Everyone is looking at the truck in fleet electrification. Range anxiety, battery chemistry, charging speed. These dominate the conversation. And they matter. But a recent industry-wide review by Joost Commandeur (TU Delft / Shell Global Solutions), together with Bart De Schutter and Neil Yorke-Smith, mapped 91 digital solutions across site energy management, fleet scheduling, and charging session management. The findings point to a different bottleneck entirely: these three systems barely talk to each other.
Commandeur and colleagues split electric fleet operations into three decision layers: energy at the site, fleet scheduling, and management of the charging session itself. And they then asked how mature the software is at each layer and how well they integrate.
The answer is uncomfortable. Charging management software (CPMS) is genuinely advanced: dynamic load balancing, priority queues, tariff-based optimization are all widely deployed. Energy management systems (EMS) are mature too, but mostly for tasks that predate electrification: solar dispatch, battery scheduling, grid limits. Fleet scheduling tools, the layer that actually decides which vehicle drives which route when, lag far behind. Many operators are still running spreadsheets.
That unevenness would be tolerable if the three layers exchanged information well. They don’t. The review describes integration between them as limited to one-way “master-slave” data flows through APIs, with no shared optimization and little bidirectional coordination. Each system makes decisions with only part of the picture. What the authors call partial observability. A polite way of saying: everyone is optimizing in the dark, separately.
This is precisely the coordination problem seen on the ground. A diesel truck only needs a route and fuel that’s always available. An electric truck needs a route, a charger, site power headroom, grid capacity, and increasingly a read on electricity prices that can swing by a factor of ten within a single day. None of that complexity is a vehicle problem. It’s an information problem. Right now, the systems that hold that information don’t share it.
What’s telling is where the innovation is concentrated. CPMS vendors are pushing hardest into smart charging and predictive analytics because they sit closest to real-time operations. Fleet scheduling, further from the electrons, has the least urgency and the least investment, even though it’s where charging needs actually originate.
Fleet scheduling is the least mature layer, still largely manual or spreadsheet-based due to low BEV penetration and conservative planning. Charging (CPMS) is the most advanced, with widespread load balancing, prioritization, and tariff optimization; energy management (EMS) is mature for traditional tasks but rarely fleet-aware.
Across all three layers, systems operate in silos, connected only by one-way APIs (e.g., OCPP) rather than unified optimization engines. This produces partially observable, fragmented decision-making, with no evidence of bi-directional coordination or closed-loop automation across layers. Relevant data such as telematics, prices, and power availability exist but aren’t shared effectively.
Future research should prioritize interoperability frameworks and multi-layer co-optimization algorithms that jointly handle site energy, charging, and scheduling under uncertainty.
Interesting news is that Einride is acquiring Flipturn for $38.4 million in shares to integrate electric trucks, charging management, and access to third-party networks. Flipturn’s software manages more than 250 MW, predicts charging needs, optimizes power delivery, and lowers energy costs. Combined with Einride’s Saga AI platform, the deal aims to remove a major barrier to fleet electrification and accelerate North American growth at scale.
So the real competitive question for the next decade of fleet electrification isn’t who builds the best battery or the fastest charger. It’s who builds, or connects, the systems that let site energy, charging, and route planning make decisions together instead of in sequence. We don’t have a vehicle problem. We have a coordination problem. And right now, almost nobody is building for that.
Walther Ploos van Amstel.
Check out: Einride’s Flipturn deal creates a charging powerhouse