Autonomous cars generate more than 300 TB of data per year

The accessibility of sensors and camera modules is making the car industry increasingly data-driven. When combined with state-of-the-art software and advanced computing, data transforms into decisions inside autonomous cars, and as autonomous vehicle technologies evolve, the data generated inside cars keeps growing. Depending on its sensor setup, an autonomous car can generate between about 1.4 TB and 19 TB of data for every hour of driving, which works out to roughly 500 TB to 7,000 TB a year for an average US driver.

Key takeaways

• An autonomous car can generate between about 1.4 TB and 19 TB of sensor data per hour, depending on how many cameras, radar units, lidar units, and other sensors it carries.
• Cameras produce most of that data, at 500 to 3,500 Mbit/s each, while radar and ultrasonic sensors add comparatively little.
• At the 367 hours a year an average US driver spends behind the wheel, that adds up to roughly 500 TB to 7,000 TB of data per car per year.
• SAE International defines 6 levels of driving automation, and Level 4 robotaxis now operate commercially while most consumer cars remain at Level 2.
• Not all sensor data stays in the car, but the data stored on board remains integral to vehicle functions, which makes in-vehicle storage a core design challenge for carmakers.

SAE levels of driving automation

Not all autonomous vehicles have the same level of automation. SAE International, a US-based association that develops standards for automobiles, created a 6-tier system in its J3016 standard that defines the levels of driving automation:

Most cars sold with driver assistance today are still at Level 2, where the driver has to supervise at all times. Level 3 has reached the road in limited form. Mercedes-Benz DRIVE PILOT is approved in Germany for speeds up to 95 km/h, and in December 2025 China granted its first Level 3 permits for pilot use on set expressway sections in Beijing and Chongqing. Level 4 is already a commercial reality in robotaxi fleets: Waymo runs around 4,000 driverless vehicles in 15 US cities and gives about 500,000 paid rides a week. No car on the market offers Level 5.

More sensors equals more data

Today, even at lower levels of autonomy, connected cars generate around 25 GB of data per hour. As more self-driving features appear inside connected cars, the architecture required to make it all possible becomes increasingly complex. This directly correlates to the number of sensors needed for an autonomous system to operate. While the number of sensors has been rapidly increasing, this rate of growth may not stay the same.

It is also important to point out that not all sensors are the same. A broad spectrum of sensors exists, each having a special purpose and quantity in a car. Depending on the sensor setup, the total amount of data generated can vary substantially. Stephan Heinrich, then at Lucid Motors, presented these estimates of sensor-generated data at Flash Memory Summit 2017:

Combined, the total bandwidth can reach up to 40 Gbit/s (about 19 TB/h). Even the lowest figure of 3 Gbit/s (about 1.4 TB/h) is a very substantial amount of data to maintain. At the top rate, a basic laptop with 240 GB of storage, enough for around 30 DVD movies, would run out of capacity in less than a minute. A phone with 32 GB of storage would be full in under 7 seconds, assuming its flash storage could even write data at the required speed.

How much data does an autonomous car generate in a year?

On a yearly estimate, the amount of data is even more staggering. According to the AAA Foundation for Traffic Safety, US drivers spent an average of 367 hours behind the wheel in 2024. Combined with the sensor data estimates above, one car could produce between roughly 500 TB and 7,000 TB of data in a single year. Autonomous and ADAS test cars, which record every sensor stream during development, generate even more per day.

However, not all of that sensor data will necessarily be stored on the car. More and more of it may be sent to the cloud, but the data stored on the car itself will remain integral to the functions of the vehicle.

White paper

The impacts of file system fragmentation on automotive storage performance

How file system fragmentation affects read and write performance and storage lifetime in vehicles, and how carmakers and suppliers can reduce its effects.

Read the paper

Final thoughts

As cars with higher autonomy levels are released to the market, cars need to handle substantially bigger amounts of data than ever before. Processing that data in a fast and seamless way is now, and will continue to be, one of the main challenges for carmakers. To tackle these requirements, carmakers are exploring both hardware- and software-based data storage solutions that can handle this data stream in the most efficient package possible. Achieving this will be an important milestone in the development of production-ready fully autonomous vehicles.

*This article was originally published in 2017, and was authored by Stan Dmitriev. It was updated in 2021 by Simon Wright, and again in October 2026.