At the critical juncture of the transition from "single point functionality" to "system intelligence" in automotive intelligence, the role of maps is undergoing profound reshaping. On April 12th, at the 2026 China Electric Vehicle Hundred People Forum, Liu Zenggang, Deputy General Manager of Baidu Maps Business Unit, systematically shared his latest strategic thinking and mass production progress in the AI automotive field.
At the end of last year, Tesla's North American Christmas edition update allowed Grok models to be installed, giving the car's infotainment a "human touch". A clear signal is released that the era of "single point intelligence" where voice, navigation, and intelligent driving fought independently in the past must be broken.
Liu Zenggang proposed that the future AI car will evolve into an all powerful "super AI intelligent agent", which needs to accurately play three roles: an experienced "AI driver", an "AI assistant" that connects car control and service ecology, and an "AI platform" that provides emotional resonance. Maps are transforming from traditional navigation tools to core infrastructure and key bridges connecting intelligent cockpits and intelligent driving.
From "passive execution" to "active understanding": AI cockpit intelligent weight construction for human vehicle interaction
Baidu Maps has released the industry's first end-to-end AI cockpit intelligent agent that has achieved mass production. This product is based on the Baidu Voice Language Large Model, deeply integrating Baidu Maps' 20-year accumulation of large-scale spatiotemporal data and AI capabilities, and combining Baidu Search's precise and efficient information retrieval capabilities. The "central agent" uniformly schedules navigation, entertainment, vehicle control, and chat sub agents, achieving a full loop from oral understanding to complex task execution.
According to data from Baidu Maps, the number of voice chat interactions increased by 13 times after the end-to-end AI voice on the car; The "AI co pilot" function launched on mobile devices has seen a 54 fold increase in daily active users compared to the previous period. It is worth noting that casual conversation has replaced traditional commands as the most frequent interaction scenario, indicating that "AI accompanied travel" is becoming a new habit for users. In terms of travel decision-making, the "travel intelligent agent" supports complex fuzzy location search, natural semantic route planning, and user preference memory, and can respond to highly colloquial and complex instructions such as "find a delicious Ningbo dish on Xierqi Street" or "go to the restaurant I visited last week".
Large scale production of lane level SR navigation: reconstruction of cabin driving integrated visual experience
In response to the pain point of visual disconnection between current navigation and intelligent driving systems, Baidu Maps has taken the lead in achieving large-scale production of lane level SR navigation for all domains. This solution integrates lane level navigation and intelligent driving environment perception data into a unified real-time rendering interface, with precise lane positioning and second level response. It covers 13 million kilometers of roads and millions of parking lot data nationwide, significantly reducing driving anxiety at complex intersections and enhancing the sense of peace of mind for both human and intelligent driving.
At present, this product has been put into mass production in the 3D three-dimensional lane level navigation jointly created by Xiaopeng Automobile, the four figure integrated navigation jointly created by Dark Blue Automobile and wutong Car Service, and the flagship model D19 to be released by Zero Running will also be carried. It is expected that more than 10 car models will be equipped with Baidu Maps lane level SR navigation by 2025. In addition, Baidu Maps has announced that it will launch an immersive parking level SR navigation in the second half of the year, bridging the gap between driving and parking in the last 100 meters of the scene, and achieving a seamless experience of "parking space to parking space" in the entire scene.
Leading the market share of intelligent driving maps: over 40% of NOA vehicles choose Baidu Maps
In the field of intelligent driving, Baidu Maps has identified three major trends: the accelerated popularization of advanced assisted driving, users' pursuit of "parking to parking" full scene experience, and intensified competition for integrated cabin driving. To this end, Baidu Maps supports the safety upgrade and interactive innovation of intelligent driving systems through global lane level data, million level parking lot data, and "driver agent" natural language control capabilities.
According to statistics from Gaogong Intelligent Automotive, in the first 10 months of 2025, Baidu Maps will have a market share of over 41.8% in the Chinese passenger car NOA assisted driving map market, meaning that 4 out of every 10 smart cars will be equipped with Baidu Maps.
For C-end users, Baidu Maps has announced the upcoming V21 May 1st version for mobile devices. The new version will further enrich lane level terrain rendering, upgrade more than 30 over the horizon safety warning signals, and expand the AI co pilot design and skill library, achieving deep integration of thousands of voice packages with precise guidance and heartwarming chat. Baidu Maps stated that it will firmly implement the "hand car integration" strategy, open up the mature innovative functions verified on the mobile end to car enterprise partners as soon as possible, and work together with the industry to accelerate the development process of China's automotive intelligence.
As the focus of competition for automotive intelligence shifts from hardware configuration to experience loop, the strategic value of maps as a key intermediate layer connecting the digital and physical worlds is being re evaluated by the industry. For car companies, how to choose a map partner that can support both cabin interaction innovation and continuous evolution of intelligent driving will directly affect the integrity of their future product experience and market competitiveness.
