ADAS: A Key to Unlocking Revenue for the Software-defined Vehicle
By Matt Fisch, AEye CEO –
Automotive OEMs are moving toward the electric and software-defined vehicle, where – like your smartphone – service updates and upgrades, as well as new revenues, will be delivered over the air. This movement toward electrification, automation, and service-based business models dramatically increases the industry’s reliance on software architectures to deliver on the promise of enhanced safety features and functionality, as well as new revenue streams. It also underlines the importance of Advanced Driver Assist Systems (ADAS) as a key revenue-driver for the software-defined vehicle.
The opportunity is undeniable. Look no further than Tesla’s $99-$199/month “Full Self-Driving” or Ford’s $75/month BlueCruise feature. Goldman Sachs forecasts electric and automated vehicles will capture $149 billion in new business by 2030, with new technologies enabling autonomous safety and convenience enhancements that could generate income of $3,750 per vehicle, a level of improvement that would boost OEMs’ average operating margins from 7% to 12%.
The ADAS Sensor Suite
These ADAS systems will require a full suite of sensors, and even more so as OEMs level up safety features to add increasingly advanced functionality, like high-speed highway automated driving. Having digital eyes on the road requires – not just cameras, radar and ultrasonics – but the addition of lidar to provide detailed, timely information about a car’s surroundings, which is essential to keeping drivers and pedestrians safe.
This is evidenced in the National Highway Traffic Safety Administration (NHTSA)’s new proposed safety standard, which would require Automatic Emergency Braking (AEB) and Pedestrian AEB systems on cars and trucks. The proposal recommends testing these systems at higher speeds, low-light, and other adverse lighting conditions, where lidar excels, to ensure high-performance standards. Lidar provides 4D vision that contributes to learning and is updatable – a key future component of ADAS systems.
OEMs: Built for Flexibility
The transition by OEMs to service-based business models for ADAS feature sets comes on the back of centralized compute, large language models, and over the air updates. And it’s designed for smart sensors that can achieve high performance metrics in a wide variety of real-world environments and use cases. Software-defined vehicles – those vehicles whose features and functions are primarily enabled through software – require software-defined sensing solutions that are flexible enough to learn over time, with the ability to update and adapt depending on the needs of the system and software.
That’s Where AEye’s Lidar Comes In
Our sensor is adaptive – meaning it has a high degree of programmability and the ability to deal with the ADAS learning curve via software. For example, our system can dynamically track the horizon, accommodating for things like slopes and potholes on the road. The sensor can also be reconfigured on the fly, based on speed or GPS cues pre-set by OEMs, to ensure safe performance regardless of where the vehicle is driving (urban, parking, or highway) and how fast it’s going. In addition, AEye’s adaptive lidar enables OEMs to easily “turn on” updates and new revenue-generating features over time, once developed and fully tested, via over-the-air software updates, without purchasing new hardware. On the business side, our underlying architecture allows AEye to partner throughout the ecosystem, with OEMs, software stack providers, AI chip companies, and traditional Tier 1s.
It’s clear to me that ADAS will lead the rollout of SaaS solutions for the software-driven car, lidar will enable many new features, and additional functionality will follow, driven by the ADAS learning curve. This is a massive market opportunity, with automakers like GM and Stelantis anticipating up to $25 billion in SaaS sales revenue annually by as early as 2030.
AEye is perfectly positioned for this evolution, with its architecture built from the ground up to be intelligent – adapting to meet the needs of the systems it serves. With cars anticipated to have 650 million lines of code – 15 to 30 times that of a fighter jet – just two years from now, it’s never been more important for sensors to be just as smart and agile as the software-driven cars they will occupy.