robotindex Waveshare JetRacer
A pre-assembled hardware kit built to run NVIDIA's own open-source JetRacer software — Ackermann front-steering, a Jetson Nano carrier board, and a CSI camera, with GPU-accelerated TensorRT road-following out of the box. This is the base "JetRacer AI Kit"; Waveshare also sells a faster Pro variant and a ROS-equipped Professional kit with lidar and an IMU. Every SKU runs exclusively on the Jetson Nano, which NVIDIA is now discontinuing — a real caveat worth knowing before buying.
waveshare.com/jetracer-ai-kit →
Image: Waveshare (official)
Specs
Battery
3× 18650 Li-ionsold separately; onboard charge/discharge protection
Weight
Not publishedno official figure found in this research
Dimensions
Not publishedno official figure found in this research
Connectivity
AC8265 Wi-Fi + BTdual-band combo module; two antennas, one radio chip
Compute
Jetson Nano B01 (4GB)2GB Dev Kit variant also sold; no Orin option exists
Camera
Sony IMX219, CSI8MP, 160° wide-angle — same sensor as Pi Camera v2
Software
Jupyter + PyTorch/TensorRTGPU-accelerated on-device inference via torch2trt
Steering
Ackermann, frontservo-actuated front axle, rear-wheel drive
Motor
37-520 DC gearmotor1:10 reduction, ~740 RPM; Pro kit uses faster RC380 motors
Compute Support
Jetson Nano is EOLNVIDIA discontinuing; module support only through Jan. 2027
Reference Price
Waveshare's own listings for the base AI Kit range from $213.99 to $632.99 depending on bundle configuration (with or without the Jetson Nano board, battery, and gamepad); the Pro AI Kit runs $255.99–$685.99, and the ROS Professional kit (adding an RPLiDAR A1 and IMU) runs $259.99–$694.99. Because so much of the price swings on which accessories are included, a fair comparison against other kits in this catalog requires checking exactly what's bundled, not just the headline number.
Insights
JetRacer's onboard inference is genuinely GPU-accelerated rather than a marketing claim: training happens off-device in PyTorch, but inference runs optimized through TensorRT directly on the Jetson Nano's GPU — a real capability edge over CPU-only rival kits. Every JetRacer SKU depends on the now-EOL Jetson Nano, which NVIDIA is discontinuing with module support only through January 2027.
Firmware & Community
waveshare/jetracer
Waveshare's own GitHub fork of NVIDIA's original open-source JetRacer project, adapted for its pre-assembled AI Kit hardware.
Verified against Waveshare's official JetRacer AI Kit, Pro AI Kit, and ROS AI Kit product pages and wiki; the original NVIDIA-AI-IOT/jetracer GitHub repository and Waveshare's own fork; NVIDIA's blog and Hackster's coverage of Jetson Nano cars at DIY Robocars events; NVIDIA developer-forum threads confirming both the Jetson Nano Developer Kit's discontinuation and community use of this specific hardware; Tianbot's separate same-named project (for the naming-disambiguation note); and Amazon/AliExpress/RobotShop listings for pricing and bundle-configuration variance. Flagged as uncertain: exact current DIY Robocars league activity level for Nano-class hardware in 2026 (historical evidence is solid, current activity is not confirmed); and precise like-for-like pricing against PiCar-X and ROSMASTER, since JetRacer's price swings heavily by bundle configuration in a way that resists a single clean number.