What CSRD Actually Requires for Waste: An ESRS E5 Data Checklist

Most CSRD readiness work on waste starts in the wrong place. Teams open ESRS E5, see the word "circular economy," and go looking for a strategy narrative. The narrative is the easy part. The part that fails assurance is a short list of weight-denominated datapoints that have to reconcile to something a third party can inspect, per waste stream, per treatment operation, for the whole reporting boundary.

This is a practitioner walkthrough of what the standard actually asks for on waste, what source data each datapoint implies, and which requirement is most likely to be the one you cannot satisfy from hauler invoices alone.

Does CSRD require me to report on waste at all?

Two filters decide this, and they are independent of each other.

The first is scope. The Corporate Sustainability Reporting Directive obliges large and listed companies to report sustainability information, and the European Commission maintains the authoritative summary of who is covered and when on its corporate sustainability reporting page. That population has been materially narrowed. Under the Omnibus I simplification package, adopted in February 2026 and in force since 18 March 2026, the requirements apply to EU companies "employing, on average, over 1,000 employees and with a net annual turnover of over €450 million"; the European Parliament tracks the file on its legislative train entry for Omnibus I. If you built a CSRD programme against the original wave sequencing, re-confirm your entry year before you spend another quarter on it.

The second filter is materiality. ESRS E5 Resource use and circular economy is a topical standard, not a universal one: its disclosure requirements apply where the topic relates to material impacts, risks and opportunities. For a manufacturer, a grocer, a hospital system, or a large hospitality portfolio, waste is very difficult to argue out of materiality. For a professional services firm with a leased office footprint, the assessment may land differently. Either way the conclusion has to be documented, because the absence of the disclosure is itself something an assurance provider will ask about.

What exactly does ESRS E5 ask for on waste?

Waste sits inside Disclosure Requirement E5-5, Resource outflows. In the version of ESRS Set 1 currently in force — Delegated Regulation (EU) 2023/2772, Annex I, ESRS E5 — the undertaking discloses, in tonnes or kilogrammes, the total amount of waste generated; the amount diverted from disposal broken down between hazardous and non-hazardous and by preparation for reuse, recycling and other recovery operations; the amount directed to disposal broken down by incineration, landfill and other disposal operations; and the total amount and percentage of non-recycled waste. EFRAG also publishes the standard in navigable form in its ESRS knowledge hub.

The revised ESRS annex adopted in July 2026 keeps the same skeleton but changes the units of presentation and adds one datapoint. Total weight of waste generated stays. Diverted and disposed quantities are expressed as proportions of total waste generated rather than absolute weights. And a new item (e) requires "the proportion of waste for which the final destination is unknown, expressed as a percentage of total waste generated."

Two application requirements in the revised annex are worth reading closely because they quietly invalidate common shortcuts. Weights "shall reflect the material’s weight in its original state and shall not be presented with further data manipulation, such as reporting it as ‘dry’ weight" — so moisture-adjusted organics tonnage is out. And incineration with energy recovery counts as "other recovery" rather than disposal only where it meets the conditions of point R1 in Annex II of the Waste Framework Directive, which is a facility-level determination, not a hauler’s marketing claim.

Why is "final destination unknown" the datapoint that will hurt?

Because in most enterprise portfolios the true answer is a large number, and until now nothing forced anyone to write it down.

Consider what a typical waste record actually contains. A hauler invoice tells you a container was serviced at a site on a date, at a price, sometimes with a weight, and usually with a service description rather than a material stream. It very often does not tell you which facility received the material. A commingled recycling load leaving a distribution centre may go to a materials recovery facility whose residual fraction is landfilled; the invoice records the whole load as recycling. Under the previous framing you could report that tonnage as diverted and nobody had a mechanism to challenge it. Under the new item (e) you are being asked to separate what you know from what you assume, and to publish the ratio.

The useful thing about a disclosed unknown-destination percentage is that it is reducible, and reducing it does not require a new sustainability strategy. It requires facility-level destination data on the loads you already generate — which comes from either a hauler willing to report receiving facilities per load, or your own weighed records at the point of generation, or both reconciled against each other.

What source data do these datapoints actually require?

Working backwards from the disclosure to the raw record, four things have to be true of every load in the reporting period.

How does this connect to my Scope 3 Category 5 number?

It is the same underlying dataset, which is the single most useful thing to understand about this work.

The GHG Protocol’s technical guidance for Category 5, waste generated in operations sets out a waste-type-specific method whose activity data is "waste produced (e.g. tonne / cubic meter) and type of waste generated in operations" plus, "for each waste type, specific waste treatment method applied (e.g. landfilled, incinerated, recycled)." That is tonnage by stream by destination — precisely the E5-5 input set, viewed through a carbon lens instead of a mass lens.

A team that builds the E5-5 data properly has already built its Category 5 inventory, and vice versa. A team that builds them as two separate projects gets two numbers that do not reconcile, which is a worse outcome than either project failing on its own. We have written separately about how the measured version of that calculation is assembled in Measured, Not Estimated and about the emission factor side in our EPA WARM guide.

What should a sustainability lead do in the next two quarters?

The sequencing that works is unglamorous and does not begin with software selection.

1. Confirm your entry year and which version of the ESRS applies to it. Everything downstream is scoped by that answer, and the answer has changed twice.
2. Document the E5 materiality conclusion, with reasoning, whichever way it lands.
3. Inventory your current waste records site by site and score each site on the four attributes above. You are looking for the count of sites where destination is unknown, not for an average.
4. Calculate your unknown-destination percentage for one recent month across the whole portfolio. This is your baseline and it will be worse than expected.
5. Fix the highest-tonnage unknowns first, by requiring receiving-facility reporting in hauler contracts at renewal and by weighing at the point of generation where the hauler cannot or will not report.
6. Only then choose the system that will hold the normalized dataset and produce both the E5-5 table and the Category 5 figure from it.

On the tooling question, for completeness: Dyrt’s Impact Engine exists to be step six. It ingests hauler invoices, utility bills, facility scale data, RFID readings and field measurements through APIs, EDI, email parsers and PDF OCR; normalizes billing periods, units, duplicates and facility IDs; and calculates against actual material flows using the emission factors the target framework requires. Where the constraint is that nobody is weighing anything at the site, a DWIT kiosk and scale produces the weighed, timestamped, stream-tagged load record the disclosure ultimately rests on. Neither one substitutes for steps one through five.

The reason to do this work in that order is that the disclosure is not really asking how circular you are. It is asking whether you know what left your buildings and where it went. That is an operational data question with an operational data answer, and it is answerable in a quarter or two if you start from the load record rather than from the report template.