Baidu Apollo: China's Largest Autonomous Driving Platform

Baidu Apollo is China's most advanced and widely deployed autonomous driving platform, encompassing the Apollo Go robotaxi service, L4 autonomous driving software stack, and partnerships with major automakers including Geely, GAC, and Ford China. Apollo Go operates in over 10 cities across China including Wuhan, Beijing, Chongqing, and Shenzhen, having completed over 8 million robotaxi orders by 2025. Baidu's Apollo RT6, the sixth-generation robotaxi designed for mass production at under $30,000 per unit, represents a breakthrough in making autonomous ride-hailing economically viable. The platform's open-source software stack has been adopted by over 200 global partners.

TL;DR

Baidu Apollo operates the world's largest robotaxi fleet with 8M+ orders in 10+ Chinese cities. The Apollo RT6 robotaxi costs under $30K, making autonomous ride-hailing commercially viable at scale.

Key Insights

Apollo Go Robotaxi

8M+ orders completed

Apollo Go (萝卜快跑) is China's largest commercial robotaxi service, operating in over 10 cities including Wuhan, Beijing, Chongqing, Shenzhen, and Hangzhou. By 2025, the service has completed over 8 million ride-hailing orders, making it the world's most-used autonomous taxi service. In Wuhan alone, Apollo Go operates a fleet of over 500 vehicles and handles tens of thousands of daily orders. The service has expanded from limited zones to city-wide commercial operation in several cities.

Apollo RT6

Under $30,000 per unit

The Apollo RT6 is Baidu's sixth-generation robotaxi designed for mass production at a cost below $30,000 per unit, dramatically lower than previous generations. The vehicle features a retractable steering wheel, integrated LiDAR and camera sensors, and Level 4 autonomous driving capability. The reduced cost is achieved through sensor miniaturization,供应链优化, and scalable manufacturing partnerships with JAC Motors. The RT6 is designed to be deployed in ride-hailing fleets, potentially replacing human-driven taxis in major cities.

L4 Autonomous Driving

Over 100M test km

Baidu Apollo has accumulated over 100 million kilometers of autonomous driving test mileage across its various test vehicles and robotaxi fleet. The platform's L4 autonomous driving software stack includes perception (multi-sensor fusion of LiDAR, cameras, radar), prediction, planning, and control modules. Apollo's autonomous driving system has demonstrated the ability to handle complex urban scenarios including intersections, construction zones, pedestrian-heavy areas, and adverse weather conditions.

Automaker Partnerships

200+ global partners

Baidu Apollo has established partnerships with over 200 global companies including major automakers (Geely, GAC Group, Ford China, Hyundai, Toyota), technology companies, mapping providers, and Tier 1 suppliers. The Apollo ecosystem enables automakers to integrate Apollo's autonomous driving capabilities into their vehicles. Baidu's partnership with Geely on the JiYue (极越) brand has produced commercially available vehicles with advanced ADAS features based on Apollo technology.

Apollo Software Stack

Open-source platform

Baidu's Apollo autonomous driving platform includes an open-source software stack that provides perception, planning, and control modules for autonomous driving development. The platform supports both L2/L3 advanced driver assistance systems and L4 full autonomous driving. Apollo's software stack has been used by numerous companies for developing autonomous vehicles, mining trucks, port logistics vehicles, and other autonomous applications beyond passenger ride-hailing.

Commercial Viability

Approaching profitability in key cities

Apollo Go is approaching commercial viability in several cities where robotaxi operations have reached sufficient scale. In Wuhan, the cost per ride for autonomous vehicles has fallen below the cost of human-driven ride-hailing in certain corridors. Baidu projects that fleet-wide profitability is achievable within 2-3 years as the Apollo RT6 is deployed at scale. The key economics improvement comes from the RT6's sub-$30K cost and elimination of driver compensation, which represents 50-60% of traditional ride-hailing costs.

Side-by-Side Comparison

FeatureBaidu ApolloPony.aiAutoXWaymo
Operational Cities10+ in China5+ (China + US)5 in China4 in US
Ride-hailing Orders8M+3M+1M+4M+
Autonomous LevelL4L4L4L4
Vehicle Cost<$30K (RT6)~$50K~$45K~$100K+
Sensor SuiteLiDAR + camera + radarLiDAR + cameraLiDAR + cameraLiDAR + camera + radar
Business ModelSelf-operated fleetFleet + licensingFleet onlySelf-operated fleet
International ExpansionLimitedUS + Middle EastLimitedUS primary
Key AdvantageScale + costDual marketFull-stackTechnology lead

Frequently Asked Questions

How safe are Baidu Apollo robotaxis?

Baidu reports that Apollo Go has maintained an excellent safety record with no fatalities and significantly lower accident rates per million miles compared to human-driven vehicles. The vehicles are equipped with redundant sensor systems (LiDAR, cameras, radar) that provide 360-degree awareness. However, like all autonomous driving systems, edge cases and unpredictable scenarios remain challenges. Baidu operates with remote monitoring centers that can intervene in complex situations, and all rides currently include safety monitors in some cities.

Can anyone ride Apollo Go robotaxis?

Yes, Apollo Go is available to the general public in its operational cities through the Baidu Maps app or the Apollo Go mini-program on WeChat. Users can hail a robotaxi much like a traditional ride-hailing service, specifying pickup and drop-off locations within the service area. The service has expanded coverage progressively, with some cities offering near-full city coverage. Fares are competitive with traditional ride-hailing services, and in some cases lower due to the elimination of driver costs.

What makes the Apollo RT6 so much cheaper than other robotaxis?

The Apollo RT6 achieves its sub-$30,000 cost through several innovations: sensor miniaturization that reduces LiDAR and computing hardware costs,供应链 optimization through partnerships with domestic suppliers, scalable manufacturing with JAC Motors, and integration of sensors directly into the vehicle design rather than retrofitting. Previous generation robotaxis cost $50,000-100,000+ because they used expensive rooftop LiDAR units and custom-built vehicles. The RT6 uses a retractable steering wheel design that eliminates the need for a dedicated driver compartment.