EU CBAM definitive from 2026; importers must file first declaration by 30 September 2027

The EU carbon border measure stopped being a reporting drill on 1 January 2026. The definitive CBAM regime is now in force, and if you import cement, iron and steel, aluminium, fertilisers, electricity, or hydrogen, the clock on real compliance is already running. This quarter is when you should move it from the later pile to the this-week pile.

What changed on 1 January 2026. The definitive regime replaced the 2023 to 2025 transitional reporting period. Now it is a live obligation, not a data-collection warm-up. The shift matters because the transitional period let importers practice; the definitive period makes the practice count. For a trader bringing any of the six goods into the EU, the question is no longer whether you will report, but whether you are structured to do it without bleeding margin.

Who must comply. Importers of CBAM goods above 50 tonnes per year must become authorised CBAM declarants. Below 50 tonnes you are exempt under the de minimis rule, except hydrogen and electricity, which have no de minimis escape. So the threshold is a real line in the sand: if your annual embedded tonnage of cement, steel, aluminium, or fertiliser stays under 50 tonnes, you can step back. If you touch hydrogen or electricity at any volume, you are in regardless. Map this before you assume you are small enough to ignore it.

The timeline you must put on the wall. The first annual declaration plus certificate surrender is due by 30 September 2027, covering your 2026 imports. Certificate sales open on 1 February 2027. From 2027, declarants must each quarter buy at least 50 percent of the embedded emissions in their imports, with the price linked to the EU ETS. Read that again: the cost becomes a quarterly cash outflow tied to carbon price, not a once-a-year accounting note. The 2026 imports you are making right now are the ones that trigger the first bill.

The indirect representative trap. Indirect representatives are not exempt. If you appointed one to handle customs and assumed that person or entity absorbs the CBAM exposure, you read the rule wrong. The obligation sits with the importer of record, and leaning on a representative does not move it. I see this mistake constantly with mid-size shippers who think a customs broker's signature transfers the carbon liability. It does not. You own it, and you should price it, not hope the broker quietly covers it.

What embedded emissions actually means for your cost. The certificate you surrender each quarter corresponds to the carbon embedded in the goods you imported, valued at the EU ETS price. If the ETS price is, say, 75 euros per tonne of CO2 equivalent, and your annual imported embedded emissions are 1,000 tonnes, the annual certificate cost runs toward 75,000 euros before any free-allocation phase-out adjustments. That is real money hitting the landed-cost line, and it is the kind of line your procurement team currently does not have a budget code for.

Here is a worked example. Suppose you import 800 tonnes of aluminium products a year, with embedded emissions of 1.5 tonnes CO2 per tonne of metal, so 1,200 tonnes of embedded CO2 annually. At an ETS price of 75 euros, that is 90,000 euros a year in certificates. Because from 2027 you buy at least 50 percent per quarter, you need roughly 11,250 euros of certificate purchasing each quarter just to stay compliant on this one material. Now add your steel and fertiliser lines and the number scales fast. The 50-tonne de minimis would have exempted you only if total embedded emissions stayed under 50 tonnes, which 1,200 is not even close to.

The first move is mapping. Pull every HS and CN code you import and check it against the six CBAM goods lists. Cement, iron and steel, aluminium, fertilisers, electricity, hydrogen each have defined CN codes, and the trick is that fabricated products made from those base materials are often in scope too. A bolt made of aluminium, a steel pipe, a fertiliser blend: all can pull you in. Do not screen only the obvious raw commodity; screen the finished and semi-finished items, because that is where the surprises live.

Start collecting embedded-emission data from suppliers now, this quarter, not next year. The method is prescribed and the data has to come from the actual production route, not a generic average you guessed. The earlier you ask, the more leverage you have, because your suppliers will get this request from every EU-bound buyer and the slow ones will bottleneck everyone. Build the data feed into your supplier onboarding so that a new vendor cannot ship to the EU without embedding the carbon figure in the commercial documents from day one.

Appoint your declarant early. The authorised CBAM declarant is the entity that submits the annual declaration and surrenders certificates, and you do not want to be hunting for one in the third quarter of 2027 while the September deadline looms. If you are the importer of record, decide whether you become the declarant yourself or appoint an indirect representative, and either way lock it in writing this year. The administrative capacity to do this well is finite, and the firms that can do it will be busy in 2027.

Budget the certificate purchases as a cash-flow item, not a footnote. Mark 1 February 2027 for the first certificate sales, and from that point plan quarterly buys of at least 50 percent of embedded emissions. If you treat this as a year-end surprise you will either miss the surrender or scramble for cash at a bad carbon price. The price is linked to the EU ETS, which moves, so build a range into your model and buy on a schedule rather than on panic. A finance team that sees this as a quarterly tax understands it; one that discovers it in September 2027 will not.

For the small importer under the threshold, do not get comfortable too early. The de minimis exempts below 50 tonnes of embedded emissions for most goods, but hydrogen and electricity have no such relief, and the threshold applies per goods group in a way that rewards knowing your exact tonnage. If you are close to the line, a single large shipment can push you over, and then the declarant requirement snaps on retroactively for the year. Keep a running embedded-tonnage count so you are never surprised mid-year that you have crossed into the regime.

The strategic play is cost pass-through. If you are the importer and the carbon cost is real, it belongs in the price you charge your EU customer, not in your margin. The contractual language should allocate the CBAM cost explicitly rather than burying it, because the price will keep moving with the ETS. Buyers who understand the mechanism will accept a transparent carbon line; buyers who are surprised by it will dispute it. Get ahead of the conversation now, while the first bill is still a 2027 event and you have time to renegotiate terms.

Common mistakes to avoid. Assuming the customs broker handles it. Assuming the de minimis covers everything. Assuming you can assemble the emission data in a week before the deadline. Assuming the ETS price will stay low. All four are wrong, and each turns a manageable compliance line into a penalty or a blocked shipment. The definitive regime has penalties, and an unauthorised or late declarant is exactly the kind of status that draws them. Treat CBAM like a tax that has already been enacted, because it has.

One more thing that changes the math over time: as the EU ETS free allocation to domestic industry phases out, the carbon price signal behind CBAM only gets stronger, not weaker. The early years may feel manageable, but the trajectory points to embedded carbon becoming a permanent and rising line in every CBAM goods import. Building the system now, while volumes are still under the first declaration, is cheaper than bolting it on once the cost is large enough to hurt. The defenders who wait are paying tomorrow's larger bill with today's unprepared process.

The practical summary is not complicated even if the regulation is. Know your tonnage, know your codes, name your declarant, collect the data, and budget the quarterly buy. None of that is exotic, but all of it is easy to defer until the deadline is close, and that is exactly the trap. CBAM is live; the only question is whether you are the importer who planned for it or the one who explained it after the fact.

The embedded-emissions calculation is prescribed, but the input quality is your responsibility. For each goods group there is a defined method, and in the early years you may use default values where installation-specific data is missing, but those defaults are usually conservative and cost you more than real data would. The smarter move is to pull the actual emissions from your supplier's production route: the fuel mix, the process efficiency, the energy source. A steel mill running on hydro power has a very different embedded figure than one on coal, and the certificate cost follows that difference straight to your quarterly bill. Start the data request now, because the suppliers with clean figures will be the ones everyone wants, and the laggards will bottleneck your declaration.

Price volatility is the part of CBAM people under-think. Your certificate cost is linked to the EU ETS price, which has swung hard in recent years and will keep moving. If you budget a flat 75 euros and the market prints 110, your quarterly buy costs nearly half again as much, and that lands in the same period as everything else. Build a range into the model, run the quarter at the high case so finance is never surprised, and consider whether your customer contract lets you pass the swing through. A buyer who agreed to a fixed carbon line at 75 will not be happy when you invoice at 110 unless the contract said the line floats with ETS. Write the float in before the price moves, not after.

The free-allocation phase-out is the slow drumbeat underneath all of this. As EU domestic industry loses its free ETS allowances, the CBAM charge on imports is designed to rise to match, so the cost you see in 2027 is the floor, not the ceiling. Planning as if this is a one-time hit is a mistake; planning as if the line grows every year is closer to right. The companies that build the system while volumes are still modest will absorb the ramp gradually, while the ones that wait will meet a larger bill with an unprepared process and no supplier data. The trajectory, not the starting number, is what should shape your 2026 and 2027 decisions.

Contract language is where the cost actually lands or gets passed. If you sell on DDP or similar delivered terms into the EU, the CBAM cost is yours to carry or to price in; if you sell on EXW or FOB, the importer of record is your customer, but you will still feel it when they demand a transparent carbon line or shift to a supplier who provides one. The practical step is to put an explicit CBAM or carbon-cost clause in the commercial terms, stating the mechanism, the data each side supplies, and how the float is handled. Buyers who understand the rule will accept this; buyers who are surprised by a new line will dispute it, and disputes cost more than the clause ever would.

For the smaller importer who thinks this is someone else's problem, the de minimis is real but narrow. Below 50 tonnes of embedded emissions you are out, for most goods, which covers a lot of modest traders. But hydrogen and electricity have no such relief, and the threshold is per goods group, so a trader importing a little steel, a little aluminium, and a little fertiliser could cross the line on the combined figure even if no single line looks large. Keep a running count of embedded tonnage by group, and the moment a single shipment pushes you over, the declarant requirement switches on for the whole year. The trap is not the threshold; it is not knowing you crossed it.

Ignoring CBAM does not make it gentle. An unauthorised declarant, or a late or missing declaration, draws penalties under the definitive regime, and at the border the goods can be the thing that gets held while the paperwork is fixed. For a trader living on thin margins and tight delivery windows, a blocked shipment is worse than the certificate cost, because the cost is known and the delay is not. The defenders who treat this as a tax that has already been enacted build the process; the ones who treat it as a rumour explain themselves after the hold. The regime is live, and the first declaration covers 2026 imports, so the clock is already running on what you are shipping today.

Record-keeping is the unglamorous sibling of compliance, and it is where audits bite. You must be able to show the embedded-emission figures, the method used, the supplier source, and the calculation behind every declaration, and those records have to be retrievable for years. Build the file as the data arrives, not in a panic before the September 2027 deadline. A clean folder per goods group, with the supplier letter, the method note, and the maths, turns a stressful audit into a box-tick. The operators who scramble for proof at deadline are the ones who discover a missing supplier figure that now has no one to ask.

A second worked example makes the steel case concrete. Suppose you import 500 tonnes of steel products with embedded emissions of 2.1 tonnes CO2 per tonne, so 1,050 tonnes of embedded CO2 a year. At an ETS price of 75 euros that is roughly 78,750 euros annually, or about 9,840 euros per quarter once the 50 percent rule bites. Add that to the aluminium line from earlier and a modest fertiliser line, and a mid-size importer can be looking at six figures a year in certificates by 2028 as the free allocation fades. The number is not apocalyptic, but it is real, recurring, and rising, and it belongs in the landed-cost model from this quarter, not from the deadline quarter.

For the finance team, the quarterly cadence is the part to internalise early. From 2027 you are not making one annual reckoning; you are buying certificates every quarter at a price you do not fully control, and each buy is tied to the embedded emissions you imported that quarter. That means your cash-flow forecast needs a carbon line that moves with the ETS, not a fixed estimate you set once. The teams that build the buy into the regular treasury rhythm avoid the year-end scramble; the ones that discover it in September 2027 treat it as a surprise and either miss the surrender or pay a bad spot price. Put the line in the model now, while the volumes are still small enough to be painless.

The indirect representative question deserves a harder look than most give it. Appointing an indirect representative does not move the obligation off your books; it shifts who files, not who owes. If your representative fails to declare or under-declares, the liability and the penalties still land on the importer of record, which is you. So the representative is a tool for capacity, not a shield for risk. Vet them the way you would vet a tax advisor: ask for their CBAM track record, their data handling, and their error history, and keep your own copy of every figure submitted. Outsourcing the paperwork is fine; outsourcing the accountability is not, and the regime is written so that the importer always answers for it in the end.

Small importers often ask whether they can simply stop importing the CBAM goods and avoid the whole regime. Walking away from a profitable line to dodge a cost you can pass through is rarely the right math; the carbon charge is a known, forecastable line, while losing the product or the customer is a permanent hit. The smarter play is to keep the trade, price the carbon transparently, and use the data you collect to pressure suppliers toward lower-emission production, which over time lowers the embedded figure and the certificate bill together. Treat CBAM as a lever on your supply base, not just a tax on your imports. The suppliers who cannot or will not provide clean emission data become a visible risk in your file, and that visibility is itself useful when you negotiate the next contract. A buyer who can show a supplier exactly what their embedded tonnage costs at the border has a concrete reason to ask for a cleaner process or a different source.

• Map every HS and CN code you import against the six CBAM goods lists (cement, iron and steel, aluminium, fertilisers, electricity, hydrogen).
• Calculate your 2026 embedded-emission tonnage per goods group to see if you cross the 50 t/yr de minimis threshold.
• Appoint an authorised CBAM declarant (or indirect representative) now; do not wait for Q4 2027.
• Start collecting supplier embedded-emission data this quarter using the prescribed calculation methods.
• Mark 1 February 2027 in your calendar for certificate purchases and budget the quarterly 50% embedded-emission buy.