Driverless cars are about to hit the road on a large scale.
Neolithics, the world's largest provider of L4-level unmanned urban distribution solutions, has completed a Series D financing round of over $600 million. This is the largest private equity financing in China's autonomous driving field to date.
Moreover, this is the company's second large-scale financing in just 10 months.
Peers in the industry are also raising funds. Jiushi Intelligence completed a $100 million Series B4 financing this month, and White Rhino received additional investment in its Series B+ financing round in August from existing shareholders such as SF Express.
Data shows that the L4-level unmanned vehicle distribution track has attracted over 4 billion RMB in total this year, outshining the trend of "equalization of intelligent driving" in the capital market.
Compared with some concepts that are still in the stage of showy technology, unmanned vehicles have actually reached the stage of large-scale and commercial verification. One of the most important landmark events is that Neolithics is expected to deliver over 15,000 vehicles this year, ten times the number of vehicles delivered last year.
Many industry insiders said that an operational scale of 10,000 vehicles is regarded as the critical point for the explosion of the unmanned vehicle industry. Only after reaching this scale can the industry form a positive cycle of cost reduction through technology, scenario verification, and scale expansion.
As 2025 is gradually coming to an end, it seems that the commercial boom of unmanned vehicles is just beginning.
01 Technical Cost
In essence, an unmanned vehicle is a robot on wheels that can move autonomously in the complex physical world. This year, there has been an obvious trend in the entire robotics field to shift from technological feasibility to commercial viability.
Now, the industry is no longer blindly pursuing full L4/L5-level autonomous driving but is more pragmatically promoting the large-scale mass production of L2+/L3-level assisted driving systems in vehicles. This can not only generate direct revenue but also feed back high-order technology iterations through real road test data from millions of vehicles, promoting a closed-loop of commercial technology.
Cost reduction through technology is the most powerful proof of the shift in the narrative logic of the robotics field, represented by unmanned vehicles, towards commercial viability.
The most expensive part of an unmanned vehicle used to be the hardware represented by sensors and computing units. Now, the prices of these hardware components have generally dropped precipitously, driving the price of unmanned vehicles down from 200,000 RMB to 20,000 RMB.
Take lidar as an example. From 2020 to 2024, the average industry price dropped from 82,000 RMB to 3,900 RMB, a cumulative decrease of 95.2% over five years. In the past two years, the price decline has accelerated. The price of mainstream products has dropped rapidly from the 3,000 - 4,000 RMB range to 1,400 RMB, a decrease of 60% - 65% in just 12 months.
This price drop is not due to ineffective involution among enterprises but rather to effective technological breakthroughs. Among leading enterprises, Hesai Technology's ATX lidar uses self-developed chips to achieve a chip-based and integrated design, reducing the unit cost to less than one-third of the previous generation products through economies of scale.
As the "brain" of an unmanned vehicle, in-vehicle computing chips have also been on a clear cost reduction path in the past two years. Among leading enterprises, NVIDIA's Orin Y chip maintains a computing power of 200 TOPS (meeting the requirements of L4-level autonomous driving) while reducing the price from 3,400 RMB for the Orin X to 2,800 RMB, and there is still room for further reduction.
The price drops of these key hardware components due to technological breakthroughs will further promote the collaborative optimization and technological improvement of chips and algorithms in unmanned vehicles. The evolution of business models has further lowered the application threshold and reshaped the industry's profit logic.
Leading enterprises such as Jiushi Intelligence no longer just sell vehicles but adopt a charging model of vehicle sales + FSD (Full Self-Driving) subscription. Under this model, customers can obtain a vehicle with a down payment as low as around 10,000 RMB and then pay a monthly subscription fee for the autonomous driving system.
For example, the bare car price of Jiushi Intelligence's E6 model is 19,800 RMB, but customers need to pay a monthly FSD subscription fee of 1,800 RMB.
The advantage of this model is that unmanned vehicle enterprises no longer rely solely on hardware sales and can accept selling hardware at a slight loss or break-even to serve the core strategic goal of quickly capturing market share.
However, this model will not cause unmanned vehicle enterprises to actually lose money because the more important profit points lie in the subsequent continuous service fee income and the mining of data value. Taking Jiushi Intelligence's E6 as an example, the total service fee over five years is 127,800 RMB, far exceeding the car price itself. Its founder said, "This business is profitable."
Like all businesses in the world, the business story of robots represented by unmanned vehicles ultimately boils down to the essence of economics - whether it can save costs for customers and improve efficiency.
02 Logistics Cost
For enterprises that use unmanned vehicles, can large-scale use really save costs on a large scale? For logistics companies, this question can now be answered affirmatively starting from this year.
The survival dilemma of China's express delivery industry in recent years is obvious to all. According to the express delivery industry data in the first half of 2025, the average price per parcel is 7.52 RMB, while in 2023, this price was 9.14 RMB per parcel.
Now, as Internet platforms are launching instant retail battles one after another, the express delivery industry will face more pressure on its transportation capacity, and its profits may also be further compressed by the e-commerce industry. In this trend, cost reduction in the express delivery industry is both inevitable and urgent.
A simplified express delivery process is: package collection → feeder transportation → central sorting → trunk transportation → central sorting → feeder transportation → delivery. Currently, the most suitable commercial scenario for the large-scale application of unmanned vehicles is the "from the distribution center to the delivery station" segment in the last-mile delivery of express parcels.
Because the traditional last-mile delivery model faces three major pressures, including the continuous increase in labor costs (in 2025, the average annual salary of delivery workers in first-tier cities reached 102,000 RMB, with an annual growth rate of 6.2%), the fluctuation of fuel prices (in 2025, the oil price remained at a high level of 7.5 RMB/L), and the continuous increase in customers' requirements for delivery timeliness.
Moreover, the traditional last-mile delivery model is limited by the rigid constraints of labor supply and price and does not have economies of scale. In contrast, the larger the scale of unmanned vehicles, the lower the cost.
Some securities firms have estimated that the single-parcel transportation cost of manned vehicles is 0.16 RMB. If unmanned vehicles are used, the single-parcel transportation cost of express delivery can be reduced to 0.05 RMB, saving 0.11 RMB per parcel and further compressing the last-mile cost, with a cost reduction potential of 69%.
Against this background, unmanned vehicle technology can change the situation of large labor cost input by reducing the last-mile delivery cost, providing a solution for the express delivery industry to break the deadlock.
Therefore, unmanned vehicles have attracted large-scale procurement orders from logistics enterprises. For example, SF Express has invested in White Rhino and carried out in-depth business cooperation with it. Currently, the daily average number of active unmanned vehicles in SF Express's system has reached hundreds, and the delivery is still in progress rapidly.
In actual use, some express delivery franchisees said that introducing three White Rhino unmanned vehicles can cover the parcel transfer of five delivery stations, with a daily delivery volume of 5,000 parcels, reducing the delivery cost by more than 50%.
An executive of ZTO Express revealed at the "2025 New Productivity Development Forum for Express Delivery and Logistics" that the company's unmanned vehicle fleet has exceeded 2,000 units. Among them, more than 1,300 units are from Jiushi Intelligence, and the other more than 900 units are from Neolithics and other enterprises.
More than 170 franchisees across YTO Express have used more than 500 unmanned vehicles, of which 67% are from Jiushi Intelligence and 33% are from Neolithics.
Yunda also said in its financial report that the company has invested more than 500 unmanned vehicles in the first half of this year, and the application scenarios in last-mile delivery have been extended to high-value-added fields such as medical supplies and fresh food cold chain. 83% of its unmanned vehicles are from Jiushi Intelligence, and 17% are from Neolithics.
China Post recently purchased 7,000 unmanned vehicles. This is the largest procurement order in the low-speed unmanned vehicle industry to date, attracting bids from enterprises such as Zhongyou Technology (providing unmanned vehicle complete vehicles by White Rhino), Jiushi Intelligence, Neolithics, and Di Shang Tie.
JD Logistics is even more aggressive, announcing that it will purchase 3 million robots, 1 million unmanned vehicles, and 100,000 drones in the next five years and fully invest them in the entire logistics supply chain scenario.
It can be seen that the logistics industry has high expectations for unmanned vehicles to reduce costs and increase efficiency.
03 Social Cost
Both Internet platforms and express delivery companies want to reduce logistics costs. But from the perspective of the whole society, what exactly is the logistics cost? Obviously, it is not just the freight on the express delivery note but includes all aspects such as infrastructure construction, environmental pollution, and public safety.
Macroscopically, there is also an important indicator, the ratio of total social logistics costs to GDP, which is used to measure the economic efficiency of a country or region.
If this indicator has a high value, it means that there is a large "friction" in economic operation, too many resources are consumed during the commodity circulation process, and there may also be problems with the economic structure of the whole society. For example, it may be overly dependent on low-value-added, high-volume and high-weight raw material industries, or the logistics system itself may be inefficient.
Currently, China's logistics cost is still relatively high. Although the ratio of total social logistics costs to GDP dropped from 18% in the past to about 14% in 2023, it is still significantly higher than the average level of 7 - 9% in developed countries.
Therefore, from the government's perspective, it is still necessary to continuously reduce the logistics cost of the whole society - this is also the main reason for the policy support that unmanned vehicles have received in recent years.
In May 2021, the Beijing High-Level Autonomous Driving Demonstration Area issued the first batch of vehicle codes for unmanned delivery vehicles in China, granting corresponding road rights to unmanned delivery vehicles for the first time and achieving another innovative breakthrough in the new product management policy in the domestic autonomous driving field.
Four years later, more than 1,900 districts and counties in China have opened roads for unmanned vehicles. The speed of road right opening for unmanned vehicles far exceeds that of intelligent driving cars. Shenzhen has even achieved city-wide testing permits for L4-level vehicles, clearing the obstacles for large-scale application.
In addition to the fact that legal liability is less likely to cause disputes (the liability subject is clearly defined as the vehicle manufacturer rather than the driver), the more important reason for the policy support for unmanned delivery vehicles is the hope that the emergence of unmanned vehicles can completely solve the core pain point of the "last mile" in the logistics industry.
Without competing for road rights with private cars, unmanned delivery vehicles can effectively improve the smoothness and resilience of the social logistics system, especially when facing the pressure of the epidemic, extreme weather, and high labor cost areas. They fill the weakest and most expensive link in the logistics chain, so they have received policy support at all levels.
But the question is, what lies behind this weakest and most expensive link? Of course, it is the delivery workers and logistics workers who cannot work tirelessly like machines and need their salary increases to keep up with the rising prices.
Perhaps the increase in delivery workers' fees has indeed led to an increase in express delivery costs, and the limit of delivery workers' efficiency has indeed restricted the breakthrough of logistics efficiency. But when we discuss the logistics cost of the whole society, it is obviously inappropriate to blame everything on people, especially in an industry like logistics that has a large demand for grass-roots labor.
04 Conclusion
With the support of big data, algorithms, and other tools, the pursuit of instant delivery by e-commerce platforms and the expansion of e-commerce shopping scale have approached or even exceeded the physical limit of single-person delivery efficiency. This means that it is not that delivery workers are not