16th April 2026
Hands-free driving wins: Why level 2+ is outpacing level 3 autonomous driving
By Michele Richichi and Rohan Hazarika
Article Summary
Unpack why hands-free Level 2+ is scaling faster and why cost, liability and regulation are slowing level 3 autonomous driving and reshaping OEM strategies.
While automakers from China to North America have publicly confirmed their intentions to pursue Level 3 autonomous driving, shows others—most notably the early-adopter German premium brands—are subtly easing off their Level 3 ambitions as their latest Level 2+ autonomous driving systems come to market.
BMW, MercedesBenz and Stellantis all went to market with highly publicized Level 3 plans, and both German automakers have Level 3 autonomous driving systems on the market today in very limited availability. But each has discovered that the combination of high cost, low consumer utility, weak demand and slow regulatory progress makes the current generation of "eyes-off" driving systems hard to justify.
In parallel, these same companies are seeing the fruits of their investments as advanced Level 2+ autonomous driving offerings (sometimes described as L2++) come to market for the first time. These systems deliver different, but no less substantial, value propositions that resonate with customers without crossing the liability threshold into highly automated L3 driving.
How regulations are shaping Level 2+ and Level 3 autonomous driving
The contrasting paths of Level 3 and Level 2+ automation stem largely from regulatory structures. Liability is the reason—the manufacturer is liable in Level 3 and above—though regulatory approaches diverge.
Under the UNECE framework, UN R157 (ALKS or Automated Lane Keeping System) Level 3 autonomous driving standards were recently expanded to 130 kph with automated lane changes, but each country must integrate them into national law before real-world use. Only a few, such as Croatia, the Czech Republic and Germany have done so, limiting Level 3 deployment mainly to German roads. Systems like Mercedes‑Benz Drive Pilot and BMW Personal Pilot L3 remain confined by narrow operational domains, speed caps and environmental conditions—since manufacturers assume liability when Level 3 is active.
By contrast, UN R171 (DCAS or Driver Control Assistance Systems) governs advanced Level 2 automation systems that retain driver responsibility. Once approved, OEMs can roll out Level 2+ autonomous driving features broadly, since regulators face fewer legal hurdles. DCAS allows hands‑free highway driving and system‑initiated lane changes while ensuring constant driver monitoring.
In mainland China, hands‑free driving is prohibited for both L2+ and L3 autonomous driving, whereas in the US, it was never formally prohibited, enabling automakers like Tesla to deploy hands‑free driving functions while maintaining driver liability.
OEMs are pausing their Level 3 autonomous driving ambitions
Level 3 autonomous driving has a weak value proposition today because it is still constrained by narrow Operational Design Domain (ODD), low speeds and patchy geographic availability. The Level 3 setup requires lidar, radar, cameras, high-performance computing (HPCs) or domain controllers—and even weather data; yet these only worked in good weather up to 60 kph on approved German motorways until recently, when speeds were raised. Priced at €6,000–€7,000, this setup faced limited utility and regular deactivation, which failed to meet customer expectations.
BMW R&D chief Joachim Post said in March 2026 that demand for Level 3 autonomous driving is not at a stage to make it profitable, but BMW would pursue it if viable business models emerge. The automaker is discontinuing its eyes-off driving Personal Pilot system in the BMW 7 Series facelift this year.
Mercedes‑Benz has made a similar decision. Its Drive Pilot remained constrained by conditions and geography even after its 95 kph expansion in Germany. The brand is expected to drop Drive Pilot from the S‑Class and EQS facelifts this year.
Stellantis also shelved its AutoDrive program in 2025, citing high cost, validation complexity and limited consumer interest.
Collectively, these withdrawals reveal a clear reality that early Level 3 autonomous driving systems work technically but are commercially unsustainable. Their limited ODDs, costly sensor stacks, software strategies and a fragmented legal environment have capped them as niche, low‑volume technology demonstrators rather than scalable products.
Level 2+ adoption surges
The shift from Level 3 to Level 2+ autonomous driving systems is mainly driven by liability shifts, high development costs and regulatory hurdles. Level 3 systems require expensive sensors, high-performance computing, extensive validation and continuous monitoring, whereas Level 2+ solutions can offer many of the same customer-facing features with less hardware and at a lower cost.
This broader industry trend is evident in major automakers’ strategies, as reflected in recent autonomous driving news. Tesla moved early to end-to-end driving with FSD v12, while Chinese OEMs like Geely have secured UN R171 approval for Driver Control Assistance System (DCAS). Brands such as Ford, GM and Rivian are also expanding hands-free Level 2 automation offerings across more models and markets.
For example, BMW’s Symbiotic Drive and Motorway Assistant enables hands-off motorway driving up to 130 kph, automated lane changes confirmed by eye gaze and urban functions like traffic-light recognition, all at a fraction of the cost of BMW’s former Personal Pilot L3. Similarly, Mercedes-Benz is taking this approach with MB. Drive Assist Pro on the new CLA, developed with NVIDIA to blend highway and urban assistance through a co-driving model under driver supervision.
From 2026–2030, S&P Global Mobility forecasts a much faster rise in Level 2+ adoption than Level 3, with Level 2+ becoming the mainstream automation path while Level 3 remains a smaller, slower-growing niche. The key reason for this is that Level 2+ is easier to scale across markets and models, while Level 3 still faces narrower operating limits, heavier legal exposure and more complex validation.
That said, the long-term outlook for Level 3 autonomous driving is improving. Better AI-native development tools, stronger compute, cheaper sensors and the spread of end-to-end architectures should make future Level 3 programs more viable than the first wave. Even so, fragmentation in regulation and the shift in liability away from the driver mean Level 3 autonomous driving is still likely to advance more slowly than Level 2+.
Unlock deeper insights into autonomous vehicle technology
From technology maturity to market adoption, S&P Global Mobility’s autonomy forecasts offer a comprehensive outlook on how automated driving is expected to evolve globally, supporting smarter decisions across product planning, investment and mobility strategy.
For more about our autonomy forecasts, get in touch.
This article was published by S&P Global Mobility and not by S&P Global Ratings, which is a separately managed division of S&P Global.