Stand on a busy Jakarta street in 2026 and the soundtrack of the city has changed in a way that surprises people who haven’t been back for a couple of years. The two-stroke motorcycle whine that was the default background of Indonesian urban life is, in pockets, replaced by the soft electric hum of delivery riders on Gesits, Volta, Selis, and the increasingly visible Chinese-built scooters that have flooded into the market. The transition is partial, uneven, and genuinely happening — and the last-mile delivery sector is where it’s furthest along.

Indonesia’s electric two-wheeler market is now somewhere around 3% to 5% of new sales depending on which dataset you trust, with the share rising. Total electric motorcycle stock on Indonesian roads has crossed 200,000 units and the rate of new registrations has accelerated meaningfully since the 2024 expansion of the government incentive program for domestically-assembled EVs. Most of these vehicles are not in private hands. They’re in fleets, and a substantial share of those fleets are doing delivery work.

Why the unit economics work for delivery

The case for electric two-wheelers in delivery work is much stronger than the case for private ownership, and the reason is straightforward: utilisation. A delivery rider in Jakarta, Surabaya, or Bandung is doing 100–180 kilometres a day across a 12-hour shift. The fuel cost saving versus a comparable petrol scooter is in the order of 30,000–50,000 IDR per shift. Multiply that by 25 working days a month and the lifetime fuel-cost differential is substantial — large enough that the higher upfront purchase price of an electric scooter pays back inside 18–24 months even before factoring maintenance differences.

Maintenance is the second factor. Electric two-wheelers have fewer moving parts, no oil changes, no spark plugs, no clutch wear. The brake and tire wear remain similar; the drivetrain costs are dramatically lower. Fleet operators tracking total cost of ownership consistently report electric models running 25%–40% cheaper than petrol equivalents over a three-year operational cycle, assuming charging access is solved.

The third factor, which doesn’t always show up in spreadsheets, is rider preference. Delivery riders who’ve switched to electric mostly don’t want to switch back. The reduced engine noise during a long shift, the absence of vibration, and the smooth torque delivery from a stop are quality-of-life improvements that fleet operators are starting to use as a hiring and retention tool in markets where rider churn is high.

The charging problem, more honestly

The single biggest constraint on faster electrification of last-mile delivery is charging infrastructure, but the constraint is more nuanced than the casual headline version suggests.

In-warehouse charging during shift changes works well in theory but only if the warehouse has enough electrical capacity to charge thirty or fifty scooters simultaneously, and most existing logistics warehouses don’t. Upgrading the electrical service to a typical Indonesian warehouse to support fleet-scale charging is non-trivial and is one of the genuine costs that fleet electrification programs underprice in their initial budgets.

Battery swap is the workaround that has gained the most traction. The model — pull into a swap station, exchange a depleted battery for a charged one in 60–90 seconds, ride away — fits Indonesian last-mile delivery patterns better than wait-for-charge. The networks operated by SwapEnergi, Volta, and Gogoro’s Indonesian operations have expanded to several thousand swap points across major Java cities. Outside Java, the swap-station density drops sharply.

Public AC and DC charging for two-wheelers remains underdeveloped because the economics for the operator are weak — a two-wheeler taking 90 minutes to charge is using infrastructure that could otherwise serve a higher-paying car customer. The dedicated two-wheeler charging that exists tends to be either fleet-owned or part of an integrated operator’s swap-station network rather than open public infrastructure.

The grid itself remains a constraint at the system level. PLN’s Statistics Indonesia data on electrification shows continued progress on coverage, but the load implications of mass two-wheeler electrification are real, particularly in cities where evening peak coincides with the post-shift charging window for delivery fleets.

Where it’s actually working

The geographic pattern of EV last-mile adoption maps closely to where the swap infrastructure is dense. Greater Jakarta is the clear leader. Bandung and Surabaya are following. Yogyakarta, Semarang, and Bali (with Denpasar’s particular tourism-driven delivery profile) have meaningful electric fleet activity. Outside these markets, the share of electric in last-mile fleets drops to single digits.

The use-case pattern also matters. Food delivery has electrified faster than parcel delivery. The shorter trips, more predictable shift patterns, and the more concentrated geography of food orders in dense urban districts all favour the electric scooter’s range and charging profile. Parcel delivery — particularly the longer suburban routes that make up much of e-commerce last-mile — is closer to the edge of what current battery technology comfortably supports for a full shift.

A consultancy I follow closely, Team400, has been doing AI-driven route optimisation work that’s relevant here — fleets running optimised routing can squeeze meaningfully more deliveries out of a single charge cycle, which expands the operational envelope for electric vehicles into use cases that wouldn’t otherwise pencil out. The compounding effect of better routing plus electrification is, I think, the genuinely interesting story rather than electrification on its own.

The next eighteen months

The trajectory through 2027 looks roughly like this. The electric share of new last-mile fleet purchases in Greater Jakarta will likely cross 50% if current incentives hold — the unit economics are strong enough that fleet operators are choosing electric on commercial grounds, not just on regulatory pressure. The non-Java cities will follow with a lag of two to three years, paced by the build-out of swap infrastructure.

The cargo three-wheelers and small electric vans for slightly larger delivery payloads are the next interesting segment. Several local manufacturers have product in market and the use case for short-range urban distribution looks structurally favourable, but the products are less mature than the two-wheelers and the operational data is thinner.

The risk to all of this is policy stability. The current incentive structure has been the difference between a viable electrification curve and a stalled one for several model years now. Whether the post-election government continues, expands, or trims the incentives is the variable that fleet operators are watching most carefully. A clean transition is possible. A messy one would be expensive for everyone, particularly the small operators who’ve already committed capital to electric fleets on the basis of the current numbers.

The soft hum of an electric delivery rider passing a Jakarta street vendor in 2026 is a small thing. Multiply it by hundreds of thousands of riders, by tens of millions of trips, and it’s how the carbon arithmetic of Indonesian urban logistics actually starts to bend.