Sonatus AI Director Powers In-Vehicle Edge AI at Scale

The cloud-based artificial intelligence (AI) systems that support software-defined vehicles today are important, but there are practical limitations to running every workload in the cloud, including responsiveness, data volume, data privacy, and dependable connectivity.

Together, these factors render current solutions insufficient to address all latency-sensitive or mission-critical in-vehicle applications. The solution is to deploy AI capabilities within the vehicle to ensure real-time decision-making, adaptive behaviors, and optimized performance.

To help the auto industry meet the challenge of moving AI to the vehicle edge, Sonatus has developed Sonatus AI Director.

This platform provides OEMs and their suppliers with an end-to-end toolchain for model training, validation, optimization, and deployment, while seamlessly integrating with vehicle data, executing models in isolated environments, and providing cloud-based remote monitoring of model performance.

As a comprehensive toolchain and in-vehicle runtime environment, Sonatus AI Director lowers 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, according to the company.

Automotive Industries (AI) asked John Heinlein, Ph.D., Chief Marketing Officer at Sonatus, how the company addresses the challenges of deploying AI at the vehicle edge.

Heinlein: Today’s in-vehicle environment is exceptionally complex: Sonatus AI Director allows AI to be deployed in more vehicles, more efficiently, securely, and at scale. Our goal was to provide clear benefits to every layer of the automotive value chain.

For vehicle manufacturers, it simplifies the deployment of AI models from multiple vendors across different vehicle models, reducing fragmentation and complexity.

Tier-1s can optimize the systems they deliver and apply AI more easily across hardware and software.

Silicon providers can help their customers take full advantage of the compute and AI acceleration capabilities of their chips.

Finally, suppliers and AI model vendors gain secure access to the input data they need from across vehicle subsystems while also protecting the intellectual property of their models.

AI: Will the system work on existing hardware platforms?

Heinlein: Yes, and in fact, that is one of the most important benefits!

A common misconception within the automotive industry is that AI’s utility is limited to high-end vehicles equipped with expensive GPUs or specialized AI accelerator hardware.

Similarly, many believe AI’s application is confined to ADAS, autonomous driving, or IVI chatbots. These narrow perspectives overlook a significant opportunity for AI to benefit a wider range of vehicles and applications.

Sonatus AI Director empowers OEMs to leverage existing compute resources for AI, eliminating the need to wait for next-generation ECUs or invest in costly AI accelerators. This accelerates time-to-market and provides a clear upgrade path as new silicon emerges.

Furthermore, Sonatus AI Director’s ability to run lighter AI models on current hardware extends its applicability across diverse vehicle price points, even benefiting many lower-end models with more modest compute capabilities.

AI: What challenges is Sonatus AI Director solving that other platforms and tools in the market are not?

Heinlein: Most MLOps platforms were built for the cloud. Even those that address edge devices tend to either be focused on vendor-specific silicon or encompass a wide range of industrial, IoT, and, in some cases, automotive targets.

Sonatus AI Director was built for the vehicle. It’s an end-to-end toolchain that understands the realities of automotive environments — hardware limits, safety-critical systems, and fragmented data.

It enables training by curating the right model-specific data, optimizes for vendor-specific silicon, manages deployment across diverse ECUs spanning multiple vehicle domains, and monitors performance in real time. That edge-awareness is what shortens time to market and lowers cost compared to generic stacks.

AI: Does it create opportunities for additional revenue streams?

Heinlein: This can be viewed from two perspectives. The first, from OEMs and model providers: simplifying the integration of AI into vehicles opens new avenues for innovation and revenue.

OEMs can achieve ongoing vehicle optimization and introduce new functionality, while specialized model vendors can more readily introduce their expertise into the automotive sector.

This collaborative approach allows both OEMs and model vendors to leverage advanced computing capabilities, potentially unlocking new revenue streams.

The second perspective, from Sonatus, is that our objective is to foster collaboration rather than accrue profits from our model partners. The purpose of Sonatus AI Director is to streamline the integration process between model developers and OEM/Tier-1 subsystem manufacturers, ensuring a smooth, efficient, and barrier-free adoption.

AI: How does Sonatus AI Director integrate with the AI Model Partners?

Heinlein: Model vendors face the challenge of deploying their models into the vehicle and accessing the needed input data from across different subsystems.

AI Director provides an intuitive framework to connect model inputs with vehicle data sources, which simplifies the integration burden for the OEM and model vendor alike.

In addition, by deploying models through isolated containers, Sonatus AI Director protects the intellectual property of model vendors, allowing for more flexible business models in which hardware and software can be more easily decoupled for faster innovation.

The combination of these two allows model vendors and OEMs to work together more quickly and easily, reducing integration costs.

AI: Where to now for Sonatus and in-vehicle AI?

Heinlein: It has been an exciting time for us, but we are just getting started!

In December 2024, we added Generative AI capabilities to our Sonatus Collector AI and Sonatus Automator AI products to significantly improve developer productivity and democratize access to vehicle data across more OEM groups.

We also announced our cutting-edge solution for improving the vehicle service experience with AI, called AI Technician, which uses an OEM knowledge base coupled with superior vehicle data to revolutionize vehicle quality and improve both the driver’s service experience and OEM service revenue.

With Sonatus AI Director, we now add the ability to deploy AI models into the vehicle edge, which can further enhance all three of those existing products.

So, while customers are free to adopt each of our products individually to suit their unique needs, the combination of them is greater than the sum of the parts.

Stay tuned, we have more innovative ideas in the works to continue accelerating the shift to AI-enabled software-defined vehicles!

Launch partners

Arm

“As vehicles become increasingly defined by AI, we’re seeing demand for intelligent processing at the edge,” said Suraj Gajendra, VP of product solutions and software, Automotive Business at Arm.

“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.”

AWS

AWS provides seamless integration with Sonatus AI Director.

This powerful combination enables 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 effortlessly deliver to Sonatus AI Director for final optimization and policy integration into production-ready vehicle model bundles.

This unified workflow dramatically accelerates the entire model development lifecycle from data collection to deployment.

NXP

Seamless Edge AI Enablement with Pre-Integrated Silicon and ML Toolchain

Sonatus AI Director is pre-integrated with NXP’s eIQ® Auto ML software and S32 processing automotive-grade IP, including its high-performance networking accelerators (PFE and LLCE), to form a unified and production-ready toolchain for in-vehicle edge AI. This integration supports the entire AI lifecycle—from model development and optimization to deployment and runtime execution—on NXP’s suite of heterogeneous processors, including application cores, real-time cores, and NPUs.

The result is a seamless, silicon-aware workflow that allows OEMs to efficiently develop and scale AI workloads on NXP hardware platforms.

Key Benefits:

• Full-stack integration between Sonatus and eIQ Auto ML software streamlines model development, tuning, and deployment
• Optimized runtime execution across diverse NXP processors with performance-per-watt efficiency
• Built-in support for NXP networking IP, enabling edge-based data capture and real-time feedback loops
• Reduced time-to-integration for OEMs deploying AI on NXP-based ECUs
• Enhanced scalability and reuse of AI models across vehicle programs using NXP silicon

Sonatus AI Director aligns perfectly with NXP’s commitment to delivering scalable, safe, and secure system solutions for SDVs. 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 our customers, helping accelerate the innovation cycle. This collaboration aims to shorten development time, streamline deployment and improve data monitoring, taking another step forward in our commitment to making AI-enabled SDVs a reality.”

A jointly developed case study on the benefits of Sonatus AI Director and NXP technologies is available at https://www.sonatus.com/resources/ai-director-partner-case-study-nxp/

Smart Eye

“By integrating Smart Eye’s highly accurate Cabin Monitoring System with Sonatus AI Director, we’re giving OEMs the flexibility to go beyond fixed rules and tailor in-vehicle alerts to real-world driver behavior,” said Martin Krantz, CEO of Smart Eye.

“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.”

Model Partners

VicOne

xCarbon Edge AI for In-Vehicle Threat Detection

xCarbon Edge AI delivers advanced in-vehicle cybersecurity by analyzing CAN traffic, Ethernet packets, and system logs directly on onboard CPUs, GPUs, and NPUs — avoiding the limitations of cloud-only VSOCs. The system surfaces only critical alerts to the cloud for fleet-wide correlation.

Sonatus AI Director delivers containerized, isolated execution, flexible deployment across varied ECU types, standardized API-driven access to relevant vehicle data, and compute-aware model optimization — ensuring scalable, efficient model integration across OEM fleets.

Key Benefits:

• Up to 60% lower cloud data transfer and processing costs
• 80% faster triage with pre-filtered, high-relevance alerts
• Improved detection for individual vehicles and fleet-level threats
• Flexible scaling across diverse ECU and hardware platforms

Sonatus is a vehicle software company that is accelerating the transition towards AI-enabled software-defined vehicles.

The diverse products comprising the Sonatus Vehicle Platform serve as key building blocks that allow automotive companies and their ecosystem to fast forward to the future of mobility and deliver continuous improvements in cost, capabilities, reliability, and user experience over the vehicle lifespan.

The award-winning Sonatus software platform is in mass production in over 5 million vehicles from Hyundai, Kia, and Genesis. Sonatus is privately-held with world-class automotive, technology, and venture investors, including Foxconn, Hyundai Motor Group’s Kia Corporation, LG Electronics, Marvell, NEC and Translink Orchestrating Future Fund, SAIC Capital, Translink Capital, UMC Capital, and Wanxiang.

Sonatus worldwide headquarters is in Sunnyvale, CA (Silicon Valley), with regional headquarters in Dublin, and offices in Bangalore, Detroit, Frankfurt, Kraków, Paris, Pune, Seoul, Shanghai, Taipei, and Tokyo. For more information, visit www.sonatus.com.

Sonatus and the Sonatus logo are trademarks of Sonatus, Inc. that are registered in the United States and are registered or pending in other countries. © 2025 Sonatus, Inc.

COMPREDICT

Software-Based Headlight Leveling

COMPREDICT’s Virtual Headlight Leveling Sensor replaces physical ride height sensors with a fully software-based, AI-driven approach compliant with upcoming UN R48-09 regulations. Integrated with Sonatus AI Director, the solution leverages existing CAN signals for pitch estimation in both static and dynamic conditions, without the addition of physical sensors. AI Director enables flexible deployment across ECUs and architectures, standardized API-based data access, performance optimization for available compute, and secure containerized runtime isolation—simplifying OEM integration.

Key Highlights:

• Up to $20 per vehicle BOM savings by eliminating ride height sensors
• Full UN R48-09 compliance for static and dynamic leveling
• Faster integration — reduced from months to days
• Hardware flexibility — deployable across varied ECU architectures

“At COMPREDICT, we are redefining how software can replace physical vehicle components. 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.  

A jointly developed case study on the benefits of Sonatus AI Director and COMPREDICT virtual sensor technology is available at https://www.sonatus.com/resources/ai-director-partner-case-study-compredict/ 

Qnovo: Context-Aware Battery Health & Safety Diagnostics

Qnovo’s Health & Safety Diagnostics (HSD) software offers 98.7% accurate battery fault prediction by combining multi-metric diagnostics. Integrated with Sonatus’ AI Director, Qnovo’s HSD is AI-enhanced with real-time contextual data allowing for driver specific battery diagnostic profiles and alerts based on past and anticipated behavior. With this contextual awareness, the software adapts to specific vehicles, drivers, and environmental conditions to deliver smarter, safer battery management. AI-Powered HSD becomes easy to deploy outside of the BMS, anywhere in the vehicle or cloud.

Key Highlights:

• Analyzes 10-12 distinct battery health metrics for earlier fault detection
• Integrating battery health and usage data doubles model parameters, enabling deeper, more accurate health and safety insights
• Personalized diagnostics based on driving behavior, terrain, and climate conditions
• Reduces false positives and broader battery-related recalls by pinpointing the exact vehicles with problems
• Extends EV battery life and lowers total cost of ownership for fleets and consumers
• Deployed across millions of kilometers with zero safety incidents and OEM production underway

“Our partnership with Sonatus transforms how quickly OEMs can deploy advanced battery safety solutions, reducing integration time to days,” said Nadim Maluf, CEO of Qnovo.

“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.”

A jointly developed case study on the benefits of Sonatus AI Director and Qnovo technology is available at https://www.sonatus.com/resources/ai-director-partner-case-study-qnovo/ 

Case studies

Field trials and implementations demonstrate the power of the Sonatus AI Director tool across silicon providers and AI model vendors:

Accelerating intelligent, secure, real-time decisions at the edge with NXP scalable AI solutions and Sonatus AI Director

Sonatus AI Director, along with NXP eIQ® Auto ML software, provides OEMs and suppliers with an end-to-end toolchain for model training, validation, optimization, and deployment, while seamlessly integrating with vehicle data. It allows the execution of ML models in isolated environments of S32 platforms, while also providing cloud-based, fleet-wide remote monitoring of model performance.

Sonatus AI Director is pre-integrated with NXP’s eIQ Auto ML software and S32 processing automotive-grade IP, including its high-performance networking accelerators, to form a unified and production-ready toolchain for in-vehicle Edge AI.

This integration supports the entire AI lifecycle—from model development and optimization to deployment and runtime execution—on NXP’s suite of heterogeneous processors, including application cores, real-time cores, and NPUs. The result is a seamless, silicon-aware workflow that allows OEMs to efficiently develop and scale AI workloads on NXP hardware platforms.

Qnovo AI-Enhanced Battery Safety on Sonatus AI Director

Qnovo’s Health & Safety Diagnostics (HSD) software achieves 98.7% accuracy in battery fault prediction using advanced modeling and machine learning.

When deployed via Sonatus AI Director, the combination enables rapid, flexible deployment and provides the model with rich contextual data—such as driving conditions and environmental inputs—to improve diagnostics beyond traditional BMS systems.

This partnership accelerates the delivery of advanced battery safety technology, reduces engineering costs, and caters to various OEM needs without custom development.

VicOne GenAI-based In-Vehicle Intrusion Detection System on AI Director

The standard vehicle cybersecurity model relies on cloud-based detection and response, requiring continuous uploads of vehicle data to a centralized VSOC. While this centralization enables better visibility and analytics, it drives up operational complexity and costs.

VicOne’s xCarbon Edge AI is a GenAI-powered, in-vehicle intrusion detection system that uses onboard processing to analyze telemetry data—such as CAN, Ethernet, and system logs, saving up to 60% lower cloud data transfer and processing costs

With resource-aware model scheduling and comprehensive data from Sonatus AI Director, it accurately detects security risks. By correlating events across multiple ECUs, its advanced threat model with pre-filtered, high-relevance alerts identifies new attack patterns missed by traditional methods, resulting in 80% faster triage.

COMPREDICT Virtual Headlight Leveling Sensor on Sonatus AI Director — Unlocking In-vehicle AI for Smarter Headlight Leveling

To meet the UN R48-09 regulation requiring automatic headlight leveling in European passenger vehicles by 2027, COMPREDICT has developed a 100% software-based Virtual Sensor using AI to replace traditional hardware ride height sensors.

The company’s Virtual Sensor uses the available CAN signals, made available via Sonatus AI Director, to calculate the current pitch angle of the vehicle. After processing by the COMPREDICT model, the pitch angle is made available to other consumers on the vehicle network to control the headlight actuators.

This solution allows OEMs to comply with the latest regulatory standards while enhancing driver safety. This scalable solution supports both certification and real-world scenarios while helping OEMs cut BOM costs by up to $20 per vehicle.