"We are standing at a crucial juncture for unmanned operations in special scenarios. The main battlefield of autonomous driving is shifting from the 'competition' on open roads to the 'deep cultivation' of special scenarios."
This judgment by Zhang Dezhao, Chairman and CEO of Velobotics, delivered at the 2026 Zhongguancun Early-Stage Investment Forum, echoes a clear shifting trend in the industry in recent years.
As he recalled: "If we go back five or six years, everyone was competing to see who could reach 120km/h on highways; later, it was who could run in urban areas; and then it was the '100-kilometer intervention rate'. At that time, the competition was focused on the technology itself, and few people talked about commercial implementation."
Now, however, the reality he observes is: "People are gradually no longer talking about how good our technology is; instead, they are talking more about which scenarios we have implemented in and what our market share is in each scenario."

In the face of the long-standing debate in the industry between the "pure vision" and "multi-sensor fusion" approaches, Zhang Dezhao demonstrated the sober insight of a practitioner.
"Many people believe that using LiDAR will drive up costs. This was the case because back in 2015, a single LiDAR unit cost around 100,000 US dollars," he pointed out. "But now, thanks to the tremendous development of the industrial foundation, the price of a LiDAR has dropped to below 3,000 RMB." In his view, the type of sensor is not the core of the issue; "the real essence lies in the underlying systems and models."
He further shared a common aspiration of the industry: "We earnestly hope that end-to-end technology will be the ultimate solution." This means completing decision-making from perception to control directly through a unified algorithmic architecture, which is expected to significantly reduce system complexity and ongoing R&D costs. Velobotics has made a forward-looking layout in this field. As early as April 2024, in collaboration with Tsinghua University and other institutions, it completed China’s first open-road test of a full-stack end-to-end autonomous driving system.
At the same time, he put forward another important trend: the upgrading of intelligence. "Through the collaboration of vehicle, road and cloud, we realize the evolution from 'single-vehicle intelligence' to 'system intelligence'." Future competition will not only revolve around the intelligence of individual devices, but more around the collaborative scheduling of entire transportation systems or machine clusters and the optimization of overall efficiency.
Last but not least is platform-based design, enabling the shift from "deep customization" to "agile reuse". The universal "autonomous driving brain" built by Velobotics can adapt to a variety of vehicle equipment quickly and flexibly, serving as a critical foundation to support the large-scale application of technology across multiple scenarios.

In terms of commercialization strategies, two models have long coexisted in the industry: one is the leapfrog approach that aims directly for the ultimate goal of full autonomy, and the other is the gradual route that advances step by step. Velobotics has clearly opted for the latter.
Zhang Dezhao breaks down the gradual development route into two specific dimensions:
Progressive expansion of application scope: From closed scenarios (e.g., industrial parks, factories) to semi-closed scenarios, and gradually to open roads; from low-speed operation scenarios to high-speed scenarios step by step.
Progressive upgrade of functional levels: Iterating from L2-level assisted driving to L3 and L4 high-level autonomous driving gradually.
"Autonomous driving follows its own objective laws—it is a marathon-style track. Only by acquiring data through scenario applications and then driving algorithm evolution with data can we steadily approach the ultimate goal."
Business Deep Cultivation: Demand-Driven, Building a Three-Tier Product System
Based on the above logic, Velobotics has built a well-structured business architecture centered on solving practical problems.
Tier 1: Safeguarding Lives – The Public Security Sector"Humanity’s most fundamental need is to protect life safety, and life is priceless." Taking emergency rescue as an example, Zhang Dezhao expounded on the fundamental value of technology. In this sector, Velobotics provides core AI controllers and high-mobility basic platforms, empowering special vehicles for fire fighting, patrol, material transportation and other purposes, which have been successfully applied in actual combat tasks such as fire fighting and emergency rescue.
Tier 2: Liberating Human Labor – The Life Service SectorWhen safety is guaranteed, improving labor efficiency becomes a core demand. Velobotics’ product matrix of unmanned cleaning and sanitation equipment has achieved full coverage of cleaning scenarios. Its cleaning robot Viggo has penetrated more than 40 countries and regions worldwide, been deployed in over 100 cities in China, with a cumulative shipment of nearly 10,000 units, and ranks first in the global market share for industrial and transportation scenarios.
Tier 3: Enhancing Experience – The Smart Mobility SectorBuilding on the first two tiers, Velobotics lays out the blueprint for future mobility, currently focusing on unmanned sightseeing and shuttle services in closed areas such as scenic spots and industrial parks. It also creates an integrated vehicle-road-cloud solution, completing strategic layout for the broader smart mobility market in the future.

Underpinning the implementation across diverse scenarios is the robust technical system and unique commercial closed loop built by Velobotics.
"We have accumulated more than 1,230 patents in total, and the majority of which are invention patents," Zhang Dezhao emphasized. Of even greater strategic value is that Velobotics’ real-world commercial operation mileage has exceeded 130 million kilometers. This not only forms a formidable barrier of operational experience, but also creates core data assets that continuously feed back into the evolution of algorithm models.

Finally, Zhang Dezhao condensed Velobotics’ strategic choices into three core tenets:Maintain strategic patience: Revere the objective laws of technological development and market acceptance, and persist in iterating through scenario applications and achieving breakthroughs step by step.Driven by dual values of efficiency: Transform every technological innovation into measurable economic and social value to realize sustainable commercial development.Uphold the original aspiration of value creation: Technology should make human life safer, easier and more fulfilling.
In the long journey of autonomous driving, which tests endurance and foresight, Velobotics chooses to root itself deeply in the industrial ecosystem. It stays committed to solving real-world pain points, quietly awaiting the blooming of commercial value. This path may not be the fastest, but it is designed to be steadier and go further.