.webp)
.webp)
“For our customers, winning in the next era of automotive requires delivering a self‑driving system built to beat the most advanced ADAS platforms in production today, both in performance and at a cost consumers can afford.”
That’s how Qasar Younis, Co‑founder and CEO of Applied Intuition, frames the stakes for automakers. The push for autonomy leadership is accelerating, driven by recent AI breakthroughs and rapidly shifting consumer expectations. Tesla’s full self‑driving capabilities and advances from leading Chinese OEMs have set a high bar for end‑to‑end (E2E) driver‑assistance performance. This presents a challenging technology race for other automakers.
To help our customers in that race, Applied Intuition today announced its Self-Driving System (SDS) for Automotive. SDS for Automotive is an end‑to‑end learned neural network ADAS platform that seamlessly integrates perception, planning, and control into a unified system—rather than discrete rule‑based modules—enabling OEMs to deliver safe, reliable human-like driving performance. It will offer a complete L2++ feature set, with a pathway to L3 and L4, supported by the Applied development and validation toolchain the world’s top OEMs already trust as they go to production.
SDS was designed by Applied Intuition's world‑class autonomy engineering and research teams, and it draws on experience from nearly a decade of autonomy work in trucking, mining, and defense. SDS integrates with Applied Intuition’s Vehicle OS and is designed from the ground up for production readiness with use of mass production sensors and embedded compute.
Key Differentiators
Advanced architecture and observability
“Our E2E learning approach combines awareness of the driving environment, understanding of driving behavior, and observability in one system,” said Peter Ludwig, Co‑founder and CTO.
SDS for Automotive provides defined intermediate outputs—such as detected objects and scene understanding—that can be inspected, validated, and traced. This enables OEMs to meet regulatory requirements and global safety standards including ISO 26262, SOTIF, and ISO/PAS‑8800:2024, while still benefiting from a state-of-the-art neural architecture.
Complete feature set, scalable over time
From remote intelligent parking and automatic emergency braking (AEB), to advanced urban driving and highway pilot, SDS is designed to deliver a full range of L2++ driver assistance features, with a pathway to L3 and L4.
Flexible, cost‑efficient hardware strategy
SDS is silicon‑ and compute hardware‑agnostic, enabling OEMs to choose the compute and sensor package that best fits their needs. Its camera-heavy design supports flexible sensor setups—from a single forward camera to 11 surround cameras—with optional radars. Lidar can be added for L3 and above for additional safety. SDS does not require HD maps, so cars can drive anywhere there are roads, keeping bill of materials (BOM) costs low and enabling global scaleup.
Transparency and OEM control
A white box autonomy stack gives OEMs full source code access and the ability to customize, validate, and integrate deeply. This lets manufacturers build brand‑tailored autonomy experiences across their lineups.
Why It Matters Now
“E2E research breakthroughs over the past 2 years have enabled us to completely rethink ADAS engineering,” said Peter Ludwig, Co‑founder and CTO. “Our E2E model, trained on large‑scale human and synthetic data, learns the entire driving process in a way that scales and adapts, delivering a driving experience as smooth and reliable as the best professional driver.”
Built with Applied's AI in ADAS research on large‑scale transformer architectures similar to those powering modern generative AI, SDS for Automotive has achieved step‑function improvements in capability every few months and continues improving rapidly. SDS is on track to push the state-of-the-art on ADAS safety and capability.
The Problem with Existing ADAS Solutions
Traditional ADAS vendors provide closed, black‑box systems that limit transparency and customization. Chipmakers suffer from inherent hardware lock‑in that limits flexibility. Some new entrants have focused on costly, L4‑heavy designs unsuited to mass production, or rely on suppliers with speculative business models unlikely to survive. OEMs need a unified open platform, engineered for automotive integration from day one, giving them full control to keep up with market competition.
Proven Expertise and Track Record
Applied Intuition’s nearly a decade of experience building autonomy in high‑demand domains flows directly into SDS for Automotive. Backed by a profitable business model and sustained investment, Applied Intuition is here for the long haul.
Much of the technology behind SDS is proven in trucking on public roads in Japan today — integrating with existing programs, enabling rapid validation, regulatory compliance, and a direct path to SOP for passenger cars.
“Our success applying this approach in trucking and off‑road programs has proven it can adapt quickly and deliver safer, more capable features,” Peter said. “That foundation positions us to bring SDS to passenger vehicles.”
Under the Hood
SDS for Automotive’s end‑to‑end neural network design combines proven production techniques for E2E autonomy with leading‑edge research for next generation advancements. Models trained on huge volumes of data form the core of the approach, with new pre‑ and post‑training techniques constantly integrated. The system’s architecture preserves traceability while still running on production-grade embedded compute and maintaining complex driving capabilities.
The end‑to‑end pipeline is production‑proven, runs on mass‑market hardware, and enables rapid loop closure for new features. The petabyte-scale data engine collects, processes, and curates multi-modal vehicle and sensor data, running on robust infrastructure that supports labeling, mining, experiment tracking, and manages compute to accelerate model iteration and deployment.
Ready to get started with SDS for Automotive?
Learn how Applied Intuition’s Self‑Driving System (SDS) for Automotive can help your team deploy benchmark‑level ADAS capabilities that match and surpass today’s leaders—and bring them to market faster.
