End-to-End AI in ADAS: Where Cost, Scale and Value Converge

For more than a decade, competition in Advanced Driver Assistance Systems (ADAS) has been defined by hardware performance, sensor configurations and increasingly sophisticated driver assistance features. Vehicle manufacturers differentiated through compute platforms, perception capability and functions such as highway navigation and automated parking.

That competitive model is now changing.

End-to-End (E2E) AI is more than a new software architecture. It is changing how automated driving systems are developed, improved and commercialised. Instead of relying primarily on hand-engineered rules and separately optimised software modules, E2E approaches use large-scale driving data to learn increasingly complex driving behaviour. This creates a fundamentally different development model, where real-world experience can feed back into the system and accelerate improvement over time.

What Makes End-to-End AI Different?

Unlike conventional ADAS, which breaks driving down into separate perception, prediction, planning and control functions, E2E AI learns driving behaviour more directly from large-scale driving data, bringing more of these capabilities together within a single model. As E2E systems evolve with world models and Vision-Language-Action (VLA) architectures, they are becoming better at understanding context, reasoning through complex situations and generalising to unfamiliar driving scenarios.

This data-driven approach also changes how driving systems can be developed and improved at scale. Performance can improve through continuous fleet data collection, model retraining and over-the-air (OTA) updates, reducing reliance on increasingly complex rule-based software. Today, its greatest commercial impact is in supervised urban assisted driving (L2++), enabling address-to-address driving while laying the foundation for future Level 3 and selective Level 4 automation.

The Learning Loop Is Becoming the Competitive Advantage

As AI foundation models become more accessible, the model itself is unlikely to provide lasting differentiation. Instead, competitive advantage will increasingly depend on the effectiveness of the learning loop, from collecting fleet data and retraining models to validating improvements and deploying updates safely at scale.

This changes the economics of ADAS. Cloud infrastructure, simulation, AI training, validation and OTA deployment are becoming as strategically important as sensors and compute hardware.

Ecosystems Are Replacing Component Competition

The shift towards E2E AI is also changing the role of technology suppliers.

Companies such as Horizon Robotics, NVIDIA and Wayve are no longer competing solely on compute hardware or software. They are building broader AI ecosystems that combine development frameworks, safety architectures and deployment support. For OEMs, selecting an E2E partner is becoming a strategic decision rather than a conventional sourcing choice, shaping how driving systems are developed, deployed and continuously improved at scale.

Horizon Robotics, NVIDIA and Wayve illustrate different approaches to the emerging E2E ecosystem. Horizon Robotics combines hardware and software in an integrated platform with a strong focus on the Chinese market. NVIDIA is building a full-stack proposition around DRIVE compute, software and its broader AI and safety ecosystem, while Wayve is taking a more hardware-agnostic approach with its AI Driver platform.

As OEMs move beyond traditional Tier-1 partnerships, technology partner selection increasingly shapes not just ADAS capabilities, but also data pipelines, development environments, validation, compute architecture and the ability to scale software improvements across the fleet.

Partnerships Are Accelerating in the End-to-End AI Ecosystem

China Has Become the Global Proving Ground

China has emerged as the world's leading commercial test bed for End-to-End AI.

Multiple OEMs now offer L2++ address-to-address assisted driving across a growing range of production vehicles, demonstrating that E2E AI is moving beyond research into commercial deployment.

Perhaps the most important lesson is that there is no single winning strategy. Some manufacturers, including XPENG, Li Auto and NIO, have invested heavily in proprietary AI capabilities, while others combine in-house development with ecosystems from Huawei, NVIDIA, Horizon Robotics, Qualcomm and DeepRoute AI. 

SBD's analysis of the Chinese market highlights the scale of this transition. In the assessed vehicle set, 14 OEMs now offer L2++ address-to-address assisted driving across more than 80 vehicle models, with E2E AI becoming the dominant architecture behind these systems.

China also demonstrates that the industry has largely moved beyond the "vision-only versus LiDAR" debate. Most production systems combine cameras, navigation and at least one LiDAR sensor to improve robustness in complex urban environments. While Tesla and some Chinese OEMs have demonstrated the vision only approach is highly capable.

The key differentiator is the speed at which OEMs can industrialise AI, scale it across platform and continuously improve through fleet data and software updates. The competitive battleground is shifting from choosing the “right” architecture to iterating faster and scaling innovations across the fleet.

Urban Assisted Driving Has Become China's Proving Ground for End-to-End AI

The Basis of Competition Is Expanding

The shift to E2E AI is changing not only how ADAS is developed, but also where value and strategic control sit across the automotive ecosystem. As driving intelligence becomes increasingly software-defined and data-driven, traditional roles are being reshaped and established value propositions are coming under pressure.

OEMs are increasingly moving to the centre of the ecosystem. Their control of the vehicle, customer relationship, fleet data and software deployment gives them greater ownership of the E2E learning loop and greater responsibility for AI development, validation, safety and continuous improvement.

For technology suppliers, the implications are equally significant. Tier-1 suppliers, compute providers, LiDAR companies and mapping providers can no longer rely solely on the differentiated performance of individual components. Their relevance will increasingly depend on how effectively they contribute to the broader E2E ecosystem through integration, data, AI development, simulation, validation, safety or other capabilities that help OEMs scale and continuously improve their systems.

This raises a fundamental strategic question for every ecosystem player: what role will it play in the E2E value chain, and how must its value proposition evolve as OEMs take greater ownership of driving intelligence?

Ecosystem player	Core challenge	Strategic opportunity
OEMs	Balance legacy ADAS investment with E2E scale, fleet-data access, AI-native SDV transition and safety accountability.	Scale from highway assist to L2++ address-to-address driving, using data and customer experience to establish a pathway to L3.
Tier-1 suppliers	Preserve strategic relevance as OEMs take greater ownership of software, system architecture and driving intelligence.	Shift towards high-value roles in E2E integration, validation, safety assurance, simulation and data-driven performance optimisation.
Compute and AI-platform providers	Differentiate beyond raw compute capability amid varied OEM requirements and fragmented regional ecosystems.	Combine silicon with software tools, safety frameworks, deployment support and ecosystem interoperability.
LiDAR companies	Justify LiDAR cost and relevance in vision-first E2E architectures.	Position LiDAR around measurable safety, redundancy and 3D-perception benefits, while expanding into data and simulation products.
Mapping companies	Maintain relevance as E2E stacks become less map-dependent	Build value through dynamic context, spatial intelligence, location APIs and machine-readable data for driving, HMI and connected services.

Table 1. Strategic implications of End-to-End AI across the automotive value chain

The Future of ADAS Will Be Defined by the Driving Experience, Enabled by Better Learning Systems

End-to-End AI is often described as the next breakthrough in perception and decision-making. But the bigger shift is in how ADAS is developed and improved, with E2E AI enabling systems to learn from real-world driving and continuously improve at scale.

The competitive advantage will increasingly come from how well an OEM can turn real-world driving situations into a safer, more natural and human-like driving experience. ADAS is becoming less about individual features and more about continuously improving that experience through learning, safe validation and fleet-wide deployment.

"The bigger question is how the automotive ecosystem will evolve around this shift. As OEMs take greater ownership of driving intelligence, partners will need to rethink their roles and value propositions, finding new ways to support OEMs in scaling E2E AI and delivering better driving experiences at the right cost, at scale and with sustainable value." Varun Krishna Murthy - Consulting Manager at SBD Automotive

How SBD can help

SBD Automotive can help benchmark your position against the wider industry and identify where action is needed most. To explore how these trends impact your strategy, architecture and supplier roadmap, get in touch with SBD Automotive for a deeper discussion. Email info@sbdautomotive.com