At the Eurobike event in Frankfurt from June 24 to 27, Canyon Bicycles GmbH is showing its futuristic Predict bike prototype, a concept designed to radically improve rider safety and pack-riding performance. Its breakthrough intelligent system architecture is designed to see what riders don’t see.

Renowned as one of the most innovative brands in the industry, Canyon designs models for road, triathlon, gravel, mountain, city, trekking, and electric bikes. What started in founder and Executive Chairman Roman Arnold’s garage has grown into the world’s leading direct-to-consumer bicycle manufacturer. Canyon bikes are available at canyon.com, through its app, and at select stores worldwide.

Rather than relying on traditional, reactive safety measures, Canyon’s Predict system uses a 360-degree sensor array to anticipate road hazards and other road users, track group-ride dynamics, advise on cornering speeds, and predict tricky surface conditions before riders notice them.

The technology is being packaged into a premium road bike with a data screen integrated into the handlebar, and the Predict bike can connect to Canyon’s augmented-reality Stingr Smart helmet with its drop-down visor and data visualization screen (see our coverage of the helmet here).

While safety tech in the automotive world has skyrocketed, cycling safety has lagged behind, according to the company. From drivers unaware of cyclists and riders riding an unsafe distance from vehicles in front, to unsafe road surfaces and chaos in the pack, it doesn’t take much to have an accident. Canyon says its new system aims to bridge that gap without ruining the pure cycling feel of a high-performance road bike.

“We considered the numbers of people killed or seriously injured while riding, or the numbers who simply don’t cycle because they don’t feel safe, and we asked ourselves what we could do to address this problem,” said Fedja Delic, Canyon’s Head of Design. “Cars have become inherently safer and motorist deaths over the last ten years have fallen, but bicycles have not seen any significant safety improvements. In fact, the proportionate and absolute number of cyclists killed or seriously injured is shown to be rising in many countries. While technology has made driving a car safer than ever before, riding a bike on the road has arguably become more dangerous than ever before. Yet with the technology available, significant bicycle safety improvements are more than possible.”

The intelligent safety system is said to transform bicycle safety from reactive to predictive by continuously perceiving the environment, understanding context, and anticipating hazards in real time.

“Road cycling needs a safety revolution,” said Mazen Jrab, Canyon’s IoT Hardware Lead. “With Canyon Predict we are transforming safety from being reactive to predictive.”

It combines 360° multi-modal sensing—camera, radar, and other distributed sensors—with on-device communications, computer, and AI processing.

By integrating a multi-dimensional motion sensor in the DT Swiss wheel hub, precise information on bike motion is provided for rider assistance and information systems. It features 6D inertial measurement sensors—accelerometer and gyroscope with battery and Bluetooth LE modules—and measures longitudinal, lateral, and upward movement including rotation for comfort and safety features such as emergency brake assist, active curve assist, lane keep assist.

A key enabler of the bike’s innovations is its Cognitive Core, an on-bike processing unit that executes a multistage Al pipeline, breaking down the complex Al task into sequenced, ordered sub tasks in a chain. It combines an LVM (large vision model) with advanced tracking algorithms for real-time object detection, tracking, and decision-making without cloud dependency.

Through the fusion of data from the sensors and integration of rider dynamics such as speed, steering, and stability, the system builds a situational model that goes beyond surrounding traffic to eliminate blind spots and remove internet dependency, enabling instant, privacy-preserving decision-making. Rather than relying on cloud computing—which introduces latency and privacy risks—the platform processes data entirely on the bike via edge AI.

It predicts future trajectories of both the rider and nearby objects, before assigning risk scores, and communicating them through intuitive feedback including directional lights, haptics, and display guidance. Onboard displays and visual cues include critical warnings about a bike and its surroundings, such as prediction, distance, terrain, and group-ride assistance, as well as the potential for community or “swarm” intelligence when multiple users are riding together.

By integrating real-time perception with rider dynamics such as speed, steering angle, and stability, the system also aims to improve control in critical situations, while ultimately minimizing both the likelihood and severity of accidents through timely guidance and interventions.

In critical situations, the rider can intervene through more than just slamming on the brakes. Adaptive hardware such as allowing the rider to remotely drop the seat post can lower the rider’s center of gravity, improve their stability, and ultimately increase their control before a crash can happen.

“I train and race on the road a lot, and there are plenty of times where data about my bike and any other safety measures about other road users would be welcome,” said Kasia Niewiadoma-Phinney, Tour de France Femmes winner and Canyon//SRAM Racing’s Polish cyclist. “Improving how safe you feel on the road and helping you react to changing circumstances benefits everyone. With this sort of new technology, it makes me eager to see where it can go in reality and what the next generation of bikes can deliver to the rider.”

To see Predict technology in action, visit the Canyon stand at Eurobike in Messe Frankfurt and/or check out this YouTube video.