In 2015, fleet management meant a GPS dot on a screen and a spreadsheet for maintenance scheduling. I’m not exaggerating. The vast majority of fleets, especially mid-size ones running 30 to 200 trucks, operated with paper logs, calendar-based oil changes, and a dispatcher who knew the routes by memory. The technology existed to do more. Most fleets just weren’t using it.
A decade later, the industry looks completely different. The fleet management market has grown from roughly $8 billion to over $30 billion, and the technology powering it has shifted from “where is my truck” to “what is my truck about to do.” That shift didn’t happen in a straight line. It happened in three distinct waves, each triggered by something specific.
2015-2017: GPS tracking becomes standard, but that’s about it
The mid-2010s were when GPS tracking stopped being a competitive advantage and started being table stakes. By 2015, the hardware had gotten cheap enough that even small fleets could justify installing trackers in every vehicle. The monthly cost per truck dropped below $30 in many cases, and the devices were simple enough that you didn’t need an IT department to manage them.
But the usefulness was limited. You could see where your trucks were. You could set geofences and get alerts when someone left a zone. You could pull historical route data. That was roughly the ceiling for most platforms.
Maintenance was still run off spreadsheets or basic CMMS software that had nothing to do with the GPS system. Driver behavior was monitored loosely, if at all. Fuel tracking meant comparing credit card receipts against mileage. The data these trucks generated sat in silos, and nobody was connecting the dots between location, vehicle health, driving patterns, and costs.
The fleet manager’s job in 2016 was still overwhelmingly reactive. Something broke, and you fixed it. A driver complained about a truck, and you pulled it into the shop. A customer called about a late delivery, and you checked the GPS to figure out what happened. Everything was backward-looking.
2017-2019: The ELD mandate accidentally built the infrastructure
The FMCSA’s Electronic Logging Device mandate changed the industry in ways that had nothing to do with hours-of-service compliance. The mandate, which became effective in December 2017 with full compliance required by December 2019, forced every interstate commercial fleet to install hardware that connected to the truck’s engine.
Before the mandate, a lot of fleets had resisted putting connected devices in their vehicles. Cost, complexity, driver pushback. The ELD rule eliminated the choice. You needed a device in every truck, and that device had to record engine status, driving time, miles, and location.
What happened next was predictable in hindsight but caught many fleet managers off guard. Once you had a connected device in every truck talking to the engine’s computer, the marginal cost of collecting additional data dropped to nearly zero. Telematics providers started packaging ELD compliance with fuel monitoring, idle-time tracking, basic diagnostic alerts, and driver scorecards. What had been sold as a compliance tool became a fleet management platform almost overnight.
The fleet management market grew by over 60% between 2017 and 2020, and a huge portion of that growth was ELD-driven adoption. Companies that had been running on GPS and spreadsheets suddenly had access to engine data, and the smarter ones started using it.
2020-2023: Data gets smart, maintenance goes predictive
The pandemic accelerated everything. Supply chains broke. Freight demand spiked. Driver shortages got worse. And fleet operators who were already collecting data from their ELD and telematics platforms started asking harder questions. Not just “where is the truck” but “why did this truck burn 18% more fuel than that one on the same route” and “how do I stop losing two trucks a month to unplanned breakdowns.”
This is when the analytics layer started maturing. AI and machine learning, which had been talked about since roughly 2016 without much real-world fleet application, started delivering actual results. The difference was data volume. A single truck generates gigabytes of sensor data per month. Across a fleet of 500 trucks over three years, that’s enough data to train models that can actually predict component failures.
The concept that gained the most traction was the digital twin, a virtual model of each vehicle that mirrors its real-world behavior based on continuous sensor data. Instead of applying generic maintenance schedules to every truck, you could build a digital twin of each vehicle that tracks how its specific components behave under its specific operating conditions. When behavior drifts from the established baseline, the system flags a developing problem weeks before a diagnostic trouble code appears.
This wasn’t theoretical. A municipal fleet of 1,400 vehicles, including 90 refuse trucks from three different OEMs, used digital twin technology and detected faults in 30% of their trucks before any DTCs were triggered. They estimated savings of about $500 per vehicle per month. A 100-truck long-haul fleet saved roughly $4,500 per truck per year by catching engine problems early and improving fuel efficiency.
The shift from calendar-based maintenance to condition-based maintenance was the single biggest change in fleet operations during this period, and it happened largely because the data infrastructure built for ELD compliance turned out to be the foundation for predictive analytics.
2023-2026: The platform era and the mixed-fleet problem
The current phase is about consolidation and integration. Fleet operators have gone from having no data to being overwhelmed by it. A typical large fleet in 2025 might be running telematics from one vendor, dash cams from another, ELD software from a third, maintenance scheduling from a fourth, and fuel cards from a fifth. None of these systems talk to each other natively.
The platforms winning market share right now are the ones that integrate hardware, AI, and operational tools into a single system, so fleet managers aren’t toggling between six dashboards to make one decision. The industry term is “fleet operating system,” and it reflects the recognition that tracking, maintenance, safety, compliance, and fuel management are all branches of the same data tree.
The other defining challenge of this era is mixed fleets. Over 83% of new vehicles now ship with OEM-embedded telematics, meaning the connectivity is built in at the factory. But most commercial fleets aren’t composed entirely of brand-new trucks. They’re a mix of 2018 models with aftermarket GPS devices, 2021 models with basic OEM connectivity, and 2024 models with full embedded telematics. Managing data across three generations of hardware and software is the unsexy but critical problem that most fleet technology companies are working on right now.
At the same time, electric vehicles are adding another data layer. EV fleet management isn’t just GPS tracking with a battery instead of a diesel tank. It requires monitoring cell-level battery health, predicting range under real operating conditions (not lab estimates), managing charging schedules around route demands, and tracking energy costs that vary by time of day and utility rate. Platforms like Intangles that were built around predictive analytics and component-level monitoring are in a strong position here, because the algorithmic challenge of predicting battery degradation is structurally similar to predicting diesel engine wear. The data inputs are different, but the analytical framework is the same.
What actually changed, if you zoom out
The technology evolution over the past decade follows a pattern that applies to most industries: first you digitize the paper (ELDs replacing logbooks), then you connect the data (telematics integrating with engine diagnostics), then you make the data predictive (AI identifying failure patterns before they trigger fault codes), then you consolidate the tools into platforms that actually change how decisions get made.
Most fleets are somewhere between stage three and stage four. They’ve digitized. They’ve connected. Some have gone predictive. Very few have fully integrated their tools into a single operational workflow where the data automatically flows into decisions about maintenance scheduling, driver assignments, route optimization, and capital planning.
That’s the gap. And the fleets that close it first will operate at a cost structure and reliability level that competitors running fragmented systems can’t match. The technology to do it exists today. The bottleneck is implementation, not invention.
The fleet manager of 2015 looked at a GPS screen and made phone calls. The fleet manager of 2026 looks at an exception dashboard that surfaces only the trucks, drivers, and routes that need attention, with recommended actions already attached. Same title on the business card. Completely different job.
