Jia Haonan reporting from Fusi Temple, Quantum Bit ROBO | WeChat Official Account AI4ROBO
Not many players in the autonomous driving sector are still this aggressive today: In terms of scale and mileage, we are the world's largest unmanned fleet, driving over 1 million kilometers per day, several times that of Waymo.
People always say logistics is "unrefined," but no one disputes that we are a robotics company.
Whether it's GPT-6, DeepSeek, or Qwen, they cannot replace me.
The entire L4 sector is calling for a path to commercialization. I can verify this issue next year, and the result of verification will be overall corporate profitability.
Compared with Robotaxi and Robotruck, Robovan must rank first in its pace toward profitability. There is no complex logic; it is simply a straightforward fact.
The above statements were made directly by Neolix CEO Yu Enyuan, one of the leading players in unmanned logistics, immediately after raising $400 million.
Yes, it is $400 million, the largest single funding round in the autonomous driving sector in 2026. Even placed within the broader landscape of embodied intelligence and physical AI, this figure is not small.
A full year prior, Neolix had officially announced raising $600 million, also the largest sum in the entire sector at that time.
Indeed, there is capital for such boldness, but regarding this unmanned logistics delivery startup, you and I likely share a fundamental fact that remains unclear:
They raised the most money, but are they doing the most valuable work?
2026: Neolix Raises the Most Funds in the Sector Again
In the hottest year for robotics and embodied intelligence, Neolix secured funding again, and it was oversubscribed.
The original plan was to raise $200 million, but the final amount doubled. The lead investors were Boyu Investment and the National Artificial Intelligence Industry Investment Fund, the latter making its first-ever investment in the unmanned delivery sector. Participating institutions included Alibaba AEF NextGen Hong Kong Fund, Monolith Capital, GGVCapital, and Vertex Growth under Temasek. Existing shareholders such as Stone Ventures, SoftBank Asia, Yizhuang State-owned Assets Investment, and Gaorong Capital continued to follow on.
Official disclosures indicate two main directions for the use of funds. First, deepening refined operational networks, accelerating the establishment of localized service hubs in major global regional centers, and building a more agile and efficient global customer fulfillment network;
Second, increasing investment in the end-to-end large model data training loop, improving the safety level and vehicle performance of autonomous driving, and continuously reducing the full lifecycle operation and maintenance costs of vehicles.
However, overseas business has independent financing arrangements, which are not included in this $400 million.
A year ago, in October 2025, Neolix just completed a Series D financing round exceeding $600 million, setting the record for the largest single private financing in China's autonomous driving sector at that time.
Combining Series D and Series E rounds over two years, Neolix has received over $1 billion;
As of this round, the cumulative financing amount has exceeded 10 billion RMB, ranking first in the unmanned delivery sector.
Raising money aggressively brings direct questions: What has Neolix actually done? Do these progressions match the status of being the "largest funded player"?
In reality, counting from the first batch of entrepreneurs in2015 and 2016, the L4 sector has existed for ten years. Both the Chinese and US markets are waiting for a sample of successful commercialization.
Therefore, it is not difficult to infer that if investors continue to invest in autonomous driving, they are definitely watching three things: whether commercial implementation holds up, whether overseas expansion can scale, and whether the capabilities validated in L2 using end-to-end models and world models can run through in L4 logistics scenarios.
With $1 Billion in Hand, What Has Neolix Actually Done?
Of course, two months ago, Yu Enyuan emphasized at WRC that Neolix has always been a robotics company. His logic is:
Logistics is generally divided into three links: transportation, loading/unloading, and handling. Everyone acknowledges that logistics is the industry most in need of embodied intelligence. Transportation can be solved by autonomous driving, while loading/unloading and handling rely entirely on robots. This demand is linked together through transportation, forming a real-world requirement.
Saying Neolix raised money as an embodied intelligence player is somewhat reasonable.
But compared to the numerous emerging robotics companies, Neolix remains "out of step."
Everyone talks about implementation. Embodied intelligence players claim easy takeover of certain positions, but Neolix talks about deep integration into the capillaries of the logistics network operations management and TCO (Total Cost of Ownership).
Everyone talks about technology. Embodied intelligence players discuss world models, VLA, etc., achieving SOTA, but Neolix talks about mapless solutions, ROI, and avoiding arms races with giants.
So let's look at it in two layers: what Neolix has done in operations, and what it has done in technology.
In terms of operations, as of October 2026, Neolix has deployed over 30,000 L4-level unmanned delivery vehicles globally, covering nearly 20 countries and over 300 cities, with cumulative autonomous driving mileage exceeding 260 million kilometers, adding over 1 million kilometers daily.
Overseas orders in 2026 exceeded 10,000 units, with fewer than 1,000 delivered; overseas revenue accounted for less than 10%. The Middle East is the fastest-expanding region, expected to operate at a scale of 10,000 units next year.
Scenarios have extended from express delivery to factories, building materials, fresh produce, pharmaceuticals, group catering, and auto parts. In the second half of 2026, order growth in factory logistics and industrial clusters exceeded expectations.
Behind these numbers, Yu Enyuan's judgment on operations focuses on the "capillaries."
Logistics is not a standardized route. Stations, factories, residential communities, merchants, and delivery points are all fragmented nodes, with unplanned routes and vastly different goods.
The more vehicles and denser the cities, the higher the complexity of scheduling, charging, resting, traffic avoidance, and safety compliance.
Scale is the ticket to entry, but scale itself does not automatically equal a moat.
The real difficulty lies in optimizing tens of thousands of granular metrics one by one, ensuring customers reduce costs while the company remains profitable. This cannot be solved by financing amounts; it can only be polished day by day through the operational system.
Therefore, unlike Robotaxi and Robotruck, and unlike other hot robotics companies, Neolix does not "predict scenarios and then deploy." Instead, it turns underlying technology and products into general platforms, allowing ecosystem partners to bring vehicles into different industries.
Patrolling public spaces is a result of this ecosystem-led approach, as is factory logistics.
The popularity of GPT-6 sparked heated debates in the embodied intelligence sector about whether large model players would "wipe out" everyone. However, companies like Neolix, which have long struggled in complex scenarios, are actually not worried.
Yu Enyuan even mentioned cooperation with MiniMax, believing that advances in multimodal large models are beneficial for players focused on implementation.
To summarize, the money spent at the operational layer goes toward building platforms, laying networks, and nurturing ecosystems.
Technically, Neolix took the lead in2025 by deploying L4-level mapless autonomous driving in commercial applications, achieving localization without relying on high-definition maps. This significantly reduced deployment costs and preparation time, which is the underlying reason it could rapidly expand across more than 300 cities.
The scheduling system released at CES 2026 is officially claimed to support stable operation of fleets exceeding 100,000 L4 vehicles.
Product metrics are also improving: comprehensive efficiency increased by more than 3 times, lifespan extended from 5 years to 8 years, and effective daily service hours increased from 12 to 18 hours. Large vehicles X20/light trucks are undergoing small-scale trial operations in closed factories and freight yards.
In terms of technical roadmap, the money is spent on modeling, product iteration, and cloud scheduling.
Moreover, they never bet solely on specific VLA/VLM terminology. The core judgment is that they cannot compete with large companies on technical investment intensity; they must find the path with the best return on investment.
Combining operations and technology, what kind of company is Neolix ultimately?
In operations, it does not pursue exclusive partnerships, does not build massive production capacity itself, shares capacity with SAIC Maxus, Liuzhou Automobile/Wuling, and GAC, and involves ecosystem partners in customer acquisition, O&M, and battery swapping/charging.
Technically, it does not compete with large companies on investment, does not chase VLA/VLM buzzwords, and seeks the path with the best ROI.
Behind these two aspects lies the same cognition: the unmanned logistics sector where Neolix operates is characterized by low leverage, high workload, and long chains. Money should be spent where it forms systemic capabilities, not where it "looks big".
So, the $1 billion indeed resulted in 30,000 vehicles, 300 cities, 260 million kilometers, mapless L4 implementation, scheduling systems, and product iterations.
More importantly, it created an operational network and a methodology for technical investment.
But it has not yet achieved overall corporate profitability, nor has it established a closed-loop commercial model overseas, nor has it fully tested the hypothesis of general vs. vertical competition.
Yu Enyuan's own positioning for the current stage is "still between 0.1 and 1".
Capital has turned into scale, scenarios, technology, and operational systems. Whether scale can turn into sustained profit is what needs to be verified next.
"Profitable Next Year!"
How can a company still between 0.1 and 1, with overseas deliveries just ramping up and technical roadmaps not finalized, bet on reaching the profitability inflection point next year?
Neolix's accounting core is not vehicle sales revenue, nor financing amounts, but TCO—Total Cost of Ownership over the full lifecycle.
This is broken down into approximately 10 key indicators, each further subdivided into dozens. Hardware, quality, safety, technology, cloud services, and O&M are all calculated. For example, vehicles must achieve a basic operational life of 8 years and 400,000 kilometers for the entire model to hold.
This involves detailed metrics related to batteries, motors, repair frequency, etc.
Consider this commercial example: Traditional manual logistics costs 1 yuan per kilometer. The unmanned vehicle's own TCO is approximately 0.25 yuan. Neolix provides services to customers at 0.5 yuan/km. Customer costs drop from 1 yuan to 0.5 yuan, a 50% reduction, leaving room for company profit.
To make this account work, Neolix must continuously improve vehicle utilization efficiency and extend service life.
Yu Enyuan stated that the new generation of products has seen comprehensive efficiency improvements of over 3 times, vehicle lifespan extended from 5 to 8 years, and effective daily service time increased from 12 to 18 hours. With vehicles lasting longer and working more hours per day, purchase and maintenance costs can be amortized over more transport tasks.
Only after the unit economics of a single vehicle are established does the effect of scale come into play.
But here is a often overlooked key point: Fleet expansion increases revenue but also burdens management, scheduling, cloud resources, safety, and service networks. As scale grows, the requirements for system management increase, and costs cannot simply be linearly extrapolated based on the single-vehicle model.
So exactly how many vehicles are needed to cross the break-even point?
Yu Enyuan did not give a clear critical fleet size for profitability, but estimated the market space to be approximately 10 million vehicles. If Neolix captures about 5% of this, i.e., 500,000 vehicles, it would move from "0.1" to "1".
Returning to the entire unmanned delivery industry, its competition will ultimately fall into the "capillaries" of the logistics network. Every delivery route, every type of cargo, every loading/unloading link, and every city's operational requirements can change the cost structure. Autonomous driving capability is certainly important, but operational efficiency, cost control, scheduling management, safety compliance, and partner systems equally determine whether a company can sell vehicles, operate them, and convert revenue into profit.
This is also the difference between Neolix and some players entering from general autonomous driving or robot bodies. Yu Enyuan's long-term experience in logistics, understanding of cargo transport, customer needs, and operational costs, has already permeated the company's product design, technical investment, and cooperation models.
Continued capital investment likely looks at Neolix's industry experience and the comprehensive capabilities formed by the long-term integration of vehicles, technology, scheduling, and operations.
Placed in the broader L4 sector, Neolix indeed has the opportunity to become one of the first startups to achieve overall profitability:
It already possesses a relatively substantial operational scale, can calculate viable TCO around specific logistics needs, and has clear domestic and overseas expansion goals.