In recent years, the primary focus within the self-driving industry has been on Robotaxi services—specifically, which company will be the first to achieve full autonomy, who will scale operations further, and who will successfully establish a comprehensive transportation network.

Robotaxi has assumed the 'leading role' in introducing self-driving technology to the broader consumer market. However, the true measure of whether this technology can be widely adopted in urban public transportation systems likely rests with buses.

A more defined trend is now emerging: Since 2025, self-driving buses in numerous countries and regions have been crossing a significant threshold—transitioning from closed testing to open roads, from technology demonstrations to fixed-route operations, and evolving from concerns about 'whether people will ride them' to 'whether they can integrate into public transportation systems.'

Globally, Seoul, South Korea, has already deployed self-driving buses for late-night, early-morning, and underserved transportation areas, with plans to further expand the fleet by 2029. Singapore is accelerating the integration of self-driving buses into its public transportation network. Japan aims to have a fleet of 10,000 autonomous vehicles, including buses with L4 autonomy, by fiscal year 2030. In Europe, cities such as CAVForth in the UK, Paris in France, Hanover in Germany, Albion in Switzerland, Rotterdam in the Netherlands, and Rogaland in Norway are continuously promoting the entry of self-driving buses into urban open roads and mixed traffic environments. In the United States, states like Louisiana are using public funds to integrate self-driving shuttles into existing public transportation systems.

This signals a crucial transformation:

Self-driving buses are evolving from 'experimental prototypes of future transportation' to 'practical additions to real-world public transportation,' potentially marking their 'golden age.'

What changes are happening with self-driving buses?

Mapping the development of self-driving technology over the past decade reveals a notable shift.

Initially, self-driving buses were primarily focused on 'proving the technology's viability.' Airports, universities, tech parks, tourist attractions, and industrial parks became primary deployment sites. With fixed routes, low speeds, and relatively manageable traffic conditions, these environments helped self-driving companies gradually validate their perception, decision-making, control, and remote operation capabilities.

Today, however, the landscape is changing, and self-driving buses are beginning to enter real urban transportation networks.

By 2026, the global self-driving bus industry will have established two core deployment regions in Europe-America and Asia-Pacific. Multiple countries will incorporate L4 self-driving buses into their urban public transportation systems, achieving a fundamental shift from 'tech exhibits' to 'practical transportation solutions.'

As a global leader in intelligent transportation, Europe has developed a mature deployment model characterized by collaboration among governments, automakers, technology providers, and operators, positioning self-driving buses as a new type of urban public transportation infrastructure.

In August this year, the Netherlands launched the L4 self-driving e-ATAK electric bus at Efteling theme park, developed by Turkish automaker Karsan and software provider ADASTEC. The bus route, operated by Dutch public transportation operator Arriva, provides regular passenger service on a 6-kilometer open road around the theme park, connecting the park entrance, hotel area, and public transportation hub. It is currently open to visitors for free. Previously, Rotterdam's Line 533 autonomous bus operated between Meijersplein Metro Station and Rotterdam Airport.

Stavanger, Norway, achieved Europe's first operation of a self-driving bus without an onboard safety officer under a regular public transportation system. The complete solution is provided by Norwegian company Applied Autonomy, relying on its xFlow platform for remote supervision and fleet scheduling management. The project, initially launched in May 2022, first deployed a pilot route in downtown Stavanger. The route spans 2.6 kilometers, with five stops along the port, passing through four roundabouts and 12 crosswalks. Initially, an onboard safety supervisor was stationed in the vehicle.

Based on continuously accumulated real-world operational data, the local road traffic administration expanded the project in stages: extending the operational route, increasing speed limits, and opening more complex sections such as tunnels. In April this year, regulatory approval was granted to remove the mandatory requirement for an onboard supervisor in self-driving buses across Europe for the first time. Instead, a remote command center was established to uniformly monitor vehicle operation status in real time and implement remote intervention when necessary.

Germany's Hamburg launched the Alike special program, leveraging policy coordination between the German Federal Ministry of Transport and local governments to promote the deployment testing and commercial operation of multiple self-driving buses, such as the HOLON Mover. The HOLON Mover's operational area covers 37 square kilometers of urban districts, connecting city parks, the Elbe Riverbank, and linking Schnelsen to Wandsbek. The project will gradually open to the public for manned experiences, integrating self-driving capabilities into the city's regular public transportation network and accumulating real-world road testing data for large-scale commercial operations.

In addition to Hamburg, Hanover and Leipzig are also rolling out self-driving bus pilot programs. Hanover's public transportation operator, stra, launched Germany's first fully autonomous medium-sized electric bus, the Karsan autonomous eATAK, for regular testing on fixed routes. Leipzig's transportation authority started earlier with the Absolut self-driving project in 2022, using Volkswagen Crafter small electric vehicles to operate between the suburban S-Bahn station and the BMW plant area on a 7-kilometer route. All vehicles are equipped with safety officers and connected to a dispatch center for unified remote monitoring.

North America stands as one of the primary regions for self-driving bus applications globally, leveraging its mature tech innovation ecosystem and government support policies to lead in technological development and scenario deployment. Governments at all levels in the United States and Canada continue to increase financial investment in self-driving technology, supporting leading self-driving companies like Waymo and Local Motors through specialized subsidies, pilot exemptions, and road network openings to deeply cultivate urban microcirculation buses, scenic area commutes, park shuttles, and other diverse scenarios, forming a virtuous industrial ecosystem of 'technological innovation + scenario deployment + policy support.'

Self-driving buses in North America are mostly concentrated on 'first-mile' and 'last-mile' shuttles, commonly used on university campuses, commercial districts, airport terminals, and fixed low-speed routes in some cities. For example, Riverside, California, deployed fully self-driving electric shuttle buses in 2025. Atlanta and other places have also launched manned public transportation pilot programs in collaboration with companies like Karsan and Beep, while Volkswagen subsidiary MOIA America and Uber initiated road testing of self-driving minibusses using the ID.Buzz model in Los Angeles.

In June this year, Turkish automaker Karsan, in partnership with technical collaborator ADASTEC, North American distributor Damera, and mobility service provider Beep, officially launched manned operations of self-driving buses in Atlanta, USA, establishing the region's first self-driving public transportation pilot project, 'ATL Spoke.' The self-driving e-JEST minibus deployed in Atlanta connects the MARTA West End Station to the Southwest Atlanta BeltLine, precisely addressing short-distance urban commuting gaps.

The Asia-Pacific region has emerged as the core market with the fastest growth and largest scale for global self-driving bus deployment. China, Singapore, South Korea, and other countries have initiated large-scale self-driving bus pilot programs, accelerating deployment across the board.

For Singapore, given its limited land resources and long-term labor shortages, self-driving public buses are not just a cutting-edge exploration but a structural necessity for the local public transportation system. As a key support, the '2040 Land Transport Master Plan' explicitly identifies self-driving technology as a core strategy for upgrading public transportation, proposing the construction of a smart transportation network with self-driving as a key component by 2040.

In March this year, Singapore's first right-hand-drive self-driving bus model was officially launched, custom-developed for the Singapore market by Chinese self-driving company Mushroom Auto. The vehicle adopts a 'front-assembly mass production + fusion of visual perception and solid-state LiDAR' technical route, leveraging a mature mass-production system and engineering capabilities to rapidly adapt the self-driving solution to the market while controlling costs. It also accumulates large amounts of proprietary data tailored to bus models, effectively shortening algorithm training and vehicle adaptation cycles. Currently, the model has passed Singapore's M1 testing certification and is capable of large-scale delivery and overseas regular operations.

South Korea's self-driving buses have already been put into actual operation and trial runs in major cities like Seoul and Busan, gradually moving toward full autonomy, regular fare collection, and diversified routes. In August this year, Seoul announced a new phase of its transportation strategy, planning to expand its self-driving bus fleet to 500 vehicles by 2030 and gradually replace some existing city buses. South Korea also has a dedicated self-driving large-scale simulation test city, 'K-City,' for continuous upgrading and testing of Level 3+ self-driving technologies.

To serve early-morning commuters, South Korea has launched dedicated 'Early Morning Companion' self-driving bus routes, with departure times 30 minutes earlier than regular bus schedules. The A160 route (Dobongsan Station-Yeongdeungpo Station), which began operations in 2024, and the A741 route (Jubong Station-Yangjae Station), launched in March this year, have collectively carried over 29,500 safe passengers. In April this year, the A148 route (Sanggye Station-Express Bus Terminal) officially commenced operations. Spanning 22.1 kilometers, the route adopts an express operation mode, stopping only at key stations for round-trip service, and is expected to reduce travel time by approximately 15 minutes per trip compared to regular bus routes.

With a complete industrial chain, diverse scenario resources, and layered support policies, China has become the core region for large-scale global deployment of self-driving buses. Self-driving buses in China follow a phased deployment pattern of 'cultural and tourism scenic areas first, followed by urban public transportation,' and have begun exporting complete technology and operational systems overseas.

2025 is regarded as a critical year for China's self-driving industry to transition from closed testing to large-scale operations. On July 20, the Ministry of Industry and Information Technology, in conjunction with the Ministry of Public Security, issued the 'Administrative Measures for Road Testing and Demonstration Applications of Self-Driving Vehicles (2025 Revised Edition),' explicitly allowing, for the first time, L4 self-driving vehicles to conduct citywide commercial operations in five cities, including Shanghai, Hangzhou, and Shenzhen.

As of now, Guangzhou has opened 21 self-driving bus demonstration routes, Chengdu has launched a cumulative total of 8 self-driving bus routes, and Hangzhou has regularly operated L4 self-driving buses in multiple districts, including Binjiang, Yuhang, Qiantang, and Xiaoshan. Shenzhen's self-driving bus routes are mainly concentrated in core areas such as the Qianhai Cooperation Zone and Luohu District, while Nanjing's L4 self-driving buses primarily serve specific areas and routes like universities and parks. Domestic self-driving bus routes generally adopt a 'reservation + fixed stop' model, with onboard safety officers and remote driving monitoring centers for emergency takeover.

Why are self-driving buses more likely to create 'public value'?

Many people used to hold a misconception that Robotaxi services would be the first to achieve large-scale commercialization.

Because taxis have higher unit prices and more flexible business models, they seem more likely to generate direct market revenue.

However, from an industrial deployment perspective, self-driving buses actually have a greater advantage: they don't need to solve 'self-driving on all roads worldwide' at once but only need to solve one route first.

A Robotaxi faces highly random starting and ending points, as well as a large amount of unpredictable urban traffic environments. In contrast, a self-driving bus mostly operates within a defined operational design domain (ODD). Routes can be planned in advance, stops can be predetermined, road conditions can be continuously collected, vehicles can be centrally dispatched, and even vehicle capabilities can be further enhanced through cloud control systems.

This significantly reduces the complexity of self-driving commercialization. More importantly, public transportation itself is a highly structured traffic product. Where passengers board and alight, which routes vehicles take, and departure and arrival times are all highly predictable.

Therefore, self-driving buses are not a 'downgraded version of Robotaxi.' On the contrary, they are likely one of the scenarios where L4 self-driving technology can most easily form a complete commercial closed loop. This is why more and more cities are incorporating self-driving buses into their public transportation plans rather than treating them as simple tech demonstration projects.

Kevin DeGood, Director of Infrastructure Policy at the Center for American Progress in Washington, D.C., pointed out that buses are more suitable for self-driving technology deployment than taxis, as the two have fundamentally different impacts on urban congestion.

'During rush hours, a single bus can transport numerous passengers, whereas self-driving taxis typically cater to only a handful,' he remarked. Self-driving taxis often encounter the problem of empty cruising, which exacerbates road traffic congestion. Conversely, self-driving buses, with their high passenger capacity, can significantly enhance overall urban traffic efficiency.

This indicates that the same self-driving technology, when implemented in buses, inherently possesses a more substantial 'public transportation multiplier effect.'

To put it simply, if self-driving technology reduces labor costs for a single vehicle, it unlocks a certain amount of transportation capacity. In the case of buses, this capacity can be utilized by a greater number of passengers. Thus, the true economic value of self-driving buses lies not merely in cost savings on a single driver but in enabling a limited set of public transportation resources to serve a broader population.

This is also why self-driving buses have garnered unanimous support from governments worldwide—not merely as a technological fad but as a pragmatic strategic choice grounded in population structure, traffic challenges, industrial upgrading, and urban governance.

Global public transportation faces a common dilemma: recruiting drivers is becoming increasingly difficult.

Research by the International Association of Public Transport (UITP) reveals that approximately 9 million individuals are currently employed in global urban public transportation, yet there remains a shortfall of about 2.4 million workers compared to the ideal level required to match population growth and public transportation demand. The European public transportation industry also estimates a driver shortage of around 10%.

The situation in the United States is equally challenging. The American Public Transportation Association (APTA) highlights that North American public transportation agencies have long grappled with shortages of bus drivers and maintenance personnel, with 85% of surveyed agencies still experiencing labor shortages and 42.7% of public transportation employees aged 55 or older.

This poses a significant challenge for global public transportation operators: Cities aspire to increase bus frequencies, extend operating hours, and address transportation issues in remote areas and 'last-mile' connectivity, but where will the drivers come from?

Self-driving technology offers an alternative solution. It does not necessarily imply the elimination of all drivers but rather a transition from the traditional model of 'one vehicle, one driver' to a model of 'self-driving + remote supervision + centralized operations.' L4 self-driving buses can operate 24/7 without a driver, enabling intelligent scheduling and remote operations, thereby completely decoupling public transportation capacity supply from traditional human labor. This is also the primary reason why countries are actively opening their road networks and introducing mature self-driving bus solutions.

Moreover, the traditional urban public transportation system exhibits notable structural shortcomings. While major bus routes can cover the core thoroughfares of cities, niche scenarios such as community microcirculation commuting, short-distance shuttle services in tourist areas, scattered routes in suburban areas, and commuting in industrial parks and ports have long suffered from sparse capacity, coverage gaps, and inconvenient travel due to dispersed passenger flows, high single-line operational costs, and difficulties in manual scheduling. These issues represent universal pain points in the global public transportation sector's efforts to enhance quality and upgrade services.

Compared to traditional buses, autonomous buses offer significant advantages, including flexible deployment, intelligent and efficient scheduling, and strong scene adaptability. They can precisely address service blind spots in traditional public transportation, establishing a multi-dimensional, grid-based public transportation system that combines "mainline public transportation with microcirculation driverless buses," thereby comprehensively improving the coverage and accessibility of urban public transportation.

This value has been fully validated through landmark implementation projects both domestically and internationally. Multiple cities in China have enhanced their urban microcirculation commuting networks by deploying autonomous buses. In the stringent overseas premium market, a domestically produced autonomous driving solution implemented in Singapore has undergone localized algorithm strategy iterations and parameter calibrations tailored to the local high-density urban road network and refined travel demands. It fully adapts to Singapore's unique localized traffic rules, such as circular island travel and right-of-way yielding, effectively undertaking the connecting and diverting functions of urban mainline transportation while addressing service gaps in short-distance commuting within specific areas. This fully proves the core value of autonomous buses in improving the global public transportation road network system.

Facts have demonstrated that autonomous buses are no longer merely a 'localized innovation' of a single country but are becoming a new type of public infrastructure commonly explored within the global public transportation system.

Four Potential Transformations in Autonomous Buses

In the short term, autonomous buses serve as an effective supplement to urban public transportation; in the long run, they will fundamentally restructure the operational models, service systems, and development landscapes of global public transportation, becoming a cornerstone of modern intelligent transportation systems.

As autonomous buses accelerate their global deployment, four significant changes may emerge in this field over the next 3-5 years:

First, the expansion from 'fixed routes' to 'dynamic routes.' Currently, the vast majority of autonomous buses still rely on fixed routes. However, with continuous advancements in autonomous driving perception and decision-making capabilities, coupled with improved capabilities of cloud-based large-scale scheduling platforms, 'regular buses + on-demand buses' will become the mainstream new model for future urban public transportation. Regular buses will continue to handle high-volume mainline commuting, ensuring the primary passenger flow channels in cities. Meanwhile, dynamic on-demand buses can flexibly adjust their routes and stops based on real-time orders and passenger flow distributions, filling in the gaps in community microcirculation and short-distance connections that mainline buses cannot cover, thus becoming a proactive, responsive mode of public transportation.

Second, the transition from the 'safety officer era' to the 'remote supervision era.' Having safety officers on board is a common transitional solution for the current commercial deployment of autonomous buses. However, this approach does not truly eliminate labor costs, as vehicle operation hours and scale are constrained by staffing levels and shift rotations, making it difficult to achieve large-scale profitability. In contrast, the remote supervision model no longer requires onboard personnel for each vehicle; instead, supervisors in a backend remote operations center can monitor multiple vehicles simultaneously. However, remote supervision also imposes higher requirements on network communications, cloud control platforms, emergency takeover mechanisms, and safety redundancy systems, necessitating simultaneous improvements in technology, regulations, and safety standards.

Third, the transformation from a 'new transportation project' to 'public transportation infrastructure.' Looking abroad, some countries no longer treat autonomous buses as standalone innovative pilot projects but as integral components of regular public transportation, incorporating them into bus operation systems, financial subsidy frameworks, and daily urban traffic management. This shift signifies that autonomous driving is moving beyond the realm of specialized innovation policies. It is no longer an experimental product of tech companies but has become urban public transportation infrastructure, assuming the same role as traditional buses in providing inclusive travel services and undergoing unified industry regulation, assessment, and maintenance management, thus truly completing its transition to normalized public services.

Fourth, the evolution from a 'transportation tool' to an 'urban governance basis.' Autonomous buses will not merely serve as travel tools but also as mobile sensing terminals for smart cities. Massive amounts of driving data, road condition data, and passenger flow data can be transmitted in real-time to urban traffic platforms, providing precise data support for urban traffic planning, road network optimization, congestion management, and public resource allocation. While serving the public travel needs of citizens, they continuously sense the operational status of the city, transforming travel carriers into mobile data entry points for urban governance. This provides objective, dynamic real-world references for refined urban governance, promoting bidirectional integration and mutual empowerment between transportation and urban governance.

Over the past decade, the dominant narrative in the autonomous driving industry has been whether 'machines can replace human drivers.' In the next decade, the more crucial question will be whether 'machines can redefine the supply models of urban transportation.'

It must be acknowledged that autonomous buses have not yet reached a mature stage today, with many projects still in pilot and validation phases. However, this does not detract from the fact that they are entering a critical window of opportunity.

More importantly, autonomous buses do not need to completely overhaul the existing public transportation system from the outset. They can prioritize addressing the most challenging issues faced by traditional public transportation: driver shortages due to an aging population, high costs of nighttime routes, lack of drivers for short-distance connections, and unresolved 'last-mile' issues after the opening of subway lines. As these 'small problems' are gradually resolved by autonomous buses, the urban public transportation system itself will undergo transformation.

The 'golden age' of autonomous buses does not signify that they have achieved fully mature commercial deployment. Instead, it means that a mode of transportation previously confined to concepts and blueprints now possesses all the necessary conditions to integrate into urban public transportation systems, transitioning from a distant future vision to real-world urban road networks.

And this transformation has already begun.