What are some trends shaping the ADAS/AD space in 2026?
- Luigi Bisbiglia
- 1 hour ago
- 4 min read
If you want to understand where the ADAS industry is really heading in 2026, this insight explores why the biggest opportunity may not be SAE Level 3 at all, but the rapid evolution of L2+ / L2++ assisted driving, urban autonomy and AI-native vehicle architectures.
The 2026 ADAS landscape is evolving around three key trends: uncertainty over the market readiness of SAE Level 3, growing value creation from SAE Level 2 assisted driving and the increasing adoption of AI-centric perception over traditional rule-based approaches.
Uncertainty over SAE Level 3 market readiness
The industry is converging on highly capable supervised driving systems rather than rushing broad L3 deployment. While limited L3 systems now work reasonably well on mapped highways under controlled conditions, scaling them globally remains extremely expensive and operationally difficult. This is why many automakers are realizing that achieving L3 autonomy is not just a technical challenge and it is more about consumer acceptance and ecosystem readiness.
Some major brands like Mercedes-Benz and BMW have opted to discontinue SAE L3 rollout in their upcoming editions, while some have postponed plans for their first rollout (e.g., Stellantis).
More value proposition unlocking at SAE Level 2 assisted driving
The industry’s focus is shifting toward delivering profitable, software-driven, assisted-driving experiences at scale.
As of 2026, the industry is finding more value in scaling SAE L2/L2+ and L2++ functionalities than aggressive SAE L3 rollouts. Advanced assisted driving systems already deliver strong customer benefits such as hands-free highway driving and automated lane changes, all without the massive cost, liability, and regulatory burden associated with L3. Most importantly, L2+ deployments can be continuously improved through OTA updates.
Urban autonomy is emerging as the new battleground and the key competitive challenge over the next 3–5 years will be less about highway autonomy and more about scalable urban driving that reliably takes customers from origin to destination with minimal intervention (albeit on a hands-on mode)
China is already commercializing this vision through Urban Navigate-on-Autopilot and Address-to-Address navigation, while the US and Europe are taking a more measured path, constrained by regulatory and liability considerations.
In China, Urban NoA is becoming a mass-market feature, expanding to entry-level and mid-level models. Map-free AI architectures are becoming standard. NIO offers one of the most robust sensor fusion (trifocal camera and long-range lidar) and the media reports suggest that NIO could become the first Chinese brand to offer SAE Level 3 ‘hands-off, eyes-off’ piloted driving outside China.
The USA is the strong leader in highway autonomy but city-wide address-to-address capability remains limited. Most brands remain focused on mapped highway operations, except Tesla.
Finally, Europe has acted so far in a very conservative way, putting a strong emphasis on highway automation. However, Mercedes-Benz, that was the first automaker to launch SAE L3 on a commercial model in both US and Europe, has announced that the service will be discontinued in favor of MB Drive Assis Pro (urban autonomy).
Tesla has received the primary regulatory approval for its SAE L2++ urban autonomy in the Netherlands and is eyeing permission to deploy the feature in more European member states.
BMW, that is the only automaker in Europe that has got commercial approval for combining SAE Level 3 (Personal Pilot) and ‘hands-off’ assisted driving (Highway Assistant) on the same model, has also announced to discontinue SAE L3.
AI-centric perception is gaining more popularity over traditional rule-based stacks
Another emerging, yet defining trend, is the migration from hard-coded, rule-based ADAS toward AI-native driving stacks.
Instead of explicitly programming every driving rule, OEMs now train models on massive real-world driving datasets, so the system learns to understand the environment statistically.
This trend is especially strong with premium brands and among newer EV-focused automakers (NIO, Xpeng), where ADAS development is becoming more like consumer AI/software development than traditional automotive engineering.
An important consequence of this trend is that the ADAS value chain is shifting from sensors to compute:
More and more OEMs are reorganizing vehicle architectures around centralized compute because future SDVs need compute platform capable of supporting ADAS, automated driving, digital twins, AI assistants and more.
The challenge is no longer collecting data but rather processing it in real-time because of the sensor proliferation, that is causing more and more data bottleneck.
OEMs increasingly view ADAS as a software business rather than a hardware feature. A powerful ADAS hardware supports OTA feature addition and capability upgrades on a post-sale basis. Many automakers can already unlock an entirely new feature remotely (e.g. Ford unlocking BlueCruise via OTA)
Future modern systems will increasingly rely on end-to-end neural networks and AI-based perception models that will serve as a foundation for assisted/automated driving.
"While SAE Level 3 is still facing uncertainty around when the market will be ready for broader adoption, SAE Level 2 is already emerging as a more immediate value pool, as automakers continue to enhance assisted-driving capabilities and monetization opportunities. On the technology side, AI-based perception is gaining momentum, as automakers increasingly move beyond traditional rule-based approaches." Luigi Bisbiglia - Senior Business Development Manager |  |
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 |
