On June 12, Trunk, a provider of L4 autonomous driving trucks and solutions, submitted its second IPO application to the Hong Kong Stock Exchange, aiming for a main board listing. This follows the expiration of its initial application in December 2025. After multiple rounds of industry consolidation in China's intelligent driving sector, very few companies have successfully transitioned from "technology demonstrations" to generating real revenue. As one of the few L4 autonomous heavy-truck companies listing in Hong Kong, Trunk's IPO is not only crucial for its own survival but also serves as a key barometer for whether China's high-level autonomous driving can cross the "valley of death."
Facing the scrutiny of the capital markets, Trunk has presented a seemingly contradictory yet compelling performance report: revenue grew by 157% over three years, gross margin jumped to 27.1%, yet cumulative net losses still exceed 570 million yuan. While its technology indicators lead the industry, it has yet to escape negative operating cash flow. This is a true snapshot of China's hard-tech startups today—struggling to break through industry waves while cautiously surviving a capital winter. As pure technology narratives lose their appeal and valuation logic shifts to hard metrics, and as more competitors like DeepWay and ZeroOne Vehicle crowd into the track, whether Trunk can use its IPO proceeds to ease operational pressures and accelerate overseas expansion to create a second growth curve will directly determine if it can remain at the table during the "pre-commercialization" reshuffle. This critical sprint is not only an ultimate test of Trunk's commercialization capabilities but also a pivotal battle for China's high-level autonomous driving industry.
From Ports to Highways: Trunk's Journey
Trunk's story began in 2017. Founder Zhang Tianlei, a Tsinghua University computer science PhD and a veteran of Baidu's early Apollo project, chose an unconventional path: instead of focusing on passenger car Robotaxis, he concentrated on commercial vehicles, specifically heavy trucks—a high-value, high-efficiency, and high-certainty scenario. At that time, while open-road passenger car scenarios offered immense potential, the technical long-tail issues were unsolvable, regulatory liability was unclear, and consumer willingness to pay was low, making a business model seem distant. In contrast, the commercial vehicle logistics track offered inherent advantages of "high value, high efficiency, and high certainty." Ports and highway routes have fixed paths and clear rules, customers have a clear willingness to pay for cost reduction and efficiency gains, and the industry faces a structural shortage of labor.
This choice essentially used "scenario certainty" to hedge against "technology uncertainty," pulling L4 autonomous driving from "technology demonstrations" back onto the track of "commercial deployment." The foresight of this decision proved accurate. Compared to the complex interactions of urban open roads, closed or semi-closed environments like ports, mines, and logistics parks have lower regulatory hurdles and higher standardization, making them naturally suitable for L4 autonomous driving. In 2018, Trunk launched the world's first unmanned electric truck at Tianjin Port. By 2020, its unmanned container truck fleet was operating routinely at Ningbo Zhoushan Port. In 2021, 60 ART (Artificial Intelligence Transport) robots were delivered to Tianjin Port's "Smart Zero-Carbon Terminal," becoming the core transport capacity of the world's first fully unmanned container terminal.
As of the end of April 2026, Trunk has delivered a cumulative total of 1,283 AiTruck smart trucks and 381 AiBox smart terminals, establishing a stable base in the port scenario. Beyond ports, starting in 2024, Trunk expanded into highway logistics, launching the Trunk Pilot (road logistics) solution. It became one of the first companies in China permitted to conduct cross-provincial routine operations of intelligent connected heavy trucks on the Beijing-Tianjin-Tanggu Expressway. According to the prospectus, in 2025, revenue from the Trunk Pilot business reached 215 million yuan, a 412% year-over-year increase, accounting for 62.5% of total revenue. Meanwhile, the port business, Trunk Port, contributed 128 million yuan, or 37%. This indicates Trunk has successfully completed a key transition from "closed-scenario validation" to "open-road monetization."
The backbone of this transition is its self-developed AiTrucker system. This system, centered around an end-to-end VLA large model and integrating multiple sensors and a high-performance computing platform, possesses human-like cognitive abilities of "see-think-act." It is not a single software algorithm but a unified, soft-hardware integrated, vehicle-cloud collaborative technology base. Its core value lies in technology reuse and scenario adaptation. Leading logistics companies like Deppon, SF Express, J&T Express, and Shentong have already integrated it into their highway transport networks. A Frost & Sullivan report indicates that by 2025 revenue, Trunk ranked fourth in China's commercial vehicle autonomous driving solutions market with a 2.7% share. In closed-road scenarios, however, its market share is 31.8%, holding the top position.
Growth Concerns and Hidden Risks
Despite the achievements, the risks are also clear. Trunk's financial structure reveals the typical characteristics of a hard-tech startup: high growth accompanied by high dependency, with high investment yet to yield positive cash flow. The prospectus shows that from 2023 to 2025 (the "reporting period"), Trunk generated revenues of 134 million yuan, 254 million yuan, and 345 million yuan, respectively. Net losses were 213 million yuan, 187 million yuan, and 171 million yuan. A notable shift in business is that highway logistics became the largest revenue source in 2025, accounting for over 62%. In this sector, the company has partnered with major logistics clients like Deppon, Shentong, SF Express, and JD.com, with autonomous driving freight routes covering regions like Beijing-Tianjin-Hebei, the Yangtze River Delta, and the Greater Bay Area.
However, behind the improving data, the issue of high customer concentration remains severe. In 2024, the top five customers accounted for 67.9% of revenue, with the largest single customer at 30.2%. This concentration decreased in 2025, but the top five customers still contributed 48.6% of revenue, making the company highly vulnerable to changes in any single client relationship. In the port and highway logistics sectors, clients are often large state-owned enterprises or logistics groups, whose investment in intelligence is significantly influenced by macroeconomic conditions and policies. If a core client changes its strategy or ends a partnership, Trunk's revenue could face drastic fluctuations. Especially in the current economic climate, large enterprises are becoming more cautious about smart technology investments, potentially slowing order pace and amplifying the operational risks associated with customer concentration. This "big customer dependency" not only weakens the company's bargaining power but also constrains its ability to dictate payment terms, exacerbating cash flow pressure.
Furthermore, the company's true profitability has not yet been established. While gross margin improved significantly from 12.2% in 2023 to 27.1% in 2025, losses continue, with cumulative losses exceeding 570 million yuan over three years. More critically, operating cash flow has been negative for three consecutive years, with net outflows of 83 million yuan, 86 million yuan, and 80 million yuan during the reporting period. The balance sheet also looks strained. By the end of 2025, Trunk's net liabilities stood at 1.165 billion yuan, and trade receivable turnover days reached 166, reflecting long collection cycles and significant capital tied up. Although cash on hand increased 161% year-over-year to 167 million yuan in 2025, this was primarily from financing activities, not operational cash generation. These issues point to a common problem: the company is still a considerable distance from sustainable profitability. As the capital market's valuation logic for autonomous driving companies shifts from "competing on technical parameters" to "competing on hard operational data metrics," Trunk must find a balance between maintaining R&D investment and improving cash flow, accelerating the conversion of its order backlog into scaled revenue to bridge the gap from technological leadership to commercial leadership.
Charting a Path Forward in a Critical Window
Fortunately, Trunk's decision to restart its IPO now coincides with a period of concentrated policy support from China's capital markets for the securitization of hard-tech assets. The autonomous driving commercial vehicle track on the Hong Kong Stock Exchange remains hot, with companies like CiDi and UISEE successfully listing, and other leaders like DeepWay, ZeroOne Vehicle, and EACON also in the process of pursuing a Hong Kong listing alongside Trunk. From an industry trend perspective, Frost & Sullivan predicts that China's commercial vehicle autonomous driving solution market for closed-road scenarios will grow from 3.4 billion yuan in 2026 to 29.6 billion yuan in 2030, a compound annual growth rate of about 72%. The open-road scenario is expected to grow from 17.3 billion yuan to 217.3 billion yuan, becoming the dominant force. Autonomous driving in commercial vehicles has gradually shifted from an option to a necessity.
As Trunk pursues its Hong Kong IPO, the stated use of proceeds is clearly focused on its core business, including: continuously enhancing core R&D capabilities, building a smart equipment and core component mass production and supply chain system, market expansion and product brand ecosystem development, strategic equity investments and acquisitions, and supplementing cash flow. The trillion-yuan logistics track is fiercely competitive, and Trunk is also eyeing overseas markets to find a second growth curve through global expansion. The company has reportedly adapted its products for markets in Southeast Asia, South America, and the Middle East.
Of course, challenges persist. In closed scenarios like ports and mines, L4 heavy trucks can operate routinely. However, on open highways, facing complex mixed traffic, extreme weather, and non-standardized road conditions, bridging the safety gap from "99% to 99.9999%" requires not only breakthroughs in frontier algorithms like end-to-end large models but also vast amounts of real-world operational data, creating a very high technical barrier. Currently, the road rights for L4 heavy trucks on national open roads are not fully granted, and cross-provincial highway operations face inconsistent access standards. More critically, in the event of an accident, a clear legal framework for defining liability between the vehicle manufacturer, technology provider, and operator is still lacking, and insurance pricing models are also incomplete. This creates a Sword of Damocles hanging over commercialization. Until the entire industry achieves cost reductions in hardware, autonomous heavy trucks will always face a difficult cost and profitability threshold.
In the autonomous driving field, the "valley of death" is a harsh industry metaphor for the transition period from a "lab demo" to "large-scale commercial deployment." Currently, L4 unmanned heavy trucks are deep in this valley, facing a four-frontal assault from technology, cost, regulation, and capital. Trunk now stands at the crossroads of large-scale commercialization. An IPO is not the finish line but a new starting point. For the entire industry, Trunk's success or failure will serve as a classic case study. If it can use this opportunity to optimize its financial structure, diversify its customer base, and accelerate global deployment, today's prospectus might just become the opening chapter of Trunk's journey to cross the "valley of death."
Comments