When we were siting our first stations in Pune, the instinct was to think like charging infrastructure. Put stations where riders congregate, where they can wait. Near a popular lunch spot, next to a housing society gate where riders gather between delivery windows. That instinct turned out to be partially wrong, and understanding why it was wrong shaped how we now think about station placement entirely.
A charging station needs to be where a rider can wait. A swap station needs to be where a rider will be next. The difference sounds subtle but it propagates through every aspect of the site selection process. Swap stations should function as in-route infrastructure, not destination infrastructure. If your rider has to make a significant detour to reach a swap station, the 90-second swap time advantage is partially offset by travel time, and the rider is more likely to defer the swap until they are critically low rather than swapping proactively when range is at a comfortable 30-40%.
The Route-First Siting Framework
Our current station siting process starts with rider route data, not map aesthetics. We look at the actual GPS trace data from riders operating in a target area and identify the corridors where rider density is highest during peak delivery hours. A corridor is not a street; it is a stretch of route segments that riders traverse repeatedly within a shift. In Pune's Baner-Balewadi area, the corridor from the Baner main road down toward Balewadi High Street and the residences in the Pashaan area appeared repeatedly in our trace data before we placed a station there.
The placement question then becomes: where along that corridor can a rider make the swap with minimum route deviation? We score potential sites on four factors. First, road accessibility: can a rider on a two-wheeler pull in and out easily without navigating a tight parking lot or a complex junction? Second, proximity to the existing route: how far off the rider's natural path is the station entrance? Third, power access: does the site have sufficient grid connectivity for our charging equipment? Fourth, space: a station needs at minimum 8-12 square meters for packs and charging equipment, plus a safe swap area.
Why Landmark Proximity Is Often Misleading
The most common mistake in swap station siting is placing stations near demand generators, meaning restaurants, offices, or shopping complexes, rather than on demand corridors. A delivery hub generates rider presence for order pickup, but it is also where riders start their routes, often with a full battery. A station near a popular restaurant cluster will see riders who just started their shifts and do not need a swap, not riders who are mid-route and low on range.
The right question is: where do riders have 20-30% battery remaining and still have 30-45 minutes of remaining shift time? That is the window where a proactive swap makes sense and where a well-placed station converts the most value. Mapping that window requires understanding average route lengths, typical battery consumption curves for the scooter models in use in that corridor, and shift start/end patterns by area.
In our Q4 2025 Pune pilot, we made one placement error by this logic. One of our 12+ stations was sited close to a major logistics hub that turned out to be a pickup point, not a mid-route stop. Riders arriving there were typically at 80-90% charge. Utilization at that station was low compared to stations placed mid-corridor. We have since prioritized repositioning that capacity in a future network expansion iteration.
Grid Capacity Is a Harder Constraint Than It Appears
Three-phase power access sounds like a simple prerequisite, but in Pune's urban and peri-urban areas it is often the binding constraint on station placement. Many of the ground-floor commercial spaces in high-rider-density corridors run on single-phase connections, which are inadequate for the simultaneous charging of multiple battery packs. A site that scores well on route proximity and accessibility can fail the power access check, and bringing in a new three-phase connection has both cost and lead-time implications that affect rollout planning significantly.
Our current approach is to prioritize sites with existing three-phase access even if they are slightly less optimal on route proximity, rather than building our rollout plan around sites that require electrical upgrades. The tradeoff is a small compromise in route optimization for a large gain in deployment speed. As the network matures and we can plan electrical upgrades with more lead time, the calculus changes.
How Swap Station Siting Differs from Charging Station Logic
Charging station siting in India has historically followed two models: public charging points near commercial areas (destination charging) and residential or office complex installations for overnight or dwell-time charging. Both models assume the vehicle will be parked for a significant time.
Swap station siting rejects both assumptions. The vehicle is parked for 90 seconds. The station needs to be on the route, not at the destination. The capacity planning is for simultaneous swap throughput at peak demand, not for how many vehicles you can queue. A swap station in a high-traffic corridor serves 40-80 swaps per day; a slow commercial charger serves 8-12 vehicles per day at best. These are fundamentally different infrastructure types being deployed in different locations for different reasons, and conflating their siting logic produces stations that nobody actually uses.
Station Siting Is an Ongoing Process
We have not finished siting our network, and we do not think there is a final state where we "complete" placement. Rider corridors shift as delivery platforms change their order density by area, as new residential developments open, and as the rider fleet in a given zone grows or contracts. A station placed optimally for a corridor in October 2025 may become less relevant if the corridor's primary delivery platform reduces coverage in that area.
This is one argument for our partner host station model, where local property owners host the swap equipment rather than us owning the real estate outright. It gives us more flexibility to reposition network capacity as corridor patterns shift, without being locked into a long-term lease on a site that has become suboptimal. For fleet operators considering their own charging infrastructure, this is the flexibility that fixed charging points cannot easily provide.