Market Commentary / CSRD & carbon credit quality On the evidence

Europe’s largest companies reported 4.84 million carbon credits. Not one disclosed the project IDs needed to verify them.

A recent analysis of the DAX40 — Germany’s forty largest listed companies — offers the first full read of how Europe’s biggest firms disclose carbon credits under the CSRD. The findings are a near-perfect illustration of the principle this portfolio is built on: a certified, registry-listed, fully disclosed credit can still be a poor one. Certification is not credibility, and traceability is not quality. Here is what the data shows, and what it means for any company reporting under ESRS E1-7.

4.84M tCO₂e in carbon credits reported by 21 active buyers across the DAX40 under CSRD FY2025.
45% of that volume — 2.17M credits — could not be traced to a specific project from disclosures and public records.
57% of the 90 projects that could be independently rated scored below BBB, the commonly cited credibility threshold.
0 DAX40 companies disclosed a carbon credit project identifier in their CSRD filing.

Source: Senken × Sylvera, “Buying Blind? What FY2025 CSRD Disclosures Reveal About DAX40 Carbon Credit Quality & Transparency” (April 2026). Figures as reported in that analysis. See attribution below.

The analysis
39 of 40DAX40 FY2025 CSRD reports read
21 buyersactive carbon credit purchasers
7 registriessearched for retirements
101 projectsidentified · 90 rated
43 countries6 continents
The Disclosure Ceiling · 01

CSRD asks how much. It does not ask which.

ESRS E1-7 requires companies to report aggregate volume, the removal-versus-reduction split, the standards used, and the EU sourcing share. Useful — but the data stops there.

What ESRS E1-7 does not require is the detail that would let anyone check the claim: project names, project IDs, countries, brokers, vintages, retirement dates, or quality ratings. As the analysis puts it, every company in its sample complies with the CSRD — yet none provides enough information in the filing to verify a single credit. The reporting is complete and the verification is impossible at the same time. That is the gap.

This matters because the missing field is small and the company already holds it. A firm that bought credits knows precisely which projects, registries, volumes, and vintages it holds — it chose them. Requiring that project-level detail in E1-7 would cost essentially nothing. The information exists; the standard simply does not ask for it. The result, in the report’s phrase, is the illusion of transparency without the substance of it.

For a company that wants its disclosures to withstand scrutiny — from auditors, boards, or the press — complying with E1-7 is therefore the floor, not the ceiling. The verifiable disclosure that regulators will eventually require, and that a credible buyer can already volunteer, is the project-level record: which project, under which methodology, at which vintage, retired on which date.

Finding One · 02

Over two million credits could not be identified at all.

When the analysts went looking for the projects behind the volumes — across seven registries, press releases, and corporate filings — nearly half the total volume left no traceable footprint.

Of the 4.84M credits reported, the analysis could attribute 46% to projects companies disclosed themselves and a further 9% through registry retirement records — leaving 45% (2.17M credits) that could not be tied to any specific project. The single largest cause is mundane and entirely legal: credits are frequently retired under a broker’s name rather than the end buyer’s, which breaks the public link between company and project. Legal, common — and fatal to independent verification.

The split, by volume
46% company-disclosed · 9% registry-traced · 45% unidentifiable

Six companies named the projects they fund — four voluntarily, two because California’s AB 1305 forces it. Their combined volume is independently verifiable. Ten more did not disclose, but their names surfaced in public registries, making their credits traceable anyway.

The report is careful on one point, and so are we: unidentifiable does not mean low-quality. It means the credit cannot be checked. Without a project ID, independent quality verification is simply not possible — the door is closed before the question can be asked.

Finding Two · 03

For most of what could be checked, the quality did not hold up.

Of the 101 projects identified, the independent ratings agency Sylvera assessed 90. The distribution is the heart of the report: most assessed projects fell below the credibility threshold.

Sylvera rates carbon projects on a scale from AAA (highest) to D, across carbon accounting, additionality, and permanence. In this sample, 57% of assessed projects scored below BBB — the level the agency identifies as the minimum for credible climate impact. The pattern is driven by low-rated renewable energy credits and cookstove programmes: precisely the methodology categories the ICVCM has so far declined to approve under its Core Carbon Principles, citing additionality concerns.

Distribution by project count (not volume-weighted), as assessed by Sylvera. A portion of the 90 are ranged or preliminary estimates rather than formal ratings; see the original report’s methodology. Bar widths are indicative.

The most instructive cases are the ones where everything except quality was right. The report highlights buyers whose portfolios are modest in volume, fully Gold Standard certified, and completely traceable in the registry — one even links to the project from its own CSRD filing — yet score below BBB across the board. One single-project portfolio sits in a category the ICVCM rejected in 2024 for lack of additionality. Being certified, traceable, and disclosed, as the analysis concludes, is not the same as being credible.

The Fix Already Exists · 04

The same company, disclosed two ways.

Two DAX40 companies file under both CSRD and California’s AB 1305. Comparing the two filings for the same buyer shows exactly what project-level disclosure changes — and how little it would cost.

Under CSRD, one of these companies reports a credit volume, a removal/reduction split, a list of standard names, and an EU sourcing share. Every figure is accurate. And from it, you cannot identify a single project, see a single methodology, assess a single controversy, or confirm a single retirement. Under AB 1305, the same company lists each project by registry ID, methodology, and vintage — at which point an independent reviewer can look each one up, check it for controversies, and assess its quality directly. Same buyer, same year, two completely different levels of accountability.

Under CSRD (ESRS E1-7)
× Can you identify a single project? No.
× Can you see the methodologies? No.
× Can you assess quality or controversies? No.
× Can you confirm credits were retired? No.
Under California AB 1305
Every project identifiable → look up in registry.
Every methodology listed → check for controversies.
Every vintage disclosed → assess quality independently.
Retirements confirmable in the registry.

Illustrative of the contrast the report draws between the two disclosure regimes for the same filer. Project specifics are in the original.

The report’s conclusion is restrained, and we share it: this is not a story about bad actors. Most of these companies are buying what the market offers and disclosing what regulation requires. The problem is that the regulation requires too little. One amendment to ESRS E1-7 — project-level disclosure alongside the aggregate volumes — would make European carbon credit purchases independently verifiable, and California has already shown it works.

What It Means · 05

The report describes the problem this portfolio is built to solve.

Strip away the specifics and the analysis lands on two propositions that sit at the centre of how the 360° Impact Portfolio sources and discloses credits.

The first is that certification is not credibility. A credit can carry a respected standard’s label, sit in a public registry, and be fully disclosed — and still fail on additionality, permanence, or measurement. A standard name records that a credit was issued under a programme; it does not record whether the tonne was real, additional, and durable. That is why we screen every offset on its own merits across six criteria — quantification, additionality, permanence, registry and double-counting, safeguards and co-benefits, and independent verification — rather than treating a certificate as the end of the question.

The second is that transparency is the precondition for everything else. The 45% that could not even be identified is the more troubling number, because an unidentifiable credit cannot be assessed at all — the quality question never gets asked. Project-level disclosure is therefore not a nicety; it is the thing that makes independent verification possible. The disclosure pack we deliver with every transaction exists precisely to close that gap: project ID, methodology, vintage, registry, retirement record, and the verification opinion behind them.

A company reporting under CSRD today can comply fully and still be unable to defend its credits to an auditor or a journalist. The fix is to disclose at the project level voluntarily — ahead of the regulation that will eventually require it — and to source against quality criteria rather than against a label. That is the position this portfolio is built to put a buyer in.

Source & Attribution · 06

About the underlying analysis.

Cited work

This page is commentary on, and a summary of, an independent analysis published by Senken × Sylvera: “Buying Blind? What FY2025 CSRD Disclosures Reveal About DAX40 Carbon Credit Quality & Transparency” (April 2026). All figures, ratings, and company-level findings referenced here are theirs as reported in that work; the interpretation and the connections drawn to the 360° Impact Portfolio’s approach are ours.

Senken is a carbon credit procurement partner. Sylvera is an independent carbon ratings provider whose project ratings (AAA highest to D) underpin the quality findings discussed above. The rating distribution includes ranged and preliminary estimates that are not formal Sylvera Ratings; ratings are point-in-time and may change. For the full dataset, methodology, and the company-by-company detail, refer to the original report.

We reproduce only the headline findings necessary to comment on the work, and encourage readers to consult the source directly for the complete analysis.

Next step

Don’t buy blind. Screen, then disclose.

The 360° Impact Portfolio screens every offset on six quality criteria and delivers the project-level disclosure pack CSRD omits. The 30-minute discovery call is the starting point.