SuperVision

Mobileye SuperVision is an advanced driver-assist system designed for automakers integrating hands-off navigation capabilities into premium consumer vehicles.

Reviewed by 7wData

On this page

Publisher review

Mobileye SuperVision is an advanced driver-assist system designed for automakers integrating hands-off navigation capabilities into premium consumer vehicles. It targets manufacturers seeking scalable autonomy solutions, particularly in China and Europe, with deployments already in Geely Group's Zeekr 001 and 009 EVs. The system combines 11 high-definition cameras with Mobileye's Road Experience Management (REM) mapping and Responsibility-Sensitive-Safety (RSS) model to enable autonomous lane changes, highway assist, and collision avoidance at speeds up to 80 mph. Over 125 million vehicles currently use Mobileye technology, with 300,000 SuperVision-equipped units already on roads.

The system operates via two EyeQ5/6 system-on-chips processing 360-degree camera feeds, leveraging crowdsourced REM data for centimeter-level localization. Key features include point-to-point automated navigation, evasive maneuver assist, and over-the-air updates - demonstrated by a 2023 rollout to 70,000 Zeekr owners in China. Mobileye projects $3.5 billion in SuperVision revenue through 2030, with 1.2 million units expected by 2026. The technology serves as a bridge to full autonomy, incorporating building blocks like parking visualization and traffic jam assist while maintaining driver monitoring requirements.

Against competitors like Waymo and Tesla, SuperVision differentiates through camera-only perception optimized for cost and scalability rather than lidar-dependent systems. Mobileye holds design wins with premium European automakers for 2025 deliveries and dominates China's ADAS market through Geely partnerships. However, the camera-first approach faces limitations in heavy rain or snow compared to multi-sensor systems. While Mobileye benefits from Intel's backing, this relationship creates dependency - the company must balance innovation with corporate priorities of its controlling stakeholder.

The system's $17 billion ADAS pipeline through 2030 reflects strong demand, but relies heavily on Chinese adoption and assumes continued regulatory approval for eyes-on autonomy. Technical trade-offs include weather-dependent perception and the need for driver attention despite hands-off operation. Mobileye addresses these through REM's crowdsourced map updates and RSS safety modeling, but faces challenges scaling to fully autonomous operation without additional sensors.

Get the AI & data signal, daily.

48k+ subscribers read this every morning. One email, both newsletters. Unsubscribe anytime.

How it works

  1. 360-degree camera perception

    Uses 11 cameras (8MP main, 3MP service) for full surround vision including 120° front and 195° parking views

  2. REM-based AV mapping

    Leverages crowdsourced Road Experience Management data for centimeter-level localization on mapped roads

  3. Responsibility-Sensitive-Safety

    Implements mathematical safety model (RSS) for collision avoidance and evasive maneuver decisions

  4. Hands-off highway assist

    Enables autonomous lane changes and traffic jam navigation up to 80 mph on approved road types

  5. Over-the-air updates

    Field-upgradable system demonstrated by 2023 feature rollout to 70,000 Zeekr vehicles in China

  6. Dual EyeQ processing

    Runs on two EyeQ5/6 system-on-chips for parallel processing of camera feeds and decision making

  7. Parking visualization

    Provides 195° service camera coverage for automated parking applications and obstacle detection

Strengths and trade-offs

Strengths

  • Projects $3.5 billion revenue from SuperVision through 2030 with 1.2 million units expected by 2026.
  • Processes 360-degree environment via 11 cameras without expensive lidar, keeping costs competitive.
  • Over 125 million vehicles already use Mobileye technology, providing deployment scale advantages.
  • Crowdsourced REM maps cover nearly every road in US and EU, enabling precise localization.

Trade-offs

  • Camera-only perception struggles in heavy precipitation compared to lidar-equipped competitors.
  • Requires driver attention despite hands-off operation, creating potential mode confusion.
  • Dependent on Intel's strategic priorities as controlling stakeholder for R&D funding.
  • Limited to 80 mph operation with reduced functionality in unmapped areas.

Pricing context

Pricing not publicly disclosed; sold to automakers for integration into production vehicles

Getting started with SuperVision

  1. Integrate hardware

    Install 11 high-definition cameras (8MP main and 3MP service) around the vehicle for 360-degree perception coverage, including front and parking views.

  2. Connect EyeQ processors

    Link the camera feeds to dual EyeQ5/6 system-on-chips for parallel processing of visual data and autonomous decision making.

  3. Enable REM mapping

    Activate Road Experience Management connectivity to access crowdsourced map data for centimeter-level localization on mapped roads.

  4. Configure RSS model

    Set up Responsibility-Sensitive-Safety parameters for collision avoidance thresholds and evasive maneuver logic based on vehicle specifications.

  5. Test highway assist

    Verify hands-off navigation functionality including autonomous lane changes and traffic jam assist up to 80 mph on approved road types.

Frequently Asked Questions

What is Mobileye SuperVision?

Mobileye SuperVision is an advanced driver-assist system for premium vehicles, enabling hands-off navigation. It uses 11 cameras with REM mapping and RSS safety models for autonomous lane changes, highway assist, and collision avoidance up to 80 mph. Currently deployed in Zeekr EVs with 300,000 units on roads.

How does SuperVision compare to Tesla Autopilot?

SuperVision uses only cameras for perception, unlike Tesla's radar-assisted system. It leverages Mobileye's REM mapping for precise localization but has weather limitations. While both enable hands-off highway driving, SuperVision focuses on automaker integration rather than direct consumer sales like Tesla's vertically integrated approach.

Which cars currently use Mobileye SuperVision?

Geely Group's Zeekr 001 and 009 electric vehicles are the primary adopters, with 70,000 units receiving over-the-air updates in 2023. Mobileye has design wins with premium European automakers for 2025 deliveries, expanding beyond its strong presence in China's ADAS market through Geely partnerships.

What are the limitations of SuperVision's camera-only system?

The camera-based perception struggles in heavy rain or snow compared to lidar-equipped systems. It's limited to 80 mph operation and requires driver attention despite hands-off functionality. Performance decreases in unmapped areas, though crowdsourced REM updates help maintain localization accuracy on known routes.

How does SuperVision's safety model work?

SuperVision uses Mobileye's Responsibility-Sensitive-Safety (RSS) model, a mathematical framework for collision avoidance decisions. Combined with 360-degree camera coverage and REM mapping, it enables evasive maneuvers while maintaining safe distances. The system processes data through dual EyeQ chips for real-time decision making.

What's the future outlook for Mobileye SuperVision?

Mobileye projects $3.5 billion in SuperVision revenue through 2030, with 1.2 million units expected by 2026. The technology serves as a bridge to full autonomy, with features expanding through over-the-air updates. Growth depends on Chinese adoption and regulatory approval for eyes-on autonomous systems.

Alternatives

How SuperVision compares

Direct head-to-head against 2 competitors. Picked by 7wData.

This tool

SuperVision

Pricing
Pricing not publicly disclosed; sold to automakers for integration into production vehicles
Target
Mobileye SuperVision is an advanced driver-assist system designed for automakers integrating hands-off navigation capabilities into premium consumer vehicles.
Strength
Projects $3.5 billion revenue from SuperVision through 2030 with 1.2 million units expected by 2026.
Watch for
Camera-only perception struggles in heavy precipitation compared to lidar-equipped competitors.

Wayve

Pricing
Custom/Contact sales
Target
Automakers seeking AI-driven autonomous driving
Deployment
Software updates
Strength
Mapless, hardware-agnostic L2+ to L4 autonomy
Watch for
Early-stage tech, limited real-world deployments

Mercedes-Benz Drive Pilot

Pricing
Bundled with premium vehicles
Target
Luxury automakers with Level 3 focus
Deployment
Factory-installed
Strength
First approved Level 3 system in Germany
Watch for
Geofenced to highways under 60km/h

User reviews

No user reviews yet. Be the first to write one.

Sources

Reporting on this tool draws on these publicly available sources.

  1. www.mobileye.com
  2. www.mobileye.com
  3. www.reddit.com
  4. www.facebook.com