Professor Zhixi Wan
28 January 2026
In retrospect, over the past five years, the global automobile industry has undergone dramatic changes in its first half, during which “energy defines the car” was the theme of competition, i.e. a race in battery technology and driving range. Looking ahead to the following five years, the focus in the next phase will undoubtedly shift to “algorithms define the car”, i.e. intelligent driving technology powered by artificial intelligence (AI), big data, and computing power. According to Morgan Stanley’s forecast, by 2030, the market worth of vehicles equipped with partially or fully autonomous driving technologies will reach as much as US$200 billion.
The interplay of advantages and constraints
Currently, the vast majority of new energy vehicles widely available in Hong Kong are still at the L2 (partial automation) stage. As the system provides only assistance, the driver must keep their hands on the wheel and their eyes on the road at all times. In the event of a traffic accident, they also shoulder full legal liability.
In April 2025, XPeng Motors, a Chinese electric vehicle company, chose to hold its Global Brand Night event at Hong Kong’s Kai Tak Cruise Terminal, announcing the mass-production schedule for its L3 intelligent driving functions. L3 refers to Level 3 driving automation (conditional automation), which operates within a specific operational design domain. The system can fully take over the vehicle’s lateral and longitudinal control, and is responsible for object recognition and event response. As a watershed in the transition to advanced intelligent assisted-driving technology, this move marks Hong Kong’s pioneering role in aligning Chinese intelligent driving technology with the global right-hand-drive vehicle market.
While intelligent driving enters its second phase, Hong Kong is well-positioned to leverage its unique “duality”. The city has a road network featuring the world’s most complex road conditions, the highest traffic density, and the most intense interactions between people and vehicles. Narrow, winding hillside roads provide an ideal testing ground for the robustness of intelligent-driving algorithms. At the same time, although the SAR Government published the Smart City Blueprint for Hong Kong as early as 2017, local streets are still dominated by traditional fuel-powered vehicles and EVs with limited driver-assistance functions. Unlike the Mainland’s proactive approach of using industrial development to expand the market and subsidies to promote technology, Hong Kong has long adhered to the policy logic of safety first and primary reliance on market forces. Such a well-established regulatory environment, coupled with the city’s unique status as the only region in China where right-hand-drive vehicles are used, constitutes a distinct advantage in connecting Chinese technology with international right-hand drive markets.
Unlocking the potential of local intelligent driving
Drawing on the growth factors of the AI industry, the barriers to the development of intelligent driving in Hong Kong can be broken down into the following formula:
Intelligence = Computing Power x Data x Algorithms
According to data from the China Association of Automobile Manufacturers and the Ministry of Public Security, as of the end of 2024, the penetration rate of new-energy vehicles in China exceeded 40%, with first-tier cities such as Shenzhen far surpassing this level. In terms of the use of intelligent driving functions, the penetration rate of Navigate on Autopilot among users of leading intelligent vehicle brands in the Mainland has already surpasseed 50%. By contrast, the typical assisted-driving experience in Hong Kong is still limited to such functions as cruise control and lane keeping. The scope of intelligent-driving sandbox testing is fairly narrow, with a severe lack of local data-training closed loops and reliance on cloud computing power.
On the other hand, despite its later start in computing-power hardware (e.g. supercomputing centres), Hong Kong possesses extremely valuable data characterized by “heterogeneity” and “high entropy”. The breakthrough lies in treating road scenarios as a strategic resource. By guiding automakers in an orderly manner to collect data, high-quality traffic-scenario datasets can be built and compliant automakers can be authorized to use them for training. Whoever possesses more high-quality Hong Kong driving data will be able to dominate the algorithmic competition in right-hand-drive markets. Perhaps the key to advancing right-hand-drive testing, while ensuring data de-identification, is to establish a closed-loop data channel dedicated to research and training that allows intelligent-driving data to flow compliantly between the two regions.
Tesla’s Full Self-Driving (FSD) system is currently the widely recognized benchmark around the world for mass-produced intelligent-driving technology. Through shadow mode, Tesla vehicles run the FSD algorithm in the background and compare its operations with the human driver’s actual actions. Should the algorithm determine that a left turn should be made but the human driver goes straight instead, the resulting discrepancy data is uploaded to adjust the model. In the Mainland, FSD has undergone extensive targeted training for scenarios such as vehicles cutting in, electric bicycles crossing the road, and construction zones, and has therefore evolved driving-style algorithms optimized for Mainland driving habits.
Even so, Hong Kong has yet to roll out autonomous vehicle technology. A primary technical challenge is the lack of relevant algorithm training and of algorithms tailored to right-hand-drive traffic patterns. For neural-network-based, end-to-end models, simple image mirroring cannot bridge the gap between left-hand-drive and right-hand-drive systems. The reason is that human drivers have fundamentally different visual habits, blind-spot distributions, and right-of-way conventions, e.g. navigating roundabouts clockwise. In addition, localizing the algorithms requires an understanding of Hong Kong driving behaviour. The algorithms need to learn the city’s unwritten driving rules, such as zipper merging at the Cross-Harbour Tunnel entrance, which requires strong game-theoretic algorithm support rather than simple rule compliance.
An opportunity to shift from a passive to proactive approach
Chinese new-energy vehicles are making a major push overseas, especially into right-hand-drive markets, including the UK, Australia, Japan, Thailand, and Indonesia. At present, almost all test tracks and research and development (R&D) centres in Mainland China operate in left-hand-drive environments. To expand into overseas right-hand-drive markets, autonomous vehicle manufacturers such as BYD and XPeng face enormous R&D adaptation costs. Not only is it costly to build simulated right-hand-drive test facilities in the Mainland but it is also impossible to replicate the complexity of real urban road conditions.
Hong Kong is the only city within China that combines exceptionally complex urban road conditions, a sound judicial environment, and a right-hand-drive traffic system. It should serve not merely as a consumer market, but as a right-hand-drive R&D centre and runway for China’s EV manufacturers to expand overseas. By actively opening up data and testing environments, Hong Kong is ideally placed to continue attracting leading automakers and advanced intelligent-driving technology companies, such as Baidu Apollo, to establish intelligent-driving R&D facilities in the city. This will not only create high-paying technology jobs, but will also enable Hong Kong to transform from a market endpoint into a hub in the global automative value chain.
As the future of intelligent driving draws near, it should not be confined to distant Silicon Valley or the broad Shennan Road. It should take shape in the hustle and bustle of Central, the energy of Mong Kok, and the expanse of the New Territories. With its strategic role as a right-hand-drive testing ground, Hong Kong is poised to gain momentum in the next phase of intelligent driving, evolving from observer into navigator. This will mark not only a triumph of technology, but also a new evolution of Hong Kong’s urban spirit.
