Closing a critical safety gap in autonomous driving. By Stefan Hamerich, Senior Director Product Management August 17, 2026 Autonomous mobility is on the rise. Robotaxi companies like Zoox and Waymo are serving hundreds of thousands of rides each week, autonomous trucking is advancing toward widescale commercial deployment, and automated driving systems continue to expand across passenger vehicles and commercial fleets. As these technologies mature, attention is increasingly turning to one question: how safely can autonomous vehicles (AVs) operate in complex, real-world conditions? One of the most important aspects of this conversation is how AVs recognize and respond to emergency vehicles. To help address this challenge, Cerence AI developed Emergency Vehicle Detection (EVD), an audio AI solution that detects approaching emergency vehicles and alerts either the driver or the vehicle system. The Safety Challenge Emergency situations are rarely straightforward. An ambulance may be approaching from behind in heavy traffic. A fire truck may be navigating through a congested urban corridor. First responders may be managing an active scene using emergency lights, traffic cones, and other signals. In July, the United States National Highway Traffic Safety Administration (NHTSA) issued a public call to action urging autonomous vehicle developers to place greater focus on interactions between AVs and first responders. The agency highlighted repeated instances where AVs interfered with emergency operations. The implications extend beyond autonomous driving. Whether operated by a human or an automated system, a car that fails to recognize an approaching emergency vehicle can delay response times, create confusion in already stressful situations, and introduce unnecessary safety risks. As automated mobility continues to scale, emergency vehicle awareness will become an increasingly important component of the overall safety stack. Why Audio Matters Many autonomous driving conversations focus on what a vehicle can see. Cameras, radar, and other sensors provide essential information about the surrounding environment. Yet emergency vehicles are often heard before they are seen. Traffic, buildings, weather conditions, and other obstacles can limit visibility. Sirens are specifically designed to provide advance warning before an emergency vehicle reaches a driver's field of view. Those additional seconds give drivers or automated systems valuable time to make decisions that can impact safety-critical outcomes. Audio offers a unique and important layer of environmental awareness that complements other vehicle sensing technologies. Enabling Better Situational Awareness for Autonomous Vehicles Cerence EVD is certified in 47 countries and supports more than 1,500 siren sounds. It is designed to recognize a broad range of emergency vehicles, including ambulances, police cars, and fire engines. It can identify sirens from more than 300 meters away using vehicle-integrated microphones. When external microphones are used, detection distances can extend beyond 800 meters. External microphone configurations also support directional awareness, helping determine where the siren is coming from. Using advanced audio processing and AI-based classification, EVD can distinguish emergency sirens from surrounding sounds such as traffic noise, music, conversations, and other common driving distractions. Proven in Today's Vehicles and Tomorrow's Mobility Platforms Emergency vehicle detection is a real-world challenge that requires solutions capable of operating across countries, vehicle types, roadway conditions, and regulatory environments. Cerence EVD is already being deployed across vehicle programs and mobility platforms. Mercedes-Benz and BMW have integrated EVD as part of their automated driving initiatives, and the technology is being applied across use cases ranging from passenger vehicles to autonomous bus deployments. Cerence also works with customers across emerging mobility categories, including robotaxis and autonomous trucking. Looking Ahead As autonomous mobility continues to evolve, emergency vehicle awareness will become a foundational capability. Vehicles must be able to understand more than lane markings, traffic signals, and nearby objects. They also need to recognize critical situations developing beyond their immediate field of view and respond appropriately when every second counts. Cerence EVD has an important role to play in that future. Serving as a vehicle's "ears," it helps drivers recognize and respond to emergency vehicles earlier, supporting safer roads for drivers, passengers, first responders, and the communities they serve. To learn more about Cerence Emergency Vehicle Detection and how it can support your mobility platform, contact the Cerence AI team or visit the EVD page.

Cerence reported fiscal third-quarter 2026 results highlighting a 20% year-over-year increase in recurring connected services revenue. The automotive AI company now has approximately 100,000 vehicles running its xUI platform, with management targeting a couple million by the end of fiscal 2027. The company secured a major xUI contract with Stellantis and its first mobile work agent customer. Variable license revenue faced headwinds from unfavorable regional production mix and difficult year-over-year comparisons. Cerence is expanding into non-automotive sectors, including industrial operations and robotics. Management expects high-single to low-double digit revenue growth in fiscal 2027, driven by xUI scaling and non-automotive initiatives. The board authorised the company's first share repurchase programme of up to $30 million. Free cash flow guidance for fiscal 2026 was raised to $76 million to $82 million.