Introduction

On October 27, the 8th Intelligent Driving Conference & 2026 Huanxiuhu Embodied AI Industry Development Promotion Conference took place in Xiangcheng, Suzhou.

Cao Cao Mobility unveiled a steering wheel-free native Robotaxi prototype on-site and revealed plans for its mass production by 2027.

On the same day, Suzhou unveiled the nation's first regional three-year action plan for the embodied AI industry: by 2029, the scale of embodied AI-related industries in the High-Speed Rail New City is projected to surpass 80 billion yuan, with 500 enterprises and 10,000 talents congregating.

Driverless Cars Are Coming (WeChat ID: DriverlessCarsAreComing) posits that Suzhou is integrating the daily reality of "hailing a driverless car" with the ambition of "transforming the entire city into a testing ground" into a unified vision.

(For further reading, please click: "Wuzhong District, Suzhou: A 13.5-Kilometer Driverless Bus Route Arrives! Tian Tong Wei Shi's L4 Autonomous Buses Unlock Smart Mobility in Urban 'Capillaries'")

I. Suzhou Redefines the Game with a New Term and Strategy

Suzhou has long been at the forefront of intelligent driving.

The intelligent connected new energy vehicle industry in the High-Speed Rail New City has achieved an output value exceeding 38 billion yuan, amassing over 280 enterprises across more than 30 sub-sectors.

More than 1,200 test and demonstration vehicles have been deployed, along with 11 driverless bus routes and over 600 kilometers of intelligent connected roads constructed. This infrastructure ranks among the top nationwide.

However, Suzhou has now elevated the concept of "intelligent driving" to a new echelon: embodied AI.

The difference extends beyond mere terminology.

Intelligent driving focuses on "how vehicles perceive roads and make decisions," whereas embodied AI addresses "how machines comprehend the physical laws of the real world."

Liu Zhiyuan from Tsinghua University succinctly summarizes it: embodied AI endows machines with "a brain that comprehends the real world" and "a body that applies physical laws."

The phrase in Suzhou's three-year action plan—"transitioning from a 'city of intelligent driving' to a nationally leading 'city of embodied AI'"—is not mere rhetoric but a strategic pivot.

The competition in intelligent driving has entered a phase of cost and operational battles, with visible limitations; embodied AI represents a broader framework, with intelligent driving as its most mature application but with boundaries extending to embodied intelligence, industrial simulation, and low-altitude economies.

Whoever establishes the foundation first will shape the next round of industrial division.

II. Cao Cao's Native Robotaxi: 200,000 Yuan as the 'Standard' Starting Point

The most attention-grabbing exhibit at the conference was the steering wheel-free autonomous Eva Cab prototype co-developed by Cao Cao Mobility and Geely.

Lacking a steering wheel or front passenger seat, it resembles a mobile living room.

Yan Yonggui, Vice President of Cao Cao Mobility, outlined a clear timeline: mass production by 2027, with "hailing a driverless taxi becoming commonplace in the future."

Gong Xin, CEO of Cao Cao Mobility, further designated Suzhou as the nation's first model city to simultaneously deploy Robotaxi and Robovan services. The Robotaxi fleet has already surpassed 200 vehicles across six cities, with the third-generation native model slated for mass production in 2027 and a cumulative deployment of 100,000 units by 2030.

The significance of the word "commonplace" lies in the cost structure.

Zhao Chenhui, CTO of Cao Cao Mobility's RoboX Division, defines 200,000 yuan as the commercialization threshold:

Vehicle costs across the industry have plummeted from over 1 million yuan to the 200,000 yuan range, yet order revenue still struggles to cover full lifecycle costs.

The core rationale of native solutions is not "making cars cooler" but redefining them from the chassis and electronic architecture upward for autonomous operations.

Retrofitted vehicles cram autonomous systems into shells designed for human driving, while native vehicles are built from the ground up with interfaces for battery swaps, cleaning, dispatching, and billing.

Cao Cao Mobility's "Green Smart Mobility Island" integrates automatic battery swaps, cleaning, vehicle preparation, intelligent dispatching, and automatic billing—essentially treating Robotaxi as an operational unit rather than a standalone vehicle.

III. Why Embodied AI Finds Its First Footing in Mobility: Because Data Here Is Most 'Valuable'

Suzhou's decision to initiate embodied AI in mobility is no accident.

The competition in embodied AI ultimately hinges on three factors: real data, real scenarios, and real industrial chains.

Intelligent driving happens to be the domain with the highest density of all three.

A Robotaxi generates daily data encompassing real-world physical interactions: cardboard boxes falling from preceding vehicles, pedestrians suddenly darting across the road, rain-blurred lane markings, and nighttime glare from oncoming headlights.

These long-tail scenarios cannot be replicated in labs or fully approximated in simulations; only real-road operations yield raw training materials.

Momenta provides a reference path.

Its R7 reinforcement learning world model debuted in mass production in April 2026, positioned as the "prologue to embodied AI," with its Robovan business already operational in Xiangcheng, Suzhou, expanding world models from passenger vehicles to autonomous delivery.

The world model's role is not "helping vehicles recognize more objects" but enabling them to predict unseen physical laws:

If a preceding vehicle drops a box of apples, the system infers the apples' rolling trajectories and dispersion range, initiating smooth deceleration and rerouting in advance—rather than waiting for sensors to detect the apples before emergency braking.

This "anticipating one step ahead" capability relies not on rule-stacking but on the world model's understanding of physical laws.

And Suzhou provides the richest soil for this logic.

Over 600 kilometers of intelligent connected roads, 7,000+ connected intersections, and 1,200+ test vehicles matter not just for "comprehensive infrastructure" but for forming a real-world testing ground.

Embodied AI demands data not annotated in offices but generated at these intersections, amid this traffic, and under real rain, snow, and glare.

The most thought-provoking aspect of Suzhou's conference is not the 80 billion yuan industrial target or the steering wheel-free prototype. It's the decision to elevate the focus just as "intelligent driving" reaches its peak in hype.

In short, Driverless Cars Are Coming (WeChat ID: DriverlessCarsAreComing) argues:

Embodied AI truly materializes only when hailing a driverless taxi becomes routine.

And Suzhou aims to ensure the entire city is prepared when that day arrives.

What do you think? References: Reports from CNR, Online Taxi Observatory, Electric New Species, Lizhi News, Xinhua Daily, and other media outlets.

#DriverlessCarsAreComing #Driverless #AutonomousDriving #DriverlessVehicles