Sunnyvale, CA-based Sonatus today announced what it says is a game-changing platform that enables OEMs to deploy AI at the vehicle edge. Its Sonatus AI Director provides OEMs and suppliers with an end-to-end toolchain for model training, validation, optimization, and deployment while integrating with vehicle data, executing models in isolated environments, and providing cloud-based remote monitoring of model performance.

Automotive AI is growing rapidly, projected to reach a market size of $46 billion annually by 2034, according to Precedence Research, and Sonatus says that in-vehicle edge AI software and services will be an increasingly important component. Its toolchain and in-vehicle runtime environment lower the barriers to edge AI adoption and innovation compared to today’s siloed approach using disparate ML development (MLops) tools, reducing effort from months to weeks or days.

“Artificial intelligence is creating opportunities for new ideas that were never before possible in vehicles,” said Jeff Chou, CEO and Co-founder of Sonatus. “With Sonatus AI Director, we are empowering OEMs to deploy AI algorithms of all types into vehicles easily and efficiently, unlocking new categories and opening up an ecosystem of innovation that connects cloud, silicon, Tier-1 suppliers, and AI model developers.”

With the new offering, Sonatus says that vehicle OEMs can gain a consistent framework that enables them to deploy models from different vendors with a single platform and across vehicle models. Tier-1 suppliers can optimize the systems they deliver to OEMs and more easily leverage AI across hardware and software technologies. Silicon providers can help their customers take full advantage of the compute and AI acceleration capabilities their chips offer. AI model vendors gain access to the needed input data from across different subsystems while protecting the intellectual property of their models.

In-vehicle edge AI, fueled by real-time and contextual vehicle data, allows OEMs to unlock new features and capabilities that enable adaptive and personalized driving experiences, proactive maintenance, improved efficiency, and optimal vehicle performance. Instead of relying solely on cloud-based models, Sonatus AI Director lets vehicle makers run AI directly in the vehicle, providing faster response, reducing data upload costs, preserving data and algorithm privacy, and ensuring continuity across intermittent connectivity.

Rather than waiting for next-generation ECU hardware, Sonatus says that OEMs can use its new solution to maximize the value of their existing compute resources, accelerating time-to-market while also providing a path to scale AI performance as new silicon becomes available. It supports a range of model types, including physics- and neural network-based models, as well as small and large language models, catering to diverse vehicle use cases. An OEM can easily manage and deploy a diverse set of AI models spanning many vehicle subsystems, realizing benefits that include cost, performance, security, and efficiency improvements.

 

Launch partners and use cases

Initial launch partners include leading automotive silicon provider NXP, compute IP leader Arm, and cloud service provider leader AWS.

“Leveraging the platform helps us to offer our customers faster integration, enhanced data access, and an edge AI infrastructure,” said Robert Moran, VP and GM Automotive Processors at NXP.

Sonatus AI Director, pre-integrated with NXP eIQ Auto ML software and tailored for NXP silicon, brings an automotive-grade AI deployment framework to NXP customers, helping accelerate the innovation cycle.

“The new Sonatus AI Director is a powerful example of how Arm is enabling edge AI today by combining Sonatus’s dynamic software architecture with the performance and efficiency of Arm-based platforms—like the new Arm Zena Compute Subsystems—to deliver the next generation of automotive intelligence and real-time decision making,” said Suraj Gajendra, VP of Product Solutions and Software, Automotive Business at Arm.

AWS provides integration with Sonatus AI Director, enabling users to collect vehicle data through Sonatus Collector AI, seamlessly transfer the telemetry to AWS SageMaker Unified Studio for advanced data preparation and model training, then deliver to Sonatus AI Director for final optimization and policy integration into production-ready vehicle model bundles. The unified workflow accelerates the model development lifecycle from data collection to deployment.

A range of subsystem expert model vendor launch partners, including Compredict, Qnovo, Smart Eye, and VicOne have seen these benefits in their use cases.

Compredict’s AI-based virtual headlight leveling sensor reduces bill of materials cost by eliminating hardware components, empowering OEMs to achieve full 2027 UN R48 compliance with a 100% software approach.

“Integrating our AI-driven Virtual Headlight Leveling Sensor with the Sonatus AI Director platform enables us to reach OEMs more quickly, scale our models while safeguarding our intellectual property, and optimize performance across various ECUs—all while delivering OEMs up to $20 in bill of materials cost savings per vehicle,” said Stefan Hassels, Head of Product at Compredict.

Integrated with Sonatus’s platform, Qnovo’s AI-powered Health & Safety Diagnostics (HSD) enables deployment anywhere in the vehicle or cloud, creating a battery management solution that adapts to specific vehicles, drivers, and environmental conditions.

“Our AI-enhanced Health and Safety Diagnostics can now be deployed anywhere in the vehicle or cloud, adapting to specific drivers and environmental conditions to deliver smarter, safer battery management with dramatically faster time-to-market,” said Nadim Maluf, CEO of Qnovo.

With Sonatus AI Director, SmartEye cabin monitoring systems can more easily customize these alerts based on holistic driver behavior by combining distraction model outputs with data from other vehicle subsystems.

“Combining our distraction detection capabilities with data from across the vehicle makes it possible to deliver safety interventions that are both more intelligent and more responsive to each driving situation,” said Martin Krantz, CEO of Smart Eye.

With dynamic model scheduling and various in-vehicle data collected by Sonatus AI Director, the VicOne xCarbon Edge AI, a GenAI-based in-vehicle intrusion detection system, can accurately infer security risks and run compute-intensive AI models even on deployed hardware.

“This joint innovation enables OEMs and Tier-1s to minimize cloud dependency, improve data access, and accelerate their response to emerging cyber threats across connected fleets,” said Max Cheng, CEO of VicOne.

 

Delivering SDV building blocks

According to Sonatus, successful in-vehicle edge AI is enabled by the capabilities of SDVs (software-defined vehicles). The vehicle software company is focused on accelerating the transition towards AI-enabled SDVs with the Sonatus Vehicle Platform.

The company has delivered key SDV building blocks from the cloud to the vehicle edge, including on-demand access to vehicle data, a critical foundational element in this innovation. Its products, such as Sonatus Collector AI and Sonatus Automator AI, can complement Sonatus AI Director and further extend the capabilities of each. Models managed by Sonatus AI Director can enhance the capabilities of Sonatus AI Technician, such as building in detectors for various types of anomalies that may assist with vehicle diagnostics.

The company will demonstrate Sonatus AI Director publicly for the first time at the IAA Mobility conference in Munich, Germany, later this month.