At the break, Tohoku's door shut and Chugoku's and Shikoku's opened — the realized table in Section 03 pro-rates exactly this period difference (the treatment of bilateral contracts relies on the framing in Doc 8 and related material ❓).

We write the regulatory outlook in stages. The ¥15 cap, the 1σ volume and the ¥0.06 fee are in force. The further cut to ¥10 and ¥7.21, however, is conditional on "no improvement in competition," and sits at the stage of decided in principle — not yet promulgated or in force. The decision framework is a review of results at 1, 2, 3 and 6 months after day-ahead trading began, on three axes: whether bid concentration near the cap has cleared, whether shortfall rates have improved, and whether any operators have exited. The one-month assessment held that improvement was seen but not fully resolved (Stable Supply WG, Doc 8), deferring the next stage, and it is expected to be judged in the second half of FY2026 after the summer peak (Jul-Aug). Each step down shrinks ΔkW revenue roughly proportionally. For a P&L heavily dependent on balancing, this single variable makes the IRR a different animal.

03 — Nine areas x three markets, on one page

We stack the three markets on one page, largest first. The engine = the balancing market (the realized figures of Section 02); the core = JEPX arbitrage (measured in Section 04); the floor = the capacity market (fixed in Section 05 as the FY2029 clearing price x derating coefficient). Since no annual table of the balancing market's per-area prices is published (see Sections 02 and 13), we treat area differences not as price but as the opening and closing of the door (bilateral contracts and shortfall rates).

Taking Section 02's realized price (the battery's realized winning rate of ¥8.8-13.5 per ΔkW per slot) and running 12 months in a JEPX-combined pattern (30 slots), the add-on is about ¥97,000-148,000 per kW per year (median ~¥122,000 = about ¥240M/yr for a 2 MW unit). We stack this on each area's floor + core (fixed in Sections 04 and 05) only for the days the door was open — full for the five areas open all year; old-regime only for Tohoku (256/365 days); new-regime only for Chugoku and Shikoku (109/365 days); and zero for Hokkaido (possibly opening from FY2026 as the bilateral contracts end).

Unit: ¥ per kW per year (nominal kW). Realized = battery realized price ¥8.8-13.5 (median ¥11.2) x 30 slots x days the door was open (range: about ¥97,000-148,000/kW/yr for the five year-round areas; the total including floor + core is roughly ¥122,000-173,000 = about ¥240-350M/yr for a 2 MW unit). Door open/closed uses, for the old regime, the annual confirmed-results note (40-90% shortfall in all but Hokkaido, Chugoku and Shikoku), and for the new regime the 28-day post-shift shortfall rates (Section 02). The balancing-platform fee deduction (¥0.03 to ¥0.06 per ΔkW per slot) is about ¥400/yr, below the displayed unit. The aggregator fee is pre-deduction (see Section 02).

What is left after the engine is cut by the rules? We take out just the two layers, floor + core.

Floor = FY2029 per-area clearing price x the 4-hour battery derating coefficient (annual average, estimate 🚩; see Section 05). Core = the measured ceiling from Section 04. Both per nominal 2,000 kW. The balancing-market add-on is not included.

04 — JEPX: where you can earn was rewritten this year

The second pillar, JEPX. This is the layer where area differences show most sharply, and the map of power shifted over the last 12 months. Below are the reference asset's arbitrage ceiling (¥/kW/yr) and its year-on-year change across the nine areas.

Source: JEPX spot-market result CSVs (FY2024-FY2026), tallied by ScienceX under the Section 01 spec. Perfect-foresight theoretical upper bound (the ceiling). Prior-year window = 2024.07-2025.06.

There are two ways to read it. By level, Kyushu is still third nationwide. Solar surplus sinks midday prices, its charging cost is the cheapest in the country, and that structure is intact — "Kyushu can't earn anymore" would be an overstatement. But by direction, it is the only one of the nine clearly shrinking. Hokuriku, Kansai, Chugoku and Shikoku are also down 3-5%, and the only areas growing are the three eastern ones. Here we correct our own earlier premise: up to the previous column we saw "only Kyushu shrinking," but recomputing on the latest window, all five western areas had slipped to small negatives together. The contraction is not Kyushu alone but a western-Japan-wide trend, with Kyushu its largest value. The main driver of the rise is the spring 2026 price surge in eastern Japan (detailed in Section 07); the structural driver of Kyushu's contraction is treated in Section 09.

Even for the same "ceiling," how you capture it differs by area. We counted the slots where the area price diverged from the system price, split into the high side (in-area is expensive = import constraint) and the low side (in-area is cheap = surplus being shut out). This direction-split tally does not exist in a primary source, so we built it by matching JEPX area and system prices across every slot.

Source: JEPX result CSVs, tallied by ScienceX (2025.07-2026.06; |area - system| > ¥0.009 used as a proxy for separation, counted by direction).

Tokyo is on the high side in 85% of slots — it even charges on the high side. Its spread cannot be made from separation, only from the intraday shape (the midday dip and the evening ramp in spring). The opposites — Shikoku, Kyushu and Chugoku — are 70-80% low-side, with far more slots sinking 5 yen or more (Shikoku 4,584, Kyushu 2,464), while no ceiling forms — the pick-at-the-floor pattern. And only Hokkaido and Tohoku hold both a deep floor and a high ceiling in-area. The mechanism of the leadership change is concentrated in this one table. Hokkaido went further: while the Kitahon interconnector was down after the 8 December 2025 earthquake off eastern Aomori (restored 31 January 2026), its high-side separation jumped to 79.8% (68.6% in normal times) — direct evidence of how thin the interconnector is.

05 — The capacity market: "¥15,112" is nominal

We fix the floor. The FY2029 main auction (published 20 January 2026, corrected 23 January) cleared at about ¥2,209.4 billion in total (+19% YoY), a record, with an all-in average unit price of about ¥13,303/kW after transitional measures; for the first time excluding the first auction, every area's clearing price exceeded the reference price (Net CONE: ¥10,075/kW). Only Kyushu sat strictly at the cap (1.5x the reference = ¥15,112.5, truncated to ¥15,112/kW); Tohoku and Tokyo cleared at ¥15,111/kW — one yen below the cap. Several trade outlets wrote that "three areas hit the cap of ¥15,112," but that is off the OCCTO original by one yen. The practical significance is nil, yet we keep it as an example of how, routed through secondary information, the numbers in the original quietly break. Laid out by area and year it is as follows; existing owners can book delivery revenue with certainty in the FY2026-28 columns.

Unit: ¥/kW (area price). ※ = multi-price-method areas. FY2029 has no multi-price. Sources: OCCTO clearing-result PDFs by year; FY2029 derating-coefficient table (published 31 July 2025). Actual receipt = clearing price x (2,000 kW x annual-average coefficient), an estimate 🚩 — formally it is built up from monthly coefficients.

The two right-hand columns are the subject of this section. When a battery registers as a firm resource (expected capacity 1,000 kW or more, at least one continuous 3-hour discharge per day), the contracted capacity is set not by nameplate but by expected capacity = installed capacity x derating coefficient. The coefficient table is the same one used for pure pumped hydro (area x month x dischargeable hours), and even a 4-hour unit is credited at only about 56-90% depending on area and month. Tokyo in particular is credited at only about 70.8% on an annual average. As a result, Tokyo, with the highest clearing price (¥15,111), has an actual receipt of about ¥21.39M — the same as Chugoku at a unit price of ¥12,388. First place in actual receipt is Tohoku, high on both unit price and coefficient. The nominal calculation "¥15,112 x 2,000 kW = ¥30.22M" is a ceiling value and cannot be used for decisions.

From a new-entrant angle, the transitional measures (a deduction for resources built before end-FY2010) end with the FY2029 auction, and new-build batteries are outside the age-deduction scope to begin with. Since the Long-term Decarbonization Auction unified the battery requirement from its third round to 30 MW or more installed and 6 hours or more of continuous discharge, a 2 MW, 4-hour unit is out of scope — the entrance to capacity revenue narrows to the main auction alone. Its next round, the FY2030-delivery main auction, is the first entry opportunity for "buyers to come," on the following timeline.

The other entrance is succession of an already-cleared project. Under Article 26 of the capacity-securing contract terms, with OCCTO's consent the contractual position can be transferred, and future capacity revenue and the delivery year carry over as is — the state of requirement fulfillment carries over too, which is a due-diligence checkpoint. In the project transfers we handle, the remaining contract years, the cleared price, and whether a reserve-utilization contract exists sit at the center of the price rationale. A new-build bid (autumn 2026) and succession of an existing one — these are the buyer's two entrances.

06 — For buyers to come: strategy changes the area

Overlay Sections 02-05, and the structure of the answer is this. For the five areas whose door stayed open all year, the realized snapshot is near-level (¥290-300M/yr for a 2 MW unit, median). So "which area to buy" is decided by three things: the opening and closing of the door (bilateral contracts and shortfall rates), what is left after the engine is cut (floor + core), and which operating strategy you use to capture it.

07 — Spring has become the biggest harvest

Now to those who already own. Seasonal common sense was rewritten in these 12 months. Cut the reference asset's gross margin by season, and in all nine areas spring (Mar-May) is the annual maximum and winter (Dec-Feb) the annual minimum.

Unit: ¥/kW per season (reference asset, perfect-foresight ceiling). Source: ScienceX, from JEPX published data.

The old rule "earn on winter evenings" does not hold, at least in this window. We check with high-price events. Tokyo's slots above ¥30/kWh number 279 in spring, 154 in summer, 35 in autumn, 4 in winter. Narrowing to ¥50 or more, it is 30 in spring, 4 in summer, 0 in autumn and winter (34 slots a year). And in the prior year, Tokyo had 0 slots above ¥30 in spring and 101 for the year — so this spring concentration is not a permanent property but a structure that first appeared in spring 2026 (Tokyo's average area price for Apr-Jun 2026 was ¥19.35/kWh, well above the ¥14.32 across the full window).

The extremes themselves are gone too. Slots above ¥100/kWh were zero nationwide across the 12 months (the prior year had just two in Chubu), and even at ¥75 or more there was only one slot in Hokkaido (¥80.0, 23 October 2025). The mechanism is traceable in primary sources — after Onagawa 2 and Shimane 2, the grid-connection of Kashiwazaki-Kariwa 6 (February 2026) restored the reserve margin and erased the scarcity of the ceiling, while solar deepened the midday floor. High prices have shifted from "a rare strike" to "a broad spring spread" — the way you earn has itself changed. As another first, Tokyo carried out its first area renewable output curtailment on 1 March 2026 (1,180 MW), so curtailment has now occurred in all nine areas. The spring floor is set to deepen further even in the east.

The one exception to the collapse of seasonal common sense is Hokkaido, where 73 slots above ¥30 stood even in winter. But 71 of those concentrate in the 8 December 2025 - 31 January 2026 unplanned outage of the Kitahon interconnector (the earthquake outage of Section 04) — a one-off event, not the seasonal norm. Reading "Hokkaido earns even in winter" as a permanent locational edge is dangerous; in a normal winter, realized values are thin like everyone else's.

Charging is spring too. Slots pinned to the ¥0.01/kWh floor (annual) are led by Shikoku 1,148 and Kyushu 1,146, with Tohoku 646 and Hokkaido 598 up 40-50% year on year. Most occur in spring. Solar surplus sinks midday and it spikes in the evening — the deepening of the duck curve carries the minimization of charging cost and the upside in selling price in the same season. We put all of this on one calendar.

08 — The daily job since 14 March

The move to day-ahead / 30-minute trading changed the very time axis of operation. The ΔkW supply once decided weekly is now a daily task: each morning, decide how many of tomorrow's 48 slots go to JEPX and how many to ΔkW. By rule, surplus power must be bid in full into spot, with the unsold portion bid into the balancing market, so "abandon JEPX entirely and put all 48 slots into ΔkW" is hard to sustain; in practice it converges on a combined pattern of 18 charge/discharge slots plus 30 ΔkW slots. Because ΔkW is a capacity (reserve) product, you can bid while holding an available state without consuming discharge energy, so it coexists with arbitrage. The ¥0.06 per ΔkW per slot fee may be included in the bid price under the guidelines, so it is not a pure out-of-pocket cost.

The skeleton of the allocation is taught by the reversal of the charging band. Measuring in Tokyo the share of the daily cheapest slot falling at midday (10:00-14:00): 35% in October, 83% in November, 87% in January, 79% in February — in autumn and winter the battery is a machine that "fills cheap at midday and discharges at the evening lighting peak." This reverses to 29% in May (the annual low), and the charging band returns to the pre-dawn trough (the nine-area average shows the same shape, 84% in November to 29% in May). In the curtailment-peak season, the range of midday negative / ¥0.01 and the evening surge is the year's widest, and midday turns into the main battlefield for discharging. This round trip between midday and pre-dawn is the backbone of seasonal operation.

Two points on the ΔkW side, too. The assessment line is a command value of ±10% and a 90% stay rate (Method II), and three non-conformances in one month for the same product lead directly to a trading suspension and a redo of the live test — with a higher ΔkW weight in winter, this is a seasonal risk. And second, in Hokkaido and Tokyo the pumped-hydro bilateral contracts ended at end-March 2026, opening room in the fast-product procurement — a tailwind for the ΔkW side this fiscal year for owners in those two areas.

This allocation call is the daily work of SOC management, assessment compliance and pay-as-bid bid-price design. Whether in-house or delegated, the selection criterion for the aggregator (operator) is the single point of "how precisely it can run the daily co-optimization" — the 10-30% fee in Section 02 is the price for that skill — and that is the conclusion in this regulatory environment. That Tokyo at +40% and Kyushu at -11% coexisted in the same 12 months is the flip side of the truth that operating skill divides the result as much as location does.

09 — Supply: where is the pipeline stacking up?

An area's 3-5 years are decided by four forces: supply (the battery pipeline), demand (large loads), grid (interconnectors) and rules. First, supply. The biggest variable eroding future spreads is, before the interconnectors, the connection rush of batteries themselves.

Grid-scale batteries (high-voltage and above). Connection contract applications = Next-Generation Grid WG, 7th meeting, Doc 1-1 (9 Feb 2026, end-December 2025 snapshot, national total 28.69 GW); connection studies = same, 4th meeting, Doc 4 (24 Sep 2025, mid-2025 snapshot, national ~143 GW). Grid-connected capacity is about 250 MW (mid-2025) rising to about 640 MW (end-2025 ⚠️; the latter needs primary confirmation), so the gap between pipeline and live operation is very large. The denominator of the cannibalization index is FY2035 peak demand in OCCTO's FY2026 demand projection (Chubu, Hokuriku and Kyushu are ❓ as the detailed table is unconfirmed).

The leader flips by date. In the mid-2025 snapshot, Tohoku's 4.00 GW of applications was the national largest, but in the end-2025 snapshot Kyushu leads with 5.90 GW — +2.85 GW in half a year, the fastest priority-connection momentum in the west. The connection-study leader is consistently Tohoku at 56.60 GW = about 40% of the nation, and Tohoku's demand is flat to slightly down. Its structure of stacking supply it cannot absorb in-area and depending on transmission to Tokyo is unchanged. By demand ratio, Chugoku rises to the top at 0.408 — applications doubled from 1.73 to 4.13 GW in half a year, i.e. 4.1 GW stacked on a 10 GW area. Hokkaido and Tohoku follow at 0.383; Tokyo, though large in absolute terms at 5.17 GW, is the healthiest on the index because demand is an order of magnitude larger. Kyushu's contraction (-10.9%) reads, consistently with primary sources, as its priority connection compounded by a high curtailment rate (4.8% actual in FY2024, a FY2026 outlook of 6.9%, the highest nationwide; national curtailment volume is projected at about 2.53 billion kWh, about 1.25x the prior year). The discipline side has lined up too — after caps on the number of connection studies and tighter land-title and deposit rules, from June 2026 connection applications require simultaneous acceptance of the generation (discharge) and demand (charge) sides. The more speculative reservations are excluded, the higher the scarcity value of the remaining slots.

10 — Demand and grid: demand rising in the east, transmission arriving late

The lead role on demand is the east. Tokyo-area peak demand rises from 54,529 MW in FY2025 to 58,880 MW in FY2035, about +4.3 GW (mainly the individual accounting of data centers and semiconductors; a roughly 250 MW-class DC campus is also starting up in Inzai and Shiroi). Hokkaido grows at +1.2% a year, the highest nationwide, and holds Rapidus in Chitose (pilot operation planned for 2027). Kyushu has TSMC/JASM's second Kumamoto fab on a schedule of first shipment end-2027 and completion December 2029. But the contracted power (MW) of Rapidus and TSMC's second fab is undisclosed in the operators' primary sources (❓ reports and estimates only), so the scale of the demand rise must be seen with a range.

Sources: OCCTO Wide-area Grid Development Committee, 96th/99th/101st meetings; 24 Dec 2025 press (one-year extension of the HVDC implementation-plan deadline); Wide-area Grid Development Plan (filed 15 Oct 2025); the eastern-region planning process; ANRE grid diagram (May 2025) and others. Kanmon saw market separation in about 24% of hours in FY2024.

Most upgrades cluster from end-FY2027 to 2039, so the windows in which grid constraints ease fundamentally within the 3-5 year forecast horizon are limited. Whether Hokkaido's and Tohoku's surplus can flow to Tokyo's rising demand hinges above all on the Tohoku-Tokyo upgrade (November 2027) and the fate of the Sea-of-Japan HVDC; whether the latter's implementation plan is filed by the deadline is the single biggest fork dividing Hokkaido's five years. On the other hand, only the Tohoku-Tokyo link completes its upgrade to 10.28 GW in November 2027, so the 56.6 GW study balance and the upgrade run almost simultaneously — rather than a uniform "saturation ahead," the Tohoku-Tokyo axis is best read as a standoff.

11 — Rules: do not mix "in force," "decided in principle" and "under deliberation"

We put the rule changes that alter revenue premises on one page, tagged by regulatory stage. The moment you mix stages, the business plan goes wrong — the triple compression in force and the reference-price doubling under deliberation differ completely in certainty.

Sources: ANRE Institutional Design Subcommittee 108th-114th and Stable Electricity Supply WG 1st; Study Group on the Simultaneous Market, second interim summary (15 Oct 2025); Next-Generation Grid WG 6th/7th/11th; the balancing platform (5 and 13 Feb 2026); OCCTO FY2030 application guidelines (draft) (30 Jun 2026).

The read is this. The headwinds on the ΔkW side are all in force and irreversible; the next fork is only whether the ¥10 cut is triggered on summer results. The tailwind on the capacity side (the reference-price doubling) is under deliberation, and whether it is reflected in the FY2030 main is first fixed by the end-July demand curve. The main pillar of the last 12 months is still ΔkW, but if realized values converge toward the floor (the market-average side), the main pillar shifts to capacity — the fork on that is held by the summer results and the end-July demand curve.

The last two rows are a tectonic shift on the development side. With the cap on connection studies and the land-title requirement, the hurdle to newly take grid slots clearly rises this summer. Conversely, the scarcity value of projects that already hold a connection-study answer or a connection contract rises in relative terms. For buyers, "a good site with a connection slot" becomes, more than ever, an asset that cannot be bought with time.

12 — Per-area reads and the decay curve (3-5 years)

Read plainly, the body of the 3-5 years is, before "area selection," the decay curve. The five open-door areas' realized snapshot is in the mid-¥140,000/kW/yr range — ¥290-300M/yr for a 2 MW unit. At what speed will this be cut by the staged cap cuts (¥15 to ¥10 to ¥7.21), the 1σ volume, and the simultaneous market (target early 2030s)? The table below is our estimate range for the landing point (the cruising level once the cuts have progressed and battery realized values have converged toward the market average), converted to 2 MW annual revenue. It is an estimate with stated premises, not a fixed value. Bull = ΔkW cap held at ¥15, capacity reference-price doubling reflected in FY2030, simultaneous market delayed. Base = one step down to a ¥10 cap, doubling reflected, competitors increasing gradually. Bear = ¥7.21 cap reached, competitors surging, simultaneous market partly applied within the horizon.

Unit: approximate 2 MW annual revenue. In ¥/kW/yr terms, roughly a ¥20,000-43,000 range (per nominal 2,000 kW). An estimate range combining primary-source rule parameters with this piece's measured data. It varies widely with a specific project's connection cost, construction-cost contribution and operating skill.

13 — Monitoring triggers, and what we still don't know

The events that could change this read can be laid out with dates. The checkboxes are for monitoring — each time a publication or ruling lands, we update the relevant section of this piece.

And we leave what we could not find out standing as unknown. The opacity of the rules is itself an input to the investment call.

Null findings (a report of what itself could not be found): The JEPX spot data is fully complete for both the current and prior years — 365 days x 48 slots — with zero missing values or interpolation. Slots above ¥100/kWh were zero nationwide in the current window (the prior year had only two, in Chubu). New-regime standalone results for secondary-① are unpublished as bids were scarce. The additional auctions for delivery FY2027-2029 have not been held (corroborating the single-main-auction policy). Hokuriku and Shikoku have no individual DC/semiconductor demand accounting.

In closing — buy the static No. 1, or buy five years of resilience?

Assembling the last 12 months' snapshot in realized terms, the five areas whose door stayed open are near-level at ¥290-300M/yr for a 2 MW unit (median), led by Kyushu at about ¥295M. Hokkaido and Tohoku, No. 1 and No. 2 in JEPX, cannot join that row because pumped-hydro bilateral contracts block the door — the arbitrage map and the ΔkW map do not overlap. The body of revenue sits not in the area but in the balancing market itself, and that body is exactly what the rules go to cut in stages. On the floor + core that survives the cut, Hokkaido and Tohoku are thick; Hokkaido turns into a time-limited candidate whose both doors open from this fiscal year as the bilateral contracts end, while Tohoku's door stays shut with 40% of the nation's connection studies stacked overhead. Tokyo, fifth on floor + core, is the only area where momentum, events, clearing room and demand all point up.

So the honest answer to "which area do you recommend" is not the No. 1 of a static ranking. The current standings and five-year resilience are different tables. Attack with arbitrage, capture with events and reserves, or fix the floor and buy selectively — decide the strategy first, and the area is decided. After that, the individual conditions — connection slot, construction-cost contribution and contract design — make a further difference of tens of millions of yen within the same area.

ScienceX brokers development rights and projects in each of the nine areas. The statistics in this piece are about the market as a whole, but what finally divides the investment call is a specific project's connection terms and contract state. From area selection to project-level judgment, we help at the same precision as the measured data.