The Fuel Card Market Is Worth $10 Billion and $1 Trillion

Ask three analysts to size the fuel card market and you can come back with three numbers spread across a factor of a hundred, every one of them defensible. The revenue issuers actually retain – the fee charged per transaction, the spread between the rate they buy fuel at and the rate they sell it to the fleet, subscriptions, reporting, charging roaming margin – was USD 9.40 billion in 2025. The fuel those cards settle across their networks is about USD 218 billion. All road fuel bought by commercial vehicles, by card, by account, by corporate credit card or in cash, is of the order of USD 1 trillion.

That is one industry measured at three perimeters, and the hundredfold spread between the smallest and the largest is the most common source of error in published estimates of this category. It is not a disagreement about growth. It is a disagreement about what is being counted – which means a great many comparisons drawn between two of those figures are not comparisons at all.

The report that reconciles them is the Global Fuel Card & Fleet Payment Solutions Market forecast, and the reconciliation is the analytical product in it, not the growth rate.

Three rungs, one industry

Only the first rung behaves like a market. It has a price, a margin, competitors, a growth rate, and a customer who can move. It is small for a structural reason: a card business earns basis points on a flow it does not own. DKV Mobility, Europe’s largest brand-independent network, reported 2024 revenue of EUR 839 million against transaction volume of approximately EUR 19.4 billion. On the throughput view it is a EUR 19.4 billion company; on the revenue view it is an EUR 839 million one. Both statements are true and only the second is a market.

Rung two is what actually passes across the rails, and it is the number most external sizing reports. It is derived rather than borrowed here: at DKV’s disclosed 4.32% gross take, USD 9.40 billion of fee revenue implies roughly USD 218 billion of settled transaction value. The assumption is that DKV’s take is representative, and Europe’s truck-heavy mix means it probably overstates the North American figure – so USD 218 billion is the top of a range, not a midpoint.

Rung three is the fuel itself, where the card is one payment instrument among several. A category quoted at USD 1 trillion and a category quoted at USD 9.4 billion can be the same industry described at rung three and at rung one. The gap between them is not a data-quality failure. It is two correct answers to two different questions.

What a factor of 23 does to a business case

Rung one to rung two is 23.2 times. Rung one to rung three is 106.4 times. Two practical consequences follow, and both are expensive.

A supplier that sizes its addressable fee pool off a throughput figure overstates that pool by roughly 23 times. It is an error that survives a business case comfortably, because a number that large does not invite challenge; it invites a slide. An investor who values an issuer off throughput reaches the opposite error: the take looks like well under one per cent of the flow, and therefore looks like a commodity with no room in it, when on the card-network base the take is 4.31 per cent – four times the impression the wrong denominator gives.

The working rule that follows is blunt, and it is worth carrying into any conversation about this category: a number that arrives without a stated perimeter should be assumed to sit on rung two, because rung two is the rung that flatters.

A fee business wrapped around a fuel business

Once the perimeter is fixed, the shape of the market becomes legible. It is a fee business wrapped around a fuel business, and the two grow at different speeds.

The active card base runs from 65.75 million to 103.20 million between 2025 and 2033, a 5.80% CAGR. Revenue runs from USD 9.40 billion to USD 18.35 billion, an 8.72% CAGR. Revenue per active card therefore rises from USD 142.97 to USD 177.81 – 24.37%. The value CAGR sits 2.92 points above the volume CAGR, and that difference is the whole investment case.

It is also visible in the accounts, which is unusual for a payments category. Eurowag reported 2024 net revenue of EUR 292.5 million, up 14.0%, serving 302,076 active trucks, up 10.0%. Revenue per active truck was EUR 968.3 against EUR 933.7 in 2023 – a rise of 3.71%. Growth came almost entirely from adding trucks, not from earning more per truck. Any model that projects a fuel card business at its revenue growth rate is projecting customer acquisition rather than pricing power, and will be wrong the moment acquisition slows.

The margins say the same thing from another angle. Corpay’s adjusted EBITDA margin of 53.57% on USD 3,974.6 million of revenue is not a processing margin – it is processing plus float plus spread, and float shrinks the moment rates fall. Eurowag, closer to a pure fee model, earned 41.6%, down from 42.4% in 2023 and 42.8% in 2022. The twelve-point gap between the two is a fair estimate of what a balance sheet is worth in this category at current rates.

Geography then decides what a card is worth at all. Europe is 44.68% of global value on 18.8% of cards; North America is 46.28% of value on 56.1% of cards. A European card earns USD 340.08 a year against a North American card’s USD 117.89 – 2.88 times, because it sits behind a 40-tonne truck crossing tax jurisdictions rather than a two-van contractor staying inside one state. That ratio is more useful to anyone deploying capital than the headline growth rate.

Electrification pulls the rungs apart

Here is the part that makes the hundredfold spread a moving target rather than a fixed confusion. The two rungs are not just measured differently; they are heading in different directions.

Most of a fuel card’s fee is charged per transaction, not per litre. An electrified van charges several times a week where a diesel van refuels once, so transaction counts per vehicle rise even as the litres behind them fall. Meanwhile the merchant rebate – the other large revenue line – is funded out of retail fuel margin, and there is no retail fuel margin at a public charger to rebate. So rung one gains from frequency and loses on rebate, while rung two loses on litres. The card is being reissued into a network where the issuer neither funds the purchase nor earns the rebate: DKV disclosed 798,000 charge cards in circulation, up around 35%, against a fuel-card base growing far more slowly.

The litre is not a stable unit either. Blending mandates change what a litre contains, which changes the excise and the tax evidence the card is there to manufacture – a subject in its own right in our ethanol fuel blending coverage. And beyond the pump, the transaction itself is being absorbed upward into the vehicle, where Plug & Charge under ISO 15118 lets a car authenticate against a contract certificate installed at or after manufacture. No card presented, no app opened, and the billing relationship belongs to whoever holds the certificate.

The practical implication for anyone holding a throughput number is that it decays faster than a fee number. Estimate this market at rung two and you are modelling a quantity that electrification is actively shrinking; estimate it at rung one and you are modelling the part that grows.

The perimeter is also where double counting hides

Rung one excludes a great deal on purpose, and each exclusion is a report of its own – which is why the headroom for counting the same revenue twice is generous.

Telematics hardware and subscriptions are excluded: already sized in our Global Fleet Management coverage at USD 25.4 billion in 2025 rising to USD 58.7 billion by 2033. Tolling revenue collected on behalf of toll authorities is excluded as a pass-through carrying a fixed agency fee. Consumer credit cards used incidentally to buy fuel are excluded. And the charging infrastructure that the roaming margin is earned on is sized separately in our EV charging infrastructure coverage, at USD 40.22 billion in 2025 to USD 147.28 billion by 2033 – the supply side of the same relationship, and the two reports should be read together rather than added together.

A “mobility payments” figure assembled from rung two of one forecast plus rung one of another is arithmetically valid and analytically meaningless. That is the failure mode the three-rung table exists to prevent.

The number to attack

Every forecast in this category rests on one assumption, and in this case it is the 24.37% rise in revenue per active card. Three things must hold at once for it to happen: the per-transaction fee is not competed to zero; transaction frequency per vehicle rises at least as fast as litres per vehicle fall; and attach across analytics, tolling and compliance lifts the non-litre share of revenue from roughly a fifth to roughly a third.

The sensitivities are worth stating in numbers rather than adjectives. If the fee is competed away and value growth collapses to the volume rate of 5.80%, the 2033 figure is about USD 14.8 billion rather than 18.35 billion. If only the attach assumption fails, it is around USD 16 billion. Those are the honest bands, and the difference between them and the base case is the value of the compliance relationship the industry is trying to build.

Credit is the risk that gets priced least. The issuer funds thirty to forty-five days between paying the merchant and collecting from the fleet, on a customer base where around 90 per cent of businesses are small or medium-sized. Eurowag’s own chairman described 2024 as a year in which a slowdown in freight demand and rising costs produced a peak of insolvencies. A card book is a concentrated bet on the solvency of the smallest operators in the economy, and the loss given default is a full fill plus fee, not a residual balance. The same dynamic, at a much earlier stage, is what our Saudi Arabia fleet management coverage shows in a market priced on transaction volume rather than credit.

What would settle it

The reason this reconciliation has to be constructed is that nobody publishes the inputs. No issuer in the category discloses transactions per vehicle and revenue per transaction together. Fee schedules are opaque, per-transaction pricing is rarely published, and revenue per active card has to be inferred from a revenue line and a card count that appear in different places in a set of accounts.

Regulation has made one price visible and left the other dark. AFIR – Regulation (EU) 2023/1804, Article 5(4) – requires the ad hoc price at publicly accessible chargers of 50 kW and above to be based on the price per kWh and displayed at the point, which puts a ceiling on the roaming spread. Nothing in EU law requires an issuer to disclose its transaction fee, its roaming spread or its revenue per card. The first jurisdiction to mandate that disclosure would do more to stabilise this market’s numbers than any forecast, because it would make the 24.37% assumption independently checkable rather than merely plausible. The tax side of the same product is already in motion: ViDA, adopted in March 2025, brings digital reporting for cross-border business-to-business transactions from 1 July 2030, which either makes the issuer the fleet’s compliance agent or removes the aggregation layer that justified the fee.

The companion reading on that half of the business – how the card became a tax document with a payment function attached, and what the German penetration data says about who is winning – is The Fuel Card Is a Tax Document That Happens to Buy Diesel.

The test to carry away

Any figure describing this category should be asked one question before it is used: which rung is it on? The answer changes the multiplier by a factor of 23, and it changes the direction of travel, because the fee rung grows while the throughput rung shrinks per vehicle.

A market quoted at USD 1 trillion and a market quoted at USD 9.4 billion are the same industry described at two perimeters. The difference between them is not the size of the opportunity. It is whether the person quoting it knows what they are quoting.