Connected vehicle data was supposed to be the new oil. Then the two companies built to refine it ran out of road: Wejo declared bankruptcy in 2023, and Otonomo, valued at $1.4 billion at its 2021 debut, was folded into a roadside-assistance company for $270 million the next year. The data itself was never the problem. The market to buy and sell it was simply smaller than anyone promised. That history shapes everything about vehicle telemetry data today, both where you can get it and whether buying it is the right move at all. This guide covers the examples, the providers worth knowing, the public datasets, and the option the marketplaces’ collapse made obvious.
What is vehicle telemetry data?
Vehicle telemetry data is the stream of measurements a vehicle records about its own location, motion, and mechanical state and transmits for remote monitoring and analysis. It is the raw material behind fleet management, usage-based insurance, predictive maintenance, and connected-car services, and it ranges from a simple GPS ping to the full sensor feed of an autonomous vehicle.
Examples of vehicle telemetry data
At its most basic, vehicle telemetry is location and motion: GPS position, speed, heading, and trip distance and time. Add the powertrain and it gets richer, engine RPM, throttle position, fuel level or EV state of charge, battery voltage, coolant temperature, and odometer reading. Layer in behavior and safety signals, harsh braking and acceleration, idling, seat belt use, and you can describe how a vehicle is being driven, not just where. Diagnostic trouble codes report faults as they happen, tire-pressure and emissions data support compliance, and at the far end of the spectrum an autonomous vehicle adds camera, LiDAR, and radar streams that dwarf everything else. A modern connected car funnels somewhere between 3 and 40 GB a day from more than 200 onboard sensors.
Vehicle telemetry data providers
Buying vehicle telemetry data is harder than it looks, which is the lesson of the marketplace era. The pure-play data exchanges that aggregated OEM feeds for resale, led by Wejo and Otonomo, could not monetize the data fast enough to survive, so the providers that remain tend to offer access tied to a clear use case rather than a raw firehose.
Car API platforms are the most developer-friendly route. Smartcar, for example, offers consent-based access to vehicle data across dozens of makes through a single integration, reaching well over a hundred million connected cars, with the vehicle owner granting permission for each application.
OEM data programs are the source of record. General Motors, Ford, and Stellantis, through its Mobilisights unit, license connected-vehicle data directly, which is the most authoritative supply but comes with the access terms and pricing of a manufacturer.
Fleet telematics vendors are a practical source if you operate vehicles. Geotab, Samsara, and Verizon Connect collect telemetry from the devices they install, and the data belongs to the operator running the fleet.
Mobility and location-analytics firms package telemetry into insight rather than raw records. HERE, INRIX, TomTom, and StreetLight Data turn aggregated vehicle movement into traffic, routing, and planning products, and in Europe neutral-server providers like High Mobility and Caruso broker OEM data under the region’s access rules.
Vehicle telemetry datasets
For research and machine learning, public datasets are often the better starting point than a commercial feed, because they are labeled and free to use. The Waymo Open Dataset and nuScenes are the standards for autonomous-driving sensor data, pairing camera, LiDAR, and radar with annotations. KITTI remains a long-running benchmark for perception tasks, Argoverse adds motion forecasting and HD maps, and BDD100K offers a large, diverse set of driving video. For lighter-weight telemetry, comma2k19 captures real highway driving from a consumer device, and Audi’s A2D2 provides multimodal sensor data with labels. These datasets are built for training and benchmarking models, not for operational analytics, which is the distinction that decides whether a dataset or a provider fits your need.
The other option: generate and control your own
The collapse of the data marketplaces pointed at something the hype obscured. Most of the organizations that need vehicle telemetry data, automakers, fleets, and insurers, already generate it themselves, in volumes that dwarf anything they would buy. For them the question was never which provider to use. It is how to collect, process, and control their own telemetry without drowning in it.
That is a different problem from shopping at a marketplace, and it is where Expanso fits. Processing telemetry at the edge, inside the vehicle or the gateway, lets a team filter and summarize the raw stream at the source, so they keep the signal worth keeping and skip the cost of moving and storing the rest. One automaker did exactly this across 2.3 million vehicles and cut transmitted data by 94%.
It also addresses the issue that quietly sank the marketplaces: ownership and privacy. Vehicle telemetry is dense with location and behavior, so enforcing governance at the source, masking and applying policy before the data moves, is what makes first-party telemetry usable without inheriting a compliance liability. Expanso is not a data provider or a marketplace; it is the automotive data layer that turns the telemetry you already produce into something you can actually use.
Frequently asked questions
What is vehicle telemetry data? It is the data a vehicle records about its own location, motion, and mechanical condition, then transmits for remote monitoring. Examples include GPS position and speed, engine RPM, fuel or battery level, harsh-driving events, and diagnostic fault codes, and in autonomous vehicles it extends to camera, LiDAR, and radar feeds.
Where can I get vehicle telemetry data? Through car API platforms like Smartcar, directly from OEM data programs such as GM, Ford, and Stellantis Mobilisights, from fleet telematics vendors like Geotab and Samsara if you run vehicles, or from mobility-analytics firms like HERE and INRIX. For research, public datasets are often the best free starting point.
What are the best vehicle telemetry datasets for machine learning? The Waymo Open Dataset and nuScenes are the leading autonomous-driving sensor datasets, KITTI and Argoverse are widely used perception and forecasting benchmarks, BDD100K offers diverse driving video, and comma2k19 and Audi’s A2D2 provide real-world telemetry and multimodal sensor data.
What happened to Wejo and Otonomo? Both were prominent connected-vehicle data marketplaces that went public via SPACs in 2021. The market for buying and selling vehicle data proved far smaller than expected, and Wejo declared bankruptcy in 2023 while Otonomo was acquired by Urgent.ly in a 2024 reverse merger, effectively ending the pure-play data-marketplace model.
How is vehicle telemetry data different from telematics data? The terms overlap heavily. Telematics is the broader technology of collecting and transmitting vehicle data, while telemetry data refers specifically to the measurements being transmitted, the location, motion, and mechanical readings the telematics system carries.
The data was never the scarce part
The story of vehicle telemetry data is that the supply was always enormous and the hard part was everything else: moving it, storing it, governing it, and turning it into something worth paying for. The marketplaces bet that buyers would pay for raw data and learned that they mostly would not. The organizations getting value today are the ones treating telemetry as something to control at the source, whether they generate it or buy it, rather than something to hoard and hope to monetize later.
Producing your own vehicle telemetry and trying to make it usable? See how Expanso processes automotive data at the edge. The data was never scarce. Control over it is the part worth having. Or book a data consultation to talk through your telemetry.
