Core conclusion: Chinese domestic brands have dominated China's urban NOA market (domestic brands accounted for 81.1% of urban NOA sales from January to November 2025). Intelligent driving has rapidly expanded from premium models priced above RMB 200,000 to entry-level models at RMB 100,000 or even 60,000–90,000. Two technical routes have taken shape: full-stack in-house R&D solutions and third-party supplier solutions. From the sales perspective, intelligent driving has evolved from an optional selling point into a core factor for vehicle purchase decisions. The free standard configuration model is squeezing the market share of one-time purchase and subscription models.
I. Product Analysis
1. Two Major Technical Routes
Route A: Full-stack in-house R&D by OEMs (Xpeng, Li Auto, BYD, NIO)
· Xpeng XNGP: One of the earliest mass-deployed urban NOA systems in China. Early versions adopted multi-LiDAR setups paired with NVIDIA chips; the new generation shifts to self-developed Turing chips and VLA mapless architecture. Strengths: smooth performance on urban roads, end-to-end intelligent driving across all scenarios. Product positioning: intelligent driving serves as the brand’s core identity, with free high-level intelligent driving standard on mainstream models. Weaknesses: slow expansion of urban coverage in early iterations and relatively high hardware costs.
· Li Auto AD Max / AD Pro: Dual-tiered solution. The Max version features LiDAR plus dual Orin chips, while the Pro version uses a pure vision solution powered by Horizon Robotics chips. Product strategy: pre-installed hardware with lifetime free OTA updates for urban NOA. It prioritizes family-oriented ride comfort, with a conservative and safe intelligent driving style that avoids sharp braking and lane changes, catering to family users.
· BYD God’s Eye: Deployed from high-end Yangwang and Denza models down to the Song and Seagull. Tiered configuration: high-end models come with standard high-level intelligent driving, while ordinary models support optional urban NOA. Features: mass production cost reduction, bringing intelligent driving down to A00-class compact cars. It supports highway NOA, urban NOA and automatic parking, with flexible hardware options including LiDAR and pure vision variants.
· NIO NAD: Fusion perception solution. New vehicles include multi-year free access, with subscription upon expiry. Its intelligent driving integrates highway, urban and battery swap parking functions. Product focus lies on high-end models, where intelligent driving acts as a premium experience add-on.
Route B: Third-party supplier solutions (Huawei ADS, Momenta, Horizon Robotics)
· Huawei ADS: Full-stack hardware and software mapless NOA solution, covering cities and rural roads nationwide. Partner brands: AITO, Avatr, Chery, etc. Version tiers: ADS SE (highway navigation assist), ADS Max (urban NOA). Business models: one-time purchase or subscription. Strengths: fast deployment, low takeover rate, enabling traditional OEMs to quickly bridge gaps in intelligent driving without strong in-house R&D capacity.
· Momenta: Supplies urban NOA solutions to multiple automakers, focusing on low-cost mass production and ranking among top third-party vendors.
· Horizon Robotics: Provides automotive-grade chips and HSD intelligent driving solutions, adopted by Changan, FAW and others, enabling LiDAR-equipped intelligent driving on models around RMB 100,000.
2. Product Segmentation by Price Band
1. Entry-level: RMB 60,000–100,000: Mainly equipped with highway NOA and automatic parking. Starting from 2026, LiDAR and optional urban NOA appear on models such as Seagull and Xingyuan, featuring pre-installed hardware with optional upgrades — a major industry breakthrough bringing intelligent driving into affordable commuter vehicles.
2. Mass-market: RMB 100,000–200,000: The most fiercely competitive segment. A large number of models come with standard urban NOA, adopting pure vision or LiDAR schemes in parallel. This is the main volume driver for high-level intelligent driving among domestic brands.
3. Mid-to-premium: RMB 200,000–400,000: Standard high-level urban NOA, multi-LiDAR and high-computing chips deliver the best intelligent driving experience. This segment forms the fundamental base for technology validation (main models of Xpeng, Li Auto, AITO, Avatr).
4. Luxury: Above RMB 400,000: Yangwang, Zunjie S800 and others are equipped with high-line-count LiDAR and drive-by-wire chassis, exploring L3 applications for highway scenarios.
3. Comparison of Intelligent Driving Business Models
1. Free standard configuration (Li Auto, Xiaomi, mainstream Xpeng models): Hardware factory-fitted, lifetime free OTA updates for software. Currently the most popular market approach, boosting conversion. OEMs embed intelligent driving costs into overall vehicle pricing to lower user decision barriers.
2. Optional / one-time unlock (ordinary BYD models, Huawei ADS Max): Users pay a lump sum at vehicle purchase to activate high-level functions.
3. Subscription model (NIO, Huawei ADS subscription): Monthly or annual billing reduces upfront purchase costs yet leads to higher total long-term expenses. Domestic users show relatively low willingness to pay for subscriptions.
4. Product Pain Points
1. Uneven experience across regions: NOA performs maturely in large cities, while performance drops sharply in tier-3/4 cities and rural areas. Mapless solutions mitigate the gap, yet unstable performance persists in extreme scenarios such as sudden pedestrian cut-ins and mixed non-motorized traffic.
2. Trade-off between hardware cost and reliability: LiDAR improves perception but raises vehicle cost; pure vision cuts costs while weakening stability under adverse weather.
3. Misconceptions among users: Many consumers confuse L2+ NOA with fully autonomous driving, imposing heavy safety communication burdens. L3 commercial deployment remains limited by regulations.
4. High OTA operation costs: Post-launch intelligent driving requires continuous data training and version iteration, demanding sustained R&D investment.
II. Sales Analysis
1. Overall Market Size and Penetration
· Basic L2 driver assistance: Penetration in new domestic passenger cars has neared 70%, becoming standard equipment and no longer a differentiator.
· Urban NOA (high-level intelligent driving): Cumulative sales reached 3.129 million units in China from January to November 2025, with an on-road penetration rate of 15.1% for passenger vehicles. Domestic brands contributed 2.5373 million units, accounting for 81.1% of total urban NOA sales, firmly dominating the high-level intelligent driving market.
· Growth trend: Urban NOA maintained year-on-year doubling growth for two consecutive years, driven by falling hardware costs, free-of-charge strategies and product downscaling from premium to RMB 100,000 models.
2. Brand Sales Structure
1. HarmonyOS Intelligent Mobility (AITO, etc.): Intelligent driving acts as a core sales leverage. Huawei ADS is a key consideration for buyers. Installed capacity has rapidly exceeded 2 million units. The partnership model enables traditional automakers to boost intelligent vehicle sales quickly.
2. Xpeng: Boasts the strongest brand recognition for intelligent driving. XNGP serves as its long-term differentiation tag. Standard intelligent driving on mainstream models drives sales of mid-to-premium vehicles.
3. Li Auto: Leverages free AD Max plus family positioning. Its intelligent driving operates conservatively with high stability, winning acceptance from family buyers and delivering steady conversion.
4. BYD: Supported by its massive overall vehicle sales base, the installed volume of God’s Eye is expanding rapidly. Tiered strategy: standard high-level systems on premium models and optional packages for entry models, covering all price ranges.
5. Others (Aion, Geely, Changan): Adopt hybrid in-house / Horizon Robotics / Momenta solutions. Volume models priced RMB 100,000–200,000 drive the base sales of high-level intelligent driving.
3. Impacts of Intelligent Driving on Sales
1. Marketing value outweighs direct profit: High-level intelligent driving is a powerful marketing asset for short-video promotion and test drives, effectively lowering customer acquisition costs. Urban NOA demonstrations during test drives significantly lift closing rates.
2. Shifts in pricing strategy: The industry is moving from "premium surcharge for intelligent driving" toward free standard configuration. Essentially, OEMs absorb intelligent driving costs into vehicle pricing to trade intelligence for sales volume, rather than monetizing software separately. Subscription models contribute limited revenue due to low domestic user adoption.
3. Customer segmentation:
· Young users in tier-1 and new tier-1 cities: Highly attentive to urban NOA and willing to pay for intelligent driving.
· Lower-tier markets and older family buyers: Prioritize basic L2 functions (ACC, lane keeping), with weak demand for urban NOA.
4. Sales-side Challenges
1. Pressure on test drive and delivery: High-level intelligent driving test drives require sales training and designated safe routes, making standardized in-store test drives difficult.
2. Risk of expectation management: Sales staff may overstate intelligent driving capabilities, triggering user complaints and safety disputes.
3. Intensified homogeneous competition: A large batch of new vehicles will feature standard urban NOA in 2026. Simply "having NOA" loses uniqueness. Competition shifts toward takeover rate, extreme-scenario performance and national coverage.
4. Cost pressure: LiDAR and high-computing chips remain costly. Fitting high-level intelligent driving on low-priced vehicles squeezes gross profit per unit.
III. Competitive Landscape and Outlook
1. Landscape: Dual-track development in the high-level intelligent driving market for domestic brands: top OEMs conduct in-house R&D (Xpeng, Li Auto, BYD); traditional automakers adopt mature third-party solutions (Huawei / Momenta / Horizon Robotics). Joint venture brands lag behind in intelligent driving products and keep losing market share.
2. Short-term trends (1–2 years)
· Urban NOA will keep penetrating downmarket, with RMB 100,000 vehicles becoming the main battlefield for standard deployment. LiDAR and pure vision will coexist long-term.
· Free standard configuration will become mainstream, while subscriptions cannot become a major revenue source in China.
· Competition focus will shift from "whether NOA is equipped" to takeover rate, performance in harsh scenarios, nationwide coverage and end-to-end large-model intelligent driving.
3. Long-term: Once L3 regulations are finalized, highway L3 will see initial commercialization. Intelligent driving will extend from vehicle selling points to Robotaxi and commercial vehicle scenarios. Meanwhile, Chinese intelligent driving solutions will go global alongside vehicle exports.
IV. Brief SWOT Analysis
· Strengths: Domestic brands lead the world in urban NOA deployment speed, supported by complete local supply chains, massive domestic road data and product pricing covering all segments.
· Weaknesses: Heavy R&D investment for intelligent driving with limited profitability; challenges in managing user safety expectations; partial reliance on overseas suppliers for underlying chips.
· Opportunities: Downmarket penetration of intelligent driving, automotive exports, iterative upgrades of end-to-end intelligent driving powered by large models.
· Threats: Homogenized price wars inflate R&D costs; tightened regulation triggered by safety incidents; risks to user trust.