Industries / Transport and vehicles / Trucking and freight / Fleet management
Fuel and charging
Fuel and charging is the work of getting energy into every vehicle at a known cost: fuel cards and the controls on them, fuel taxes, fuel-saving programs, and the switch to electric vehicles.
6
Milestones achieved, from 1983 to 2025.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1983
Cards that record each fill
Fleet drivers paid for fuel with general-purpose cards until the early 1980s. Comdata launched a fuel card in 1981 at about 10,000 truck stops, by the trade magazine Automotive Fleet's account. Wright Express, a spinoff of a heating-fuel company in Portland, Maine, began in 1983 with a card that trucks could use at three unattended stations in Portland. As oil companies put electronic card readers in their stations in the late 1980s, Wright Express worked with them so that each purchase recorded the odometer reading, a driver ID or PIN, and the date, time and amount, so a fleet could see what each vehicle used.
Credit
Comdata; Wright Express (now WEX), from an idea of Parker Poole III of the A.R. Wright Company
1996
One fuel tax return
A truck that crosses state lines owes fuel tax to each state it drives through. In 1983 Arizona, Iowa and Washington set up the International Fuel Tax Agreement (IFTA), under which a carrier holds one license, files one tax return each quarter with its home ("base") jurisdiction, makes one payment and faces one audit, and the base jurisdiction passes the other members their shares. A 1991 US law, the Intermodal Surface Transportation Efficiency Act, required the 48 contiguous states to collect fuel use taxes in line with IFTA by 30 September 1996, and Alberta became the first Canadian province to join in 1992.
Credit
Arizona, Iowa and Washington; IFTA, Inc.; the US Congress
2004
Fleets benchmark their fuel
The US Environmental Protection Agency (EPA) and 15 "charter partners" began work in 2002 on a voluntary program for freight, and EPA launched it as SmartWay in 2004. Carriers and shippers measure and benchmark their freight fuel use and emissions with EPA's tools, shippers use the results to choose more efficient carriers, and EPA verifies fuel-saving equipment. By EPA's estimate, partners have saved 397 million barrels of oil and US trucking companies $55.4 billion in fuel costs since 2004.
Credit
US EPA and its charter partners
2017
Companies pledge electric fleets
The Climate Group, a nonprofit, launched EV100 in 2017 for companies committing to fully electric fleets. In its 2026 report it counted 400,000 battery-electric vehicles deployed by members worldwide, with 1.7 million more to switch by 2030, and said more than 70% of members added no diesel or petrol vehicles to their fleets the previous year. The counts are the members' own, gathered by the Climate Group.
Credit
The Climate Group and the EV100 member companies
2020
India's big order stalls
In September 2017 Energy Efficiency Services Ltd (EESL), a state-run company under India's Ministry of Power, chose Tata Motors to supply 10,000 electric cars for government use, an order Autocar India called the world's largest single electric vehicle purchase; Mahindra & Mahindra supplied part of the first batch. The tender, a public call for bids, asked for more than 130 km per charge on the official test. In June 2018 a senior official quoted by the newspaper Mint said the cars did not cover even 80 to 82 km in city driving, and EESL said only 150 had been delivered; Tata Motors replied that it had "not heard range as a major concern." By February 2020 EESL had dropped the 10,000-car tender and asked for bids on 1,000 sedans with more than 180 km of range and fast charging.
LessonA range certified in a test can fall short of a fleet's daily use, and trip records can show the gap before an order is placed.
autocarindia.comautocarindia.comautocarindia.comautocarindia.com
Credit
EESL; Tata Motors; Mahindra & Mahindra
2025
California's fleet rule shrinks
The California Air Resources Board (CARB) adopted its Advanced Clean Fleets rule in April 2023, in force from 1 October 2023. It required state and local government fleets, large private fleets that CARB calls "high priority" fleets, and drayage trucks, which haul containers to and from ports and rail yards, to move to zero-emission vehicles in stages. The private-fleet parts needed a waiver from the federal government, and on 13 January 2025, a week before a new US administration took office, CARB withdrew its request and stopped enforcing them. The board approved repealing those parts on 25 September 2025, and the repeal was filed with California's Secretary of State on 10 September 2026; the rules for government fleets remain.
LessonThe parts of the rule that depended on federal approval fell when the federal government changed, while the parts for the state's own and local government fleets, which needed no waiver, stayed.
Credit
CARB
Watch for
AI checks each fuel purchase
WEX says its SecureFuel service, rolling out from 3 August 2026, combines fleet-card purchases with live data from vehicles' telematics, the data a vehicle sends over a mobile network, through providers such as Geotab and Samsara. AI fraud models use it to decline suspicious purchases before they go through, for example by checking where the vehicle is and confirming that its fuel level rose afterward. WEX says its payment network covers about 95% of US fuel stations.
Achieved whena named fleet or an outside auditor publishes fraud losses and wrongly declined purchases from before and after using it.
Claimed by
WEX, August 2026
Open
01
A public fuel-fraud test bench
A public set of at least 100,000 simulated but realistic fleet fuel purchases, each paired with the vehicle's GPS track and fuel-level readings, with known kinds of misuse planted in it: fills larger than the tank, purchases while the vehicle was elsewhere, fuel for a second vehicle on the same card, and purchases outside working hours. Then score simple rules, open-source anomaly detectors and AI models on how many misuses they catch and how many honest purchases they would wrongly decline.
The briefCredit · unclaimed02
EV readiness, checked afterward
A follow-up study of fleets that went electric after a telematics-based suitability study. For each vehicle, compare what the study predicted (daily range needed, charging time, cost per mile) with the first year of actual use, including vehicles sent back to gasoline.
The briefCredit · unclaimed
Open 01 · the brief
A public fuel-fraud test bench
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public set of at least 100,000 simulated but realistic fleet fuel purchases, each paired with the vehicle's GPS track and fuel-level readings, with known kinds of misuse planted in it: fills larger than the tank, purchases while the vehicle was elsewhere, fuel for a second vehicle on the same card, and purchases outside working hours. Then score simple rules, open-source anomaly detectors and AI models on how many misuses they catch and how many honest purchases they would wrongly decline.
Why it is still open
The figures come from the companies that sell the checks. Motive says fuel card fraud can reach 19% of a fleet's spending; WEX launched SecureFuel in August 2026 and Geotab describes tank-capacity checks at the pump, all without published catch rates. Searches in November 2026 found no public, labeled dataset or shared test for fleet fuel fraud.
Built with what exists
Python to simulate routes, shifts and fills; fuel station locations from OpenStreetMap; tank sizes from makers' published specifications; open-source detectors such as scikit-learn's isolation forest; a language model such as Claude to review each flagged purchase and state its reason, in the way the record reviewer in the Note "Seven more flagged" reviews drug-making records. No real cards or payments are used.
First step
List the kinds of misuse from vendors' guides and generate a first 10,000 purchases for 50 simulated vans.
Done looks like
The data, the generator and the scoring code public, with catch rates and wrongful-decline rates for at least five methods on each kind of misuse, and one outside team rerunning the test and matching the published scores.
Open 02 · the brief
EV readiness, checked afterward
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A follow-up study of fleets that went electric after a telematics-based suitability study. For each vehicle, compare what the study predicted (daily range needed, charging time, cost per mile) with the first year of actual use, including vehicles sent back to gasoline.
Why it is still open
The published studies are forecasts. In 2019 NREL (now NLR) and Sawatch Labs ran such studies for five US government and utility fleets and reported suitability scores, not later results. Geotab's 2021 study of 91,252 vehicles managed by Enterprise Fleet Management estimated $167 million in savings over four years if 45% of them went electric. EESL's order in India stalled over range (see Achieved above). This research found no study comparing the forecasts with what happened.
Built with what exists
Public fleets' records of electric vehicle energy use, downtime and cost, obtained by records request; the fleets' published suitability reports; NLR's open-source vehicle simulator FASTSim to rerun each forecast; Python.
First step
Find five public fleets that published a suitability study before 2023 and request their electric vehicle records since then.
Done looks like
At least 300 vehicles from five fleets, with forecast and actual daily range use, energy cost per mile and charging problems side by side, and the share of vehicles each study placed wrongly.