Carbon-Aware Routing: Optimizing for Emissions Alongside Cost and Time

Most routing engines were built to minimize distance, time, or cost — and emissions were assumed to track roughly with all three. That assumption holds less often than it used to. A shorter route through dense urban traffic can emit more than a longer route that keeps a vehicle at a steady highway speed, and a marginally more expensive rail leg can cut emissions by an order of magnitude relative to the truck leg it replaces.

Carbon-aware routing treats emissions as a modeled variable, not a byproduct — which means feeding the solver real emissions factors per vehicle type, fuel, load, and road grade, rather than assuming distance is a sufficient proxy.

A routing engine that only sees distance and time will find the cheapest path, not the lowest-carbon one — and on many lanes, those two paths are not the same route at all.

§ 02Emissions factors have to vary by more than vehicle class

A generic 'emissions per kilometer' constant hides more than it reveals: load factor, idling time at congested junctions, gradient, and even ambient temperature (which affects both combustion efficiency and EV range) all move the real number meaningfully. Routing decisions built on a single static factor per vehicle type will systematically misjudge exactly the routes where the true emissions profile differs most from the average.

§ 03The trade-off is usually smaller than planners expect

In deployments we've supported, meaningful emissions reductions — in the 8-15% range — have typically been achievable with cost increases in the low single digits, once the carbon-cost trade-off is made explicit in the objective function rather than treated as an afterthought applied to an already-finalized route plan. The routes that are 'free' emissions wins (same cost, lower emissions) are usually a bigger opportunity than most networks have actually gone looking for.