How to build a commercial system for your pipeline

Europe: price is no longer the whole award

Article 26 of the Net-Zero Industry Act requires Member States to apply specified non-price criteria in auctions for renewable energy. The Commission’s guidance, published in 2026, states that this applies from 30 December 2025 to at least 30% of the volume auctioned per year in each Member State, or alternatively to at least 6 GW a year. The volume is the capacity or generation the auction is designed to support, counted across technologies, and it counts in the year the auction is published rather than the year the result is announced. 124

Responsible business conduct, cybersecurity and the ability to deliver on time are pre-qualification. A project that fails them does not bid. Resilience and a sustainability criterion may be pre-qualification, award criteria, or both. Ireland’s consultation on Article 26 states the award weighting directly: price must weigh 70–85%, non-price factors 15–30%, and each non-price award criterion at least 5%. 3 Member States still choose the detailed tests, so the same plant can face two evidence lists.

Great Britain: the price is public, and so is the date

GB CfD rounds remain pay-as-clear price auctions. AR7a awarded 4,905 MW of solar across 157 projects at £65.23/MWh in 2024 prices, against an administrative strike price of £75. Onshore wind cleared at £72.24 against £92. AR8 keeps those caps, puts solar in its own pot, and gives government sight of anonymised bids before budgets are set. 56 A supplier comparison that prices a plant as if the cap were the revenue will not survive the model.

Battery revenue has a second public anchor. The 2029/30 T-4 Capacity Market cleared at £27.10/kW/year, less than half the £60 of the previous T-4. 7 A floor or toll can sit on top of that, but the capacity payment in the close file has to be the payment the auction actually produced, not last year’s.

The grid queue is no longer a waiting room. NESO replaced a first-come queue of over 700 GW with a 283 GW pipeline of generation and storage. A Gate 2 transmission offer is signed within 90 days. 89 That date is an input to the offtake, the EPC programme and the financing case. If the comparison uses a different date, the contract is already wrong.

A developer active in both markets has two scoring systems and one plant. The EU bid needs evidence of delivery and resilience. The GB bid needs a price that can clear, a capacity case and a connection date. The supplier comparison under both has to use the same scope, programme and exclusions. That discipline is the commercial system.

There is no public dataset of commercial hours on European or UK renewable projects, so this section is analysis, not a measurement. It draws on how the auctions above are structured and on Ignis practitioner experience of live projects. We see the same five breaks.

The people who can see these breaks are the same people needed to decide them: the commercial lead, the technical lead and whoever will sign. Their weeks go on version control. The calls that move the deal, margin, risk and the terms that will still be true at close, get whatever time is left. A commercial system reverses that order. Agents can chase the versions and draft the first comparison. A named person still decides the price, the risk and the term.

Ignis sets these five layers up on the client’s own evidence. The descriptions below are the ones we use on the site, because they are also the ones a close has to satisfy. Nothing here requires a rip-and-replace of the model, the data room or the contract tool. It requires that those tools stop being the only place a number can live.

What each layer does on a live contract

Process is the path the team already walks, written down: grid offer, consents, supplier tender, comparison, negotiation, offtake, due diligence, contracts, financial close. Each step has an owner and a definition of done. Without that, an agent will automate whatever version of the path lives in one person’s head.

Context is the memory. Past comparisons, the clauses that were conceded, the exclusions that turned out to matter, and the public rules the current deal has to meet: a CfD cap, a Capacity Market price, an EU non-price test, a Gate 2 deadline. Context is not a dump of old files. It is the set a person is prepared to rely on.

Data is the single set of numbers. Capacity, dates, price, indexation and scope. The tender, the model, the contract and the close file read from it. When a number changes, the other documents are flagged, not quietly edited. This is the layer that stops a 49.9 MW plant becoming 50 MW in one pack and 48 MW in another.

Governance is who may change a price, accept a risk or send a mark-up. The approval sits on the record with the evidence that was in front of the person who gave it. A full audit trail is what a lender asks for when two drafts disagree, and what a board needs when it asks why this term was accepted.

Workflows are the repeatable work: building the comparison grid, chasing missing interface schedules, drafting the first mark-up against the last agreed position, checking that dates match. Agents can run that work. They stop at the decision. The person who owns margin, risk or a binding term still makes the call.

The comparison that closed the last project becomes the starting point for the next tender. The team spends the new project on what changed.

The example is hypothetical. It is a 60 MW solar project with an illustrative 11% load factor, considering a CfD at the AR7a clearing price, an EPC comparison, and a co-located 30 MW battery that might enter the Capacity Market. Public numbers are the price anchors. Everything else is an assumption and is marked as such. We do not use a market EPC price, because we are not publishing one. The point of the tables is the order of magnitude, so a team can see which decisions deserve senior time.

Read the figure from the top. A negotiation that gives away £1/MWh to close a point of interface risk has a price. So does a programme that slips a month because an exclusion was never aligned with the grid date. The Capacity Market line is smaller than a year of solar revenue, which is the point of putting the numbers in one place: the team can see which argument is worth the week.

The same arithmetic is what a European non-price score is competing with. If 15–30% of an award is evidence rather than price, the comparison has to show whether the supplier can actually produce that evidence, not only whether the headline is lower. A resilience statement that the EPC contract then contradicts is not a scoring advantage. It is a close problem waiting for the lender.

1. Name the path. Write the steps from the current gate to financial close, and the person who owns each one. If two people think a step ends in different documents, that is the first fix.
2. Restate every offer onto one scope. Do not rank prices until interfaces, exclusions, indexation and the programme are on the same lines. The ranking is a decision. The restatement can be drafted by an agent and checked by the technical lead.
3. Lock the numbers that are already public. The CfD cap, the last clearing price, the Capacity Market result and the connection date are not opinions. They go in the data layer and the model reads them.
4. Approve changes, do not email them. A price, a date or a risk allocation changes only with a named approval. The previous value stays visible. That is the audit trail a close question will ask for.
5. Leave the record where the next project can find it. The comparison, the rejected clause and the reason are context. The next tender starts there. Ignis is built to hold that record on your evidence, with agents doing the repeatable work and your team keeping the calls.

Investors do not need another narrative of how busy the tender was. They need to see why this price, this date and this risk were accepted, and that the signed contract is the one the model describes. That is a commercial system. The alternative is to rebuild it, project by project, until the pipeline is limited by memory rather than by sites.

Frequently asked questions

How was this analysis done?

Rules and prices. The Net-Zero Industry Act coverage rule is taken from the Commission guidance of 2026 and the Commission’s own description of the support-scheme rules, with the implementing regulation as further context. The 70–85% and 15–30% award weights, and the 5% minimum per non-price criterion, are taken from the Government of Ireland’s consultation describing Article 26, not from a claim that every Member State has already published identical scoring. GB prices are from the DESNZ AR7a results, the AR8 Price and Pot Notice, and Ofgem’s Capacity Market report for 2025/26. The grid figures are from NESO’s December 2025 releases and the connections handbook.

Illustrative calculation. Figure 4 and Tables 1 and 2 use a hypothetical 60 MW plant. Load factor, discount rate, battery size and de-rating are assumptions. The arithmetic is generation times price, a 20-year annuity at 7% (factor 10.59), and a one-twelfth split for a month of delay. They are not a forecast and they are not a measured project.

What this paper does not do. It does not report client results, quote a developer, or invent an EPC market price. It does not claim that Ignis has changed a close outcome. The Ignis system is the record described in section 3. Performance claims would need a measured project, and this paper does not contain one.

Not covered. A full legal reading of every national auction, corporate PPA pricing, and offshore wind CfD pots. The companion paper, Bids are now won on evidence, not hours, treats the bid desk. This one treats the path from comparison to close.