日本の洋上風力、新しい配点では「迅速性」は240点中10点しかない

Japan’s first offshore wind auction was decided on two numbers: how cheap the bid was, and how soon the turbines would turn. The consortium that won all three zones in December 2021 walked away from all three in 2025, taking about 1.7 GW off the board and forfeiting security deposits estimated at around ¥22 billion (Shulman Advisory).

On 5 June 2026 the Ministry of Economy, Trade and Industry and the Ministry of Land, Infrastructure, Transport and Tourism published the answer. They revised the operational guideline for the occupancy tender system in general sea areas, after a joint committee process that began in September 2025 and a public consultation that ran from 22 January to 22 February 2026 (METI, 5 June 2026).

The press release lists four changes: setting an assumed supply price band, reweighting the deliverability score, scoring deliverability more finely, and lowering the weight of speed while allowing more schedule flexibility.

The revised guideline itself carries the numbers.

Price and deliverability now carry equal weight

The guideline states that the allocation between the supply-price evaluation and the deliverability evaluation is, for the time being, 1:1. It adds that once a track record has accumulated, and taking European precedent into account, a reweighting toward price may be considered on the assumption of mature deliverability (revised operational guideline).

So the reweighting is written as temporary. The ministries say they will look at it again once developers have a delivery record to point at.

The price band stops the race downward

Under the revised design the ministries set an “assumed supply price band” between the supply-price ceiling and a price that assumes a developer applies realistic ingenuity on top of the level considered necessary to complete the project.

The operative rule is a flat score. Where a bid comes in at or below the ceiling minus that band, the price score is a uniform 120 points. The band itself is set at the time of each tender, with the business environment of the day taken into account.

Press reports through late 2025 described this as a price floor. The adopted text does something narrower. It caps the reward rather than disqualifying the bid. Below the band, a lower price buys a developer nothing in the scoring and costs it margin in the build, which is the trade that produced the 2021 result.

The deliverability table, as published

The guideline sets the deliverability allocation at 120 points, split 2:1 between the ability to execute and local coordination. The reasoning given is that execution capability is essential to certain delivery, while local understanding is indispensable to building offshore wind at all.

Ability to execute, 80 points:

• Project plan, speed: 10 points
• Project plan, foundations of the plan: 20 points
• Project plan, execution of the plan: 25 points, split between delivery structure and track record at 6, funding and revenue plan at 14, the plan up to commissioning at 20, and the plan after commissioning at 5
• Stable power supply and supply-chain formation: 25 points

Local coordination and economic effect, 40 points:

• Ability to coordinate with the heads of relevant administrative bodies: 10 points
• Coexistence with nearby shipping routes, fisheries and others: 10 points
• Regional economic effect: 10 points
• Domestic economic effect: 10 points

Add the 120 available on price and the maximum is 240. Speed is 10 of them.

How the 10 speed points are calculated

The clause that changes behaviour is not the 10 points. It is the formula behind them.

Speed is scored relatively. The shortest construction period among bidders sets the full mark, and each bidder is scored against it. The guideline then adds a gate. Where a bidder’s combined score for the foundations and execution of its plan falls below half of the 45 points available for those two items, its speed score is zero. At or above half, the speed score is the period-based figure multiplied by that bidder’s own foundations-and-execution ratio.

It also closes the obvious loophole. The benchmark for the shortest construction period is drawn only from bidders whose foundations-and-execution total exceeds half marks, so an implausible schedule attached to a weak plan cannot set the pace for everyone else.

In 2021 an early commissioning date was an asset you could assert. Now it is a multiplier on a plan the evaluators have already scored, and an aggressive date in a thin submission scores nothing.

The method for everything other than speed moves the same way. Each evaluation item gets a checklist, split 1:1 between a basic standard that asks whether a matter is addressed and an advanced standard that asks whether the treatment is detailed, specific, appropriate and reliable. The score is the allocation multiplied by the share of boxes met.

Supply-chain resilience is now a scored item

The 25 points for stable power supply and supply-chain formation have no equivalent in the original design. The guideline describes what they measure: stable supply of turbine components, fast recovery when equipment fails, and the establishment of a domestic and regional industrial base. It asks bidders to show resilience across both hardware and the systems around it.

That item is now worth more than the plan up to commissioning is worth on its own, and two and a half times what speed is worth. A developer that can name its component suppliers, its spares position and its repair response is scoring on a line that did not exist when Round 1 was bid.

The 14 points for the funding and revenue plan read the same way. The guideline asks evaluators to weigh the financial plan against recent economic trends, material prices and expected future change in the operating environment. Round 1 was lost to a cost curve nobody had modelled. The new table makes modelling it part of the score.

The prefectural governor is in the room

Three of the four local-coordination items carry the same instruction: hear the opinion of the prefectural governor. It applies to the ability to coordinate with relevant administrative bodies, to coexistence with shipping routes and fisheries, and to regional economic effect.

That is 30 of the 240 available points reached through a channel the developer does not control. The governor’s view is solicited as part of the evaluation, not gathered by the bidder and reported. For a foreign developer without a standing relationship in Akita or Chiba, that is a practical constraint on how early the local work has to start, and it is not work that can be compressed into the months before a submission.

What is still unwritten

The press release is explicit that two of the most commercially significant items are not in the operational guideline at all. Bid-winning restrictions, and the rules that apply when a selected operator withdraws, will be set out in the occupancy tender guideline drawn up for each sea area.

For anyone modelling a Japanese bid, that is where the remaining risk sits. The scoring is public and stable now. The consequences of walking away, and any limit on how many zones one party can hold, are per-zone documents that have not been published yet.

Why this matters beyond Japan

Every offshore wind market that ran a low-price auction between 2019 and 2022 is dealing with a version of the same problem. Bids were written against a cost curve that inverted, and the contracts that looked cheapest turned out to be the ones least likely to get built. Japan published its arithmetic, which makes the correction easy to read.

The 2021 design paid for a low number and an early date, and it got both from a bidder who could not deliver either. The 2026 table pays for the plan, and makes the date a function of it.

The same design question sits underneath grid connections, hydrogen support, data-centre power contracts and every other auction where a government buys delivery years in advance from a private balance sheet.

Tech for Impact Summit 2027 runs on 18 and 19 May 2027 at the United Nations University in Tokyo, co-hosted by UNU. Senior leaders from business, policy and capital spend two days on questions of this kind. Participation is by invitation. If this is the work you do, ask us for one at tech4impactsummit.com/apply.