Carriers moving the same cargo over the same trade can differ by 20% to 35% in well-to-wake emissions intensity, according to VesselBot's Decoding Maritime Emissions report for the third quarter of 20261, 5. The study tracked 83,817 voyages by 6,730 container-carrying vessels1.
Fleet-wide intensity was almost unchanged at 235.1 g CO2e/TEU-km, up 0.5% on the year1. Total emissions fell 1.9% to 55.5 million tonnes of CO2 equivalent, while transport work rose 1.2% to 712 billion TEU-km, one TEU being the standard 20-foot container moved one kilometre5.
A flat average says little about the ship booked for next month's consignment, and less about what the carbon in that booking costs.
Where the gap comes from
Five of six vessel-size classes improved their intensity, from 3.4% for Panamax ships to 12.9% for very large container ships1. Feeders went the other way, rising 1.3% to 285.9 g CO2e/TEU-km, and they made up 65% of tracked voyages1.
Utilisation explains much of the spread. Voyages below 50 g averaged 80% utilisation, against 32% for voyages at 800 g or more1. On China and East Asia to North Europe, Evergreen had the highest absolute emissions of the three carriers examined, at 14,273 tonnes of CO2 equivalent per voyage1. It also had the lowest intensity, 47.4 g, about 17% under the lane average of 57.3 g1. Its ships carried 20,400 TEU on average at 96% utilisation1.
Two limits apply. VesselBot examined only three carriers on that trade5. Its figures are modelled estimates, drawn from AIS ship tracking and digital twin models of fuel consumption1, 5. A single quarter is not a carrier ranking.
What the gap is worth per TEU
The EU ETS covers half the emissions of a voyage that starts or ends outside the EU, and all emissions between EU ports and in port2. Shipping companies surrender allowances for 40% of 2024 emissions, 70% of 2025 emissions, and 100% from 20272. EU carbon permits traded at EUR 90.16 on 4 November, on a contract that tracks the benchmark3.
The calculation is straightforward. Take a voyage's emissions per TEU, halve it for the extra-EU rule, and multiply by the permit price. VesselBot's September case study, which priced allowances at EUR 80, put the ETS cost of CMA CGM's 58 direct Singapore to Rotterdam, Valencia and Piraeus voyages at about EUR 26 to EUR 31 per TEU4. If two carriers on a lane differ by 20% to 35% in intensity, their ETS cost per box differs by about the same percentage, given the same routing and permit price.
One caution. The intensity figures are well-to-wake, while the ETS charges the CO2 ships report, plus methane and nitrous oxide from 20262. Well-to-wake also counts the fuel's production6, so the two do not map one-for-one, and the ETS gap may be narrower or wider than the intensity gap. Treat any per-TEU figure as an order of magnitude.
The surcharge is not the cost
Carriers publish carbon charges as named line items. VesselBot worked through a 100 TEU Bremerhaven to Norfolk booking on a Maersk vessel with an ETS surcharge of EUR 64 per TEU4. For the Asia to Europe voyages above it compared an Energy Transition Surcharge of EUR 56 to EUR 70 per TEU with the modelled cost4.
Both charges are flat per TEU and both also cover FuelEU Maritime, so neither follows a single voyage's exposure4. For EU and EEA to US voyages, VesselBot put the trade average at EUR 32.4 per TEU across 344 voyages in the first half of 20264. The Bremerhaven to Norfolk port-pair average was EUR 28.2 and the modelled cost of an actual sailing EUR 12.1. The standard deviation around the trade average, at EUR 32.6, was as large as the average itself4.
FuelEU sits on top. It caps the yearly average greenhouse gas intensity of energy used by ships above 5,000 gross tonnage calling at European ports6. The limit starts with a 2% cut in 2025 and reaches an 80% cut by 2050, on a well-to-wake basis6. The first reports, covering 2025, were due to verifiers on 31 January this year6.
What a buyer can see before booking
EU MRV generates per-voyage data. Companies must monitor emissions, fuel use, distance, time at sea and cargo carried for each voyage, and have the annual report verified2. What the Commission publishes is an annual report on the monitored fleet, not a shipment-level feed2.
The IMO system gives less. The Data Collection System gathers fuel consumption, and the Carbon Intensity Indicator turns it into an A to E rating for ships of 5,000 gross tonnage and above7. That is a ship's annual operating band, not the carbon in your container.
Pre-booking estimates therefore come from models. VesselBot combines AIS positions with digital twin models to estimate fuel burn, and argues that benchmarking a carrier is not the same as estimating a shipment5.
Common formats are only now arriving. A BIMCO-chaired group published version 1.0 of two voluntary formats in October, Operational Vessel Data and Vessel Emissions Reporting, naming ETS cost allocation and FuelEU pooling among the uses8.
What to ask in the 2027 round
Ask which vessel and service will carry the cargo, and for that ship's CII rating7. Ask for modelled intensity in g CO2e/TEU-km on your lane, the utilisation assumed, and whether the number is verified or modelled1, 5.
Split the carbon charge into allowance cost, FuelEU cost and any margin, because published surcharges bundle them4, 6. Ask for ports and routing as part of the price. In VesselBot's sample, a Singapore to Felixstowe to Zeebrugge rotation put only 0.5% of voyage emissions in ETS scope, because the UK is outside the scheme4.
Then agree a reset clause. The permit price rose 7.5% in the past month and is 10.6% above its level a year earlier3.
Watch the Q4 report for whether the gap persists, and whether feeders close any of it1. Watch voyage times too. Some 166 medium and large container ships transited Suez in July to September, against 16 a year earlier, yet China to North Europe still averaged 33.2 days1. Port time on North Europe to the US East Coast rose 52.1% to 2.1 days, and longer port stays burn fuel without adding transport work1, 5.