DeepWay

1.1 Identity, Business Model, and Strategic Context

DeepWay Technology Co., Ltd. was incorporated in July 2020 as a joint venture between Baidu, Inc. and Lionbridge Financial Leasing Group. The company is headquartered in Hefei, Anhui Province, China (Hefei headquarters formally established November 2023) with R&D operations in Beijing and assembly plants operated through contract manufacturing partners. The corporate website and HKEX application proof describe DeepWay's mission as developing "smart NEV heavy-duty trucks" that are forward-engineered from blank-sheet for electric powertrains and L2–L4 autonomous driving, rather than converting diesel incumbents. The business model has three intended phases. Phase 1 (current primary driver): vehicle sales of the Xingchen and Xingtu tractor series, with hardware revenue comprising over 97% of 2025 revenue. Phase 2 (scaling): Tianji software subscriptions—the Suixing L2 ADAS subscription (priced at RMB 3,000–5,000 per year) had over 7,500 activated vehicles and a 30%+ uptake rate as of the May 2026 filing. Phase 3 (target from 2027+): the Tianji Duxing Level 4 single-vehicle autonomous freight network, currently in development. The company also operates Tianji Tianshu, a fleet management SaaS. As of December 31, 2025, DeepWay had delivered 11,531 vehicles cumulatively (509 in 2023, 3,002 in 2024, 8,020 in 2025), ranking #9 in China's NEV heavy-duty truck market for 2025. Revenue was RMB 3,961.1M (~US$582M) in 2025, representing 101% year-on-year growth. Net loss narrowed to RMB 649.1M in 2025 (from RMB 675.1M in 2024), while operating cash flow turned positive at RMB 835.2M. Gross margin improved to 4.9% (from 0.5% in 2024). [CO001, CO002, CO006, CO007, CO008, CO009]

1.2 Leadership, Governance, and Shareholders

DeepWay was co-founded by Wan Jun, who serves as Chairman and CEO. Wan Jun also founded Lionbridge Financial Leasing Group, the trucking logistics firm that co-established DeepWay with Baidu. The CTO is Tian Shan, who previously led autonomous commercial vehicle research within Baidu's intelligent driving unit and brings deep technical continuity between Baidu's Apollo platform and DeepWay's Tianji intelligence suite. As of December 31, 2025, the company employed 1,308 people across all functions, with 412 (31.5%) in R&D. The shareholder structure reflects the joint-venture origin. Wan Jun's group (including Lionbridge entities) collectively holds approximately 20.44% of the outstanding shares. Baidu retains a 13.48% stake as of the HKEX application proof. Financial investors from the Series A (led by Qiming Venture Partners), Series A+ (led by Shandong Weiqiao Pioneering Group and SoftBank China Venture Capital), and Series B (co-led by Zhongan Capital and Puhua Capital) rounds hold significant stakes. Pre-IPO round investors include PuHua Capital (led Round 1 at $163M), Stone UAE, and NGS Super Australia. HKEX filings and financial statements were audited by a Big Four auditor. The company had no publicly named CFO or independent board directors in materials reviewed as of June 2026. Key-person risk is elevated on both Wan Jun (external relationships, strategy) and Tian Shan (full-stack technology architecture). Baidu's white-box Apollo IP license creates a strategic dependency that could affect governance if Baidu's stake falls below a license threshold. [CO003, CO004, CO005, CO022, CO029, CO033]

1.3 Funding History, Valuation, and Capital Strategy

DeepWay has raised a total of approximately RMB 3.96B (over $580M USD equivalent) across five disclosed rounds: Series A (RMB 460M, Aug 2021), Series A+ (RMB 770M, Mar 2023), Series B (RMB 750M, Dec 2024), Pre-IPO Round 1 (~RMB 1.177B, Jan 2026), and Pre-IPO Round 2 (total pre-IPO exceeding $310M USD, closed Apr 2026). The Pre-IPO Round 1 was the largest single round for an autonomous trucking company globally in the five years to 2026, according to company press materials. The pre-IPO valuation is not explicitly stated in the application proof; no official post-money equity valuation has been confirmed in public sources reviewed. The company filed its first HKEX listing application on November 6, 2025 (which expired), and refiled on May 7, 2026 with CICC and CMB International as joint sponsors. The IPO proceeds are intended to fund vehicle R&D, Changxing Three-Electric Smart Factory completion, and overseas market expansion. Capital efficiency concern: cumulative net losses exceed RMB 1.7B against a gross margin of 4.9% in 2025. The R&D expenditure was RMB 386.6M in 2025 (9.8% of revenue), and the company would need material margin expansion to achieve profitability at scale. An adverse analysis notes that the $173M and $310M figures refer to Round 1 and cumulative pre-IPO totals respectively—a distinction frequently blurred in press coverage that may mislead investors about round sizing. [CO013, CO014, CO015, CO016, CO017, CO018]

1.4 Milestones, Scale, and Market Presence

DeepWay's milestone arc spans from a two-party joint-venture announcement in late 2020 through commercial truck deliveries (first to CATL for trunk transport in June 2022), overseas market entry in Thailand (September 2024), and a pre-IPO institutional round exceeding $310M in April 2026. The company accumulated over 200M autonomous-driving kilometers across its fleet and 400+ service centres nationwide. The 2025 operating year was a step-change: deliveries grew 167% to 8,020 units, operating cash flow turned positive, and the Pre-IPO rounds brought in global institutional capital. The Tianji Suixing L2 ADAS subscription crossed 7,500 active vehicles. Overseas commercial operations now span Thailand, Australia, Malaysia, Singapore, New Zealand, Kazakhstan, and the UAE, with the first overseas commercial operation commencing in Thailand in September 2024. Intellectual property: DeepWay held 195 patents, 265 trademarks, and 108 software copyrights as of December 31, 2025. Road-test licences for L4 autonomous operation were obtained in Beijing and Tianjin. The Tianji Yanxing IPTS platooning system entered commercial service in 2023. Key customer risk: the top-5 customer concentration ratio was 39.5% in 2025 (improved from 82.5% in 2023), and the single largest customer accounted for 12.7% of 2025 revenue, creating meaningful revenue concentration. [CO024, CO025, CO026, CO028, CO036, CO037]

1.5 Exhibits

2.1 Market Boundary and Status-Quo Alternatives

DeepWay should not be analyzed as a generic autonomous-vehicle company, because the broad AV category mixes passenger cars, robotaxis, consumer ADAS, and software stacks whose buyers, budgets, and rollout rules differ materially from heavy-duty freight. The most decision-useful boundary starts with China road freight because that is the spend pool into which fleets, shippers, and logistics operators actually buy tractors, operating efficiency, and freight services. Inside that market, DeepWay’s practical wedge is the new-energy heavy-duty truck transition: fleets already replacing diesel assets with battery-electric tractors can add smart-cab hardware, assisted-driving subscriptions, and eventually higher-autonomy freight services without changing the core procurement workflow. The excluded spend is just as important. Passenger AV, warehouse robots, pure routing software, rail or ocean freight, and generic EV charging infrastructure all touch logistics or autonomy narratives but do not directly express the buyer job DeepWay is solving. The status quo substitute set is therefore concrete: diesel heavy-duty trucks with human drivers, conventional electric trucks still operated manually, and labor-heavy relay networks on long-haul corridors. Industry reporting summarized by IDTechEx notes that sub-1,000 km routes traditionally require two drivers while longer relay structures can involve four to six drivers, making labor, safety, and utilization central to purchase logic rather than peripheral features. DeepWay’s own positioning and filing disclosures support this phased framing: sell the truck first, monetize assisted driving second, and only then try to capture a share of freight service economics once corridor permissions and operational confidence are sufficient.[CM001, CM002, CM003, CM015, CM018, CM023]

2.2 Market Sizing and Growth Trajectory

The clearest top-down market anchor is not autonomous software at all but China road freight transport. Mordor Intelligence estimates the China road freight transport market at $500.9 billion in 2026, growing to $668.55 billion by 2031 at a 5.95% CAGR, which establishes the enormous spend base from which any truck OEM, fleet software provider, or autonomy operator must win share. A second lens is China’s broad autonomous-vehicles market: GII Research, distributed through Yahoo Finance, projects $22.84 billion in 2025 growing to $218.95 billion by 2034 at a 28.55% CAGR. That figure confirms the speed of automation investment but materially overstates what is relevant to DeepWay because it includes passenger-car and non-freight categories. A more operationally relevant third lens is the new-energy heavy-duty truck market. Industry coverage tied to 36Kr reports 231,100 China new-energy heavy-duty truck sales in 2025, up 182% year over year, with 28.89% penetration and expectations for roughly 35% in 2026 and above 50% by 2030. That electrified installed base matters because it creates the hardware substrate on which autonomy features can attach. The fourth lens is freight-autonomy-specific and much more aggressive. EqualOcean, citing a Beijing think tank, frames a 2030 China heavy-duty fleet of 6.27 million vehicles and 853.9 billion yuan in autonomous truck revenue potential. ARK Invest offers an even broader global ceiling, projecting $320 billion of autonomous over-the-road truck delivery revenue by 2030. These are useful directional signals, but they are not apples-to-apples with road freight spend or vehicle-sales data. DeepWay’s actual SAM is narrower: repeated corridor freight, new-energy truck replacement budgets, assisted-driving attach on activated fleets, and eventually route-specific autonomous service revenue where permits and remote-operations models exist. The correct diligence posture is therefore to preserve multiple lenses simultaneously and resist collapsing them into one heroic TAM number.[CM004, CM006, CM007, CM008, CM009, CM010]

2.3 Buyer Segmentation and Adoption Economics

DeepWay’s buyer map is multi-stage because the product is really a stack: a smart new-energy tractor, a paid assisted-driving layer, and a future autonomous-freight service. In the first phase, the direct buyer is usually the fleet owner, leasing arm, or transport operator making vehicle capex decisions. The user is a combination of driver, dispatcher, and safety/operations manager; the payer is often the fleet P&L owner or CFO-equivalent evaluating total cost of ownership, financing, utilization, and residual value. In the second phase, the budget case shifts toward operating expense: the fleet has already bought the tractor and is now deciding whether the assisted-driving subscription, energy optimization, or remote-operations tooling pays back through labor savings, fuel savings, reduced accidents, or higher asset utilization. DeepWay’s disclosed 7,500+ activated L2 vehicles, 30%+ subscription rate, and 200M+ cumulative mileage suggest the company has moved beyond a pure concept stage and into an early subscription-conversion phase, though public materials do not disclose per-vehicle unit economics. Segment-wise, the most plausible early adopters are high-frequency line-haul carriers, contract logistics operators, industrial captive fleets, and OEM- or dealer-linked fleet ecosystems that can standardize routes and maintenance. The adoption path is also narrower than a simple “sell trucks to everyone” story. Fleets usually start with route-constrained electrification, install or activate assisted-driving functions on predictable trunk lines, and only then experiment with more autonomous freight operations when regulation, insurance, and organizational trust permit. That sequencing matters for market sizing because the ultimate freight-service upside may be large, but budget ownership and willingness to adopt are initially governed by familiar fleet capex and TCO logic rather than speculative software multiples.[CM018, CM019, CM024, CM025, CM026, CM031]

2.4 Growth Drivers and Adoption Constraints

The strongest growth driver is structural inefficiency in Chinese logistics. China Daily, citing the China Logistics Information Center, reported total logistics costs of 18.2 trillion yuan in 2023, or 14.4% of GDP; the Swedish Transport Analysis Agency’s study notes that this ratio remains materially above developed-market norms. That gap creates a direct macro incentive for any technology that can lower labor intensity, improve asset utilization, and reduce energy or accident costs. Electrification is a second major driver because the new-energy heavy-duty truck transition is already real rather than hypothetical: 2025 sales and penetration data show fleets are replacing tractors now, which gives autonomy vendors a hardware install base instead of asking buyers to fund a category from zero. Third, China’s policy environment is more commercially constructive than many Western markets: national intelligent connected vehicle planning, public-road testing precedents for heavy-duty trucks, and designated demonstrations shorten the path from pilot to paid operations. The constraints are equally material. Capital intensity remains high because DeepWay must fund vehicle programs, software, and eventually a heavier operations stack before the L4 revenue layer is proven. Regulatory permissions are still corridor-specific, not nationwide blanket approvals, so the real SAM is constrained by where fleets can actually deploy higher-autonomy freight. Trust and liability are non-trivial: safety-driver-light or driver-out operations ask fleets and shippers to accept a new risk profile in a low-margin industry. Finally, market-sizing evidence is contradictory by design. Broad AV reports, new-energy truck sales data, freight-spend studies, and autonomy-scenario forecasts answer different questions. That does not invalidate the opportunity; it means investors should underwrite DeepWay on staged commercialization milestones and route economics, not on the largest headline forecast available.[CM005, CM012, CM013, CM014, CM021, CM022]

2.5 Exhibits

3.1 Competitive Landscape and Market Structure

DeepWay does not compete in a single narrow peer set. The landscape spans seven practical alternatives for freight buyers and fleet operators: direct China highway-autonomy truck peers, global autonomous-trucking stack vendors, electrified freight orchestrators, legacy truck OEMs and fleets moving into new-energy vehicles, the human-driver status quo, internal build or mixed-stack deployments, and likely entrants from larger autonomy or OEM ecosystems. In China, the most immediate peer set is Inceptio and KargoBot, because both are actively commercializing heavy-truck autonomy on domestic corridors and selling a concrete operating improvement story to fleets. Pony.ai is adjacent but increasingly relevant because it has brought robotruck revenue, SANY truck manufacturing ties, and dual-listing capital into the freight segment. Plus.ai, Aurora, Kodiak, and Einride matter less as immediate domestic substitutes today and more as strategic benchmarks for what best-in-class OEM partnerships, driverless safety narratives, and freight-network orchestration can look like. The status quo remains formidable: assisted-driving trucks, conventional fleet operations, and human drivers are still the default procurement path, while large fleets can multi-home across OEMs, telematics, and autonomy modules instead of standardizing on a single full-stack vendor.[CP001, CP002, CP003, CP004, CP041]

3.2 Direct Competitor Profiles

DeepWay's own profile is unusual because it is not only an autonomy-software vendor. The company says it forward-engineers the complete electric heavy truck and embeds assisted driving from the chassis up, which supports low aerodynamic drag, low incremental L2 BOM cost, and a monetization path through both vehicle sales and subscriptions. Its disclosed scale by filing period, 7,500-plus L2-activated vehicles, 30%-plus subscription attach, 200 million-plus kilometers, and CNY 1.5 billion H1 2025 revenue, shows meaningful customer adoption but not yet profitability. Inceptio is the strongest like-for-like China benchmark: it combines 4,000-plus L2+/L3 trucks, more than 400 million kilometers of autonomous mileage, large OEM integrations, and a public commercialization narrative around safety and fuel savings. KargoBot is the most concentrated L4 challenge, with a narrower but more advanced corridor-led program that claims positive unit economics in Ordos. Pony.ai brings balance- sheet and capital-markets strength from its broader autonomy business, while Plus.ai, Aurora, Kodiak, and Einride show how global peers are using OEM and carrier partnerships to reach production corridors without building a complete truck platform themselves.[CP005, CP006, CP007, CP008, CP009, CP010]

3.3 Capability, Pricing, and Distribution Comparison

Capability comparison is not just a race to the highest autonomy level. DeepWay's advantage is integrated electric-truck design with low-cost L2 deployment and measurable subscription attach. Inceptio leads on disclosed mileage and OEM manufacturing leverage. KargoBot leads on public L4 corridor specialization and platooning economics. Pony.ai blends trucking with a much larger robotaxi platform, giving it cash, engineering brand, and cross-program learnings. Plus.ai and Aurora show the power of global OEM distribution, but their disclosed evidence in this pack is stronger on partner breadth than on China route density. Public pricing remains opaque across nearly all peers: rather than software list prices, most vendors sell some combination of truck ASP, subscription attach, OEM-integrated ADAS cost, fuel savings, labor reduction, or utilization uplift. That opacity raises diligence importance around realized pricing, discounting, and contract structure. Distribution power is correspondingly shaped by OEM partnerships and manufacturing position. DeepWay's vertical integration is a strength if a buyer wants a tightly optimized electric truck, but it can be a limitation for fleets that prefer modular sourcing and multi-homing across chassis, software, and operations providers.[CP012, CP014, CP017, CP018, CP020, CP021]

3.4 Moat Durability and Displacement Risk

DeepWay's moat is most credible where electric-truck design, autonomy economics, and service access are tightly coupled. The company can argue that its complete-truck architecture lowers sensor and compute integration friction, keeps incremental assisted-driving cost low, and creates a better operator experience than retrofit approaches. That is real differentiation, but the durability of the moat is contested. Inceptio's larger disclosed mileage base and OEM share may compound faster than DeepWay's integrated-hardware edge because data flywheels and production partnerships are harder to compress once they scale. KargoBot shows that a more corridor-specific L4 strategy can reach attractive economics without solving every freight use case first. Pony.ai shows that adjacent autonomy platforms with much deeper cash pools can extend into trucking. Plus.ai and Aurora indicate that OEM-centric, geography-spanning models may commoditize parts of the stack. TuSimple's collapse into CreateAI after regulatory and governance turmoil is the clearest adverse precedent: cross-border autonomy programs can lose strategic focus quickly, and freight buyers may discount ambitious roadmaps if trust, capital, or regulatory posture deteriorate.[CP019, CP024, CP026, CP030, CP031, CP032]

3.5 Exhibits

4.1 Revenue Scale, Mix, and Quality

DeepWay's top-line ramp is real and unusually fast for an autonomous-trucking company, but the composition of that revenue matters more than the absolute headline. The HKEX filing shows revenue began only in June 2023 with the first batch deliveries of Xingchen I, then climbed from RMB 425.6 million in 2023 to RMB 2.0 billion in 2024 and RMB 4.0 billion in 2025. That growth came almost entirely from complete-truck hardware sales: truck hardware represented 99.9% of 2023 revenue and still 99.2% in 2025, while software and other revenue reached only 0.8% of sales in 2025. Delivery volume expanded from 509 units in 2023 to 8,020 in 2025, but estimated blended revenue per delivery fell materially over the same period, implying that scale has been won through a lower ASP mix, a richer product ladder, pricing pressure, or all three. The filing supports strong public traction on units and revenue, including a 516-unit backlog at filing date and ninth-place rank in China's 2025 NEV heavy-duty truck market, but it does not disclose realized subscription ARPU, contract discounting, or a list-to-net bridge that would let an investor separate hardware ASP, software take-rate, and channel effects with confidence.[CI001, CI002, CI003, CI004, CI005, CI006]

4.2 Cost Structure, Unit Economics, and Concentration

The 2025 income statement shows improvement, but not yet evidence of durable high-quality unit economics. Gross profit rose to RMB 194.9 million and gross margin improved to 4.9%, versus only 0.4% and 0.5% in 2023 and 2024, yet the company still posted a RMB 649.1 million net loss and a -16.4% net loss margin. R&D remains a defining cost center: DeepWay spent RMB 386.6 million on R&D in 2025, equal to 9.8% of revenue, and the filing says all R&D is expensed as incurred with no capitalization. Sales and marketing also scaled with the go-to-market buildout, reaching RMB 223.3 million and 428 heads in 2025, while estimated S&M spend per delivered truck fell sharply as volume ramped. Revenue quality still carries concentration risk. In 2025, the top five customers generated 39.5% of revenue, the largest customer contributed 12.7%, the largest customer's subsidiary also supplied batteries, and the top five suppliers accounted for 79.4% of purchases. Distributor sales were only 2.5% of 2025 revenue, which suggests DeepWay still depends primarily on direct sales and its 110-plus service-center footprint rather than a broad external channel. The net result is that DeepWay looks more like an early-stage EV OEM with assisted-driving monetization optionality than a software-first autonomy company with already-proven recurring economics.[CI012, CI013, CI014, CI015, CI016, CI017]

4.3 Cash Flow, Balance Sheet, and Capital Dependency

The cash-flow statement improved sharply in 2025, but the balance sheet still requires careful normalization. Operating cash flow moved from RMB -714.6 million in 2023 and RMB -429.6 million in 2024 to RMB 835.2 million in 2025, while financing inflow reached RMB 1.63 billion and year-end cash rose to RMB 1.61 billion. At the same time, investing cash outflow expanded to RMB 1.16 billion, trade and other receivables climbed from RMB 321.2 million in 2023 to RMB 1.58 billion in 2025, and net liabilities worsened to RMB -1.90 billion because of pre-IPO financial liabilities measured at fair value through profit or loss. Those FVTPL liabilities alone were RMB 2.96 billion at the end of 2025. Management also states plainly that it expects to continue incurring net losses in 2026. Public capital support remains substantial: Baidu still held 13.48%, the initial pre-IPO close was US$173 million, and the expanded pre-IPO round took disclosed financing above US$310 million after earlier Series A+ and Series B raises. However, the draft application proof leaves the line-item use of IPO proceeds redacted, so public investors still cannot underwrite the exact split between operations, technology investment, overseas expansion, working capital, and other obligations.[CI026, CI027, CI028, CI029, CI030, CI031]

4.4 Peer Context and Financial Verdict

Peer comparison clarifies both DeepWay's strengths and its limitations. DeepWay's RMB 3.96 billion of 2025 revenue is much larger than the public 2025 revenue disclosed by Aurora, WeRide, or Pony.ai, but that difference should not be misread as proof of superior autonomy economics because DeepWay is booking complete truck hardware revenue at scale, whereas those peers still report more software, services, or earlier-stage commercial mix. DeepWay's 4.9% gross margin remains far below WeRide's 30.2% and does not resemble software-like economics yet. Public traction is corroborated on revenue, deliveries, backlog, and cumulative L2 mileage, but critical underwriting inputs remain private, including subscription ARPU, realized discounting, warranty cost, receivables aging, sales-cycle length, CAC, payback, and contract renewals. External commentary also remains skeptical: Electrek highlighted that DeepWay still was not profitable despite more than 12,000 cumulative delivered electric semis globally. The financial verdict is therefore mixed: DeepWay has achieved genuine manufacturing and commercial scale, but public evidence still supports an EV-hardware-plus-ADAS story with improving yet thin gross margins, meaningful capital intensity, and several unresolved data gaps before a long-term margin path can be underwritten.[CI029, CI035, CI041, CI046, CI047, CI048]

4.5 Exhibits

5.1 Product Portfolio and Tianji Software Suite

DeepWay's core commercial product is the Xingchen series of forward-engineered new-energy heavy-duty trucks. The first-generation Xingchen Gen-I entered pilot operations in 2021 following the company's July 2020 founding, establishing proof of concept for the blank-sheet AV-first chassis architecture. The second-generation Xingchen Gen-II reached commercial production in 2023 with substantially improved powertrain efficiency and onboard compute headroom, and by 2025 had achieved 8,020 unit deliveries — making DeepWay the ninth-largest NEV heavy truck seller in China. The Xingtu (StarWay) platform, positioned for broader market accessibility, extends the product portfolio into additional fleet segments. Overseas, the Xingchen Gen-II obtained NHVR certification in Australia and commenced commercial operations in Thailand in September 2024, with distribution agreements covering Malaysia, Singapore, New Zealand, Kazakhstan, and the UAE by 2025. The Tianji suite is DeepWay's proprietary software stack layered on top of the vehicle hardware platform. It comprises four commercial modules: Suixing (Level 2 active safety assistance), Yanxing (Intelligent Platoon Transportation System for 2–3 truck convoys), Duxing (Level 4 solo autonomous development programme), and Tianshu (fleet SaaS for real-time telemetry, dispatch, and route optimization). As of December 2025, more than 7,500 vehicles are running the Suixing L2 subscription, representing a 30%+ take rate across the active fleet. The Tianshu SaaS creates contractual touchpoints with fleet operators independent of hardware refresh cycles. Yanxing IPTS entered commercial service in 2023 leveraging 5G-V2X communications for coordinated platoon management. Duxing remains in supervised road-test mode with ICV licences in Beijing and Tianjin only; no commercial L4 revenue has been disclosed in the HKEX filing.[CE001, CE002, CE003, CE004, CE005, CE006]

5.2 Technology Architecture and Apollo Integration

DeepWay's most structurally differentiated technical asset is its forward-engineered Electrical/Electronic Architecture (EEA). Unlike legacy truck OEMs that retrofit AV sensors onto existing diesel platforms, DeepWay's Xingchen chassis was designed from the outset for autonomous operation, with centralized domain controllers, redundant power networks, and native sensor mounting points. This approach means the vehicle's CAN/Ethernet backbone was architected for low-latency AV communication rather than retro-fitted as an add-on. The standard sensor suite per vehicle comprises 10 cameras (covering surround and forward long-range fields of view), 5 millimetre-wave radars, and 3 infrared sensors — 18 sensor units in total — providing 360-degree environmental coverage. The onboard computing platform delivers 500+ TOPS, sufficient to run real-time perception, prediction, and planning simultaneously. DeepWay is the only commercial vehicle company to hold a white-box licence to Baidu's Apollo autonomous driving software stack, meaning access to the platform source code rather than a compiled API. Apollo provides the mature core perception, prediction, and planning pipelines developed over Baidu's decade-plus of AV R&D; DeepWay adds commercial-vehicle-specific layers including heavy-truck dynamics models, load/weather handling, and driver-assist calibration profiles on top. The Three-Electric system (battery pack, drive motor, and motor controller) was designed for long-haul heavy freight — the most energy-intensive segment of the commercial vehicle market. A dedicated Three-Electric manufacturing facility in Changxing, Zhejiang broke ground in February 2025. Until that facility reaches production, assemblies are contracted to JAC Motors and Shandong Reach, both established Chinese vehicle manufacturers. DeepWay has accumulated more than 200 million autonomous and semi-autonomous kilometres across its active fleet as of the May 2026 HKEX filing. The company holds ICV (Intelligent Connected Vehicle) road-test licences in Beijing and Tianjin municipalities enabling supervised L4 development on public roads.[CE009, CE010, CE011, CE012, CE013, CE014]

5.3 IP Portfolio, R&D Investment, and Technology Roadmap

As of December 31, 2025, DeepWay holds 195 patents, 265 trademarks, and 108 software copyrights, a portfolio assembled in under five years of operations. Patent coverage spans vehicle hardware (chassis design, Three-Electric components, sensor mounting systems), software (perception algorithms, EEA firmware), and system-level innovations (platoon coordination protocols, fleet optimization models). R&D investment has grown substantially in absolute terms — from RMB 219M in 2023 to RMB 387M in 2025 — while declining as a share of revenue as the top line scales (51.5% in 2023, 9.8% in 2025). The 412 R&D employees (out of 1,308 total) represent 31.5% of the workforce, a notably high ratio for a company delivering physical hardware at volume. The technology roadmap follows a three-phase commercialisation trajectory disclosed in the HKEX filing: Phase 1 (current) — vehicle sales with Tianji subscription attach; Phase 2 — expand the IPTS platoon network to create autonomous-freight corridors; Phase 3 — transition to a full L4 autonomous freight network where DeepWay operates trucks as a service. The credibility of Phase 3 is constrained by the current development-only status of Duxing: no commercial L4 deployment has been announced and the HKEX filing discloses no target timeline for commercial L4 operations. The open-source nature of the Apollo platform creates both an advantage (proven codebase, active ecosystem) and a dependency risk: any change in Baidu's licensing terms for the white-box Apollo access would require DeepWay to replace the core AV software stack, an undertaking that would consume significant R&D and potentially delay the commercial roadmap. Competitive erosion in the L2 ADAS segment from domestic peers (notably HAOMO, Hesai, and platform integrators working with SAIC and FAW) represents the nearer-term risk to subscription revenue growth.[CE021, CE022, CE023, CE024, CE031, CE032]

5.4 Exhibits

6.1 Named Customer Deployments — China

DeepWay's strongest commercial proof resides in China's short-haul port-logistics and express-delivery sectors. The most significant 2026 deployment is a 200-truck order from Platypus Logistics, a world-leading port-logistics enterprise with operations across Shanghai, Ningbo, Guangzhou, and Shenzhen. On May 7, 2026 the first 50 DeepWay Xingtu 6×4 electric trucks (505 kWh battery) were formally handed over to Sixiangjia, a company within the Platypus ecosystem, with the remaining 150 units scheduled for H1 2026 delivery. Platypus selected DeepWay on the basis of four operational factors: ultra-low energy consumption (industry-leading real-world efficiency), safety from ground-up engineering rather than diesel-to-electric conversion, native TMS digital integration with Platypus's capacity-dispatching system, and high uptime reliability in intensive port-cycle operations. Daily operating mileage on Platypus routes ranges from 200 to 400 km. The Platypus selection is commercially significant because the company uses data-driven vehicle procurement; adoption by a data-driven operator constitutes independent validation of total cost of ownership, not merely a marketing relationship. In October 2024, DeepWay and CATL jointly delivered 100 Shenxiang Xingchen battery-swappable heavy-duty trucks to Xinchenghui Logistics Co., Ltd. in conjunction with the inauguration of a new CATL Electric Truck Charging and Battery Swapping Center in Dujiangyan, Sichuan Province. These trucks target Southwest China's gravel-transport sector, achieving 190 km range on a full 49-tonne load per charge and 400–500 km per day through battery-swap cycles. Energy consumption of 0.925 kWh/km at full load was cited as a key procurement driver. On April 21, 2025, DeepWay and STO Express held a delivery ceremony in Shanghai for the first batch of new-energy heavy-duty trucks — the first such delivery in China's national express-delivery industry. These vehicles are equipped with CDTL's distributed drive system (EA5000N), achieving 1.15 kWh/km ultra-low energy consumption for intercity transport. The milestone marked a critical precedent for zero-emission adoption in China's parcel-logistics sector. SF Express is separately confirmed as a DeepWay customer per industry analyst databases, though the specific fleet size and contract terms are not publicly disclosed.[CU001, CU002, CU003, CU004, CU005, CU006]

6.2 International Market Customers and Distribution

DeepWay entered international markets in earnest in 2025 and accelerated customer acquisition through 2026. The most commercially advanced international deployment is Foodstuffs North Island in New Zealand, which became the first operator of a DeepWay Star electric truck in a temperature-controlled commercial supply chain outside China. Foodstuffs commissioned the vehicle from its Palmerston North Distribution Centre in March 2026. The truck performs a daily dual-cycle route: a 200 km round trip delivering chilled and frozen groceries to Kapiti Coast stores in the morning, followed by a 280 km round trip to other lower North Island stores in the afternoon. A 90-minute midday recharge sustains the 460 km daily cycle. The truck has a 600 kWh lithium iron phosphate battery, 224 kW continuous / 492 kW peak power, 30,000 Nm continuous torque, and a 50,000 kg Gross Combination Mass. Foodstuffs previously trialled the vehicle in 2025 over 1,000 km with the refrigerated trailer. Chief Executive Chris Quin linked the adoption to supply chain resilience against fuel price volatility as well as environmental targets (42% emissions reduction by 2030). In the New Zealand distribution channel, Etrucks New Zealand was formally appointed as official national distributor in July 2025 and launched the DeepWay Star at the EROAD Fleet Managers' Day in Hamilton on July 30, 2025. The vehicle's retail price of NZD 255,000 + GST (after the government's Low Emissions Heavy Vehicle Fund subsidy) represents the most affordable battery-electric semi-tractor available in New Zealand. The distribution agreement was signed between Etrucks director Ross Linton and DeepWay global operations manager Chiara Wang at DeepWay's Hefei headquarters. Etrucks previously distributed the high-end Windrose E1400; the DeepWay Star fills a cost-effective tier, broadening addressable fleet segments. In Australia, DeepWay signed a dealership agreement with an Australian partner in May 2025 and confirmed its first order — making Australia DeepWay's fourth Asia-Pacific market after Singapore, Thailand, and Malaysia. The Australian distributor conducted real-world testing before signing and validated the trucks for high efficiency and safety performance. DeepWay International General Manager Lin Facai committed to customized products and full lifecycle support for the Australian market. The NGS Super pension fund's participation in the April 2026 pre-IPO round reflects both capital commitment and an intention to leverage its extensive Australia–New Zealand resources to accelerate DeepWay's regional commercialization.[CU010, CU011, CU012, CU013, CU014, CU015]

6.3 Customer Segmentation, Adoption Trajectory, and Recurring Revenue

DeepWay's primary customer segments are large commercial logistics operators in China's short-haul, fixed-route bulk-freight market — a segment estimated to account for 40–50% of China's total road freight volume and the most viable deployment scenario for near-term autonomous operation at scale. The Chinese customer base spans express delivery (STO Express, SF Express), port logistics (Platypus, Xinchenghui), gravel/bulk cargo, and regional trunking. DeepWay holds over 95% market share in China's express-delivery electric-truck segment and ranks among the top sellers in Shanghai, Xinjiang, and Shandong. Adoption has followed an accelerating trajectory: 509 units delivered in 2023 (first commercial year), 3,002 in 2024, and 8,020 in 2025 — the latter ranking ninth in China's new-energy heavy-truck market. As of early 2026, cumulative deliveries reportedly exceeded 12,000 units across China, Thailand, New Zealand, Australia, and other markets, with cumulative customer-operated distance exceeding 200 million km by mid-2025 (as reported by the New Zealand distributor). DeepWay's revenue grew 101% in 2025 to approximately RMB 3.96 billion (~$582M), driven almost entirely (>90%) by vehicle sales. A key dimension of the customer model is the recurring revenue from Tianji Suixing ADAS subscriptions. DeepWay is the first company globally to make an integrated driver-assistance system standard across its entire vehicle lineup; fleet operators pay subscription fees for active use of the autonomous-driving assistance features. As of mid-2025, over 2,000 vehicles had the subscription activated with a rising paid subscription rate described by investors as a strong commercial signal. DeepWay's 400+ service centers and 1,000+ certified engineers in China underpin the after-sales and subscription service network required to sustain repeat business. However, publicly available evidence on customer retention is limited. No NRR, GRR, or formal churn data has been disclosed. Long-term fleet-level subscription renewal rates and any evidence of contract length or lock-in structure are not publicly available. These represent material diligence gaps for evaluating the quality and durability of the customer base.[CU019, CU020, CU021, CU022, CU023, CU024]

6.4 Retention, Expansion, and Concentration Risks

DeepWay's customer expansion strategy relies on the "data flywheel" — increasing autonomous-miles across the installed base feeds algorithm refinement, enabling safer and eventually driverless operation that justifies higher per-truck or per-km subscription pricing over time. This logic underpins the long-term case for land-and-expand within logistics operators, but the commercial evidence for actual upsell from L2 ADAS to future L4 autonomous-freight-robot services remains entirely theoretical as of mid-2026. Concentration risk is material. The customer base is dominated by Chinese logistics operators in a single vertical (bulk/port freight). The top two fleet customers (Platypus at 200 trucks and Xinchenghui at 100 trucks) together represent approximately 300 confirmed production deployments in China — a large share of the Chinese named-customer evidence. International customers are in early-commercial or pilot stage; Foodstuffs NZ represents one truck in a real-world deployment, and the Australian order has not been publicly quantified. No customer has publicly disclosed fleet-expansion orders beyond the initial contracts. From the adverse perspective, robotics.press notes in a competitive-intelligence report that DeepWay's self-reported operational metrics are self-certified rather than audited, and that the $310M pre-IPO round figure conflicts with earlier tracked data (RMB 1.177B ≈ $173M first tranche vs. a cumulative figure that some databases tracked at $282–354M). The report characterises the gap between 6,400 L2-ADAS trucks and commercially viable L4 deployment as involving unresolved regulatory, insurance, and technical-validation challenges. No independent customer outcome study (G2, Gartner Peer Insights, academic third-party audit) corroborating DeepWay's TCO claims has been identified in public sources as of the research date.[CU028, CU029, CU030, CU031, CU032, CU033]

6.5 Exhibits

7.1 Regulatory, Legal, and National-Security Risks

DeepWay's core commercial thesis depends on China granting broad commercial L4 truck deployment authorization. As of June 2026, that authorization is restricted to five pilot cities — Shanghai, Guangzhou, Shenzhen, Chongqing, and Hangzhou — where safety-driver removal was conditionally permitted from August 2025. Outside these corridors, commercial L4 freight operation is not permitted. A mandatory national L3/L4 standard is scheduled to take effect in July 2027, which will impose harmonized safety testing, certification, and deployment approval requirements across China. Non-compliance with these standards would prevent commercial operation in any geographic market. The transition creates both a timing risk (if certification takes longer than projected) and a cost risk (from required system upgrades to meet national standards). China's data privacy and cybersecurity laws — the Personal Information Protection Law (PIPL), the Data Security Law (DSL), and the Cybersecurity Law — impose strict requirements on cross-border data transfers, with autonomous-vehicle sensor and telematics data classified as data of potential national importance. These laws require that certain data be stored domestically, subject to security assessments before export, and processed only by licensed entities. For overseas deployments, host-country regulators impose their own approval requirements. The UN Global Technical Regulation for automated driving systems, adopted in June 2026, introduces a new layer of international compliance standards that DeepWay must meet for fleet deployments in regulated markets. International deployments face a compounding risk: Chinese autonomous-driving technology embedded in commercial trucks operating on Western transport corridors has attracted geopolitical scrutiny. The Carnegie Endowment noted in November 2025 that Chinese autonomous-vehicle platforms operating in partner-country supply chains raise data-sovereignty and infrastructure-vulnerability concerns for host governments. The US and EU have begun to extend screening mechanisms to Chinese technology companies operating physical transportation infrastructure. Even if DeepWay obtains host-country operational approval, post-approval restrictions or mandatory audits may materially constrain the overseas revenue model. No L4 commercial-trucking insurance framework has been standardized in China, creating an unresolved risk-transfer mechanism for autonomous fleet operators.[CR001, CR002, CR003, CR004, CR005, CR006]

7.2 Technology, Operational, and Supply-Chain Risks

DeepWay's current product range is commercially deployed as L2 Advanced Driver Assistance (ADAS) — not Level 4 autonomous — across its ~12,000-unit installed fleet. The commercial transition from L2 fleet-data collection to Level 4 autonomous freight operation involves engineering milestones (full-stack L4 validation, sensor redundancy certification, fail-safe architecture), regulatory milestones (commercial deployment permits in each jurisdiction), and insurance milestones (insurance coverage frameworks) that are all unresolved as of mid-2026. robotics.press has noted that DeepWay characterizes its L2 ADAS as an on-ramp to L4, but all autonomous-miles metrics are self-reported and have not been independently audited ahead of the HKEX IPO filing. The validator dimension is critical: investors relying on the data flywheel narrative (accumulating training miles to accelerate L4 safety) cannot independently verify the claim. JAC Motors is the primary OEM manufacturing partner for DeepWay's Xingtu and Star platforms. The manufacturing dependency concentrates production risk in a single partner's capacity, quality systems, and logistical availability. Any disruption to JAC's production capacity would directly impair DeepWay's ability to fulfill the 1,400-unit backlog reported as of mid-2025. CATL's 600 kWh LFP battery pack is the only battery platform across the full product line; no secondary supplier has been publicly disclosed. Battery pricing fluctuations — which are driven by lithium carbonate and LFP cathode material spot prices — compress gross margins directly. A rapid decline in CATL battery pricing (as occurred in 2023–2024 across the Chinese EV industry) could trigger fleet customers to defer purchases awaiting cheaper next-generation packs, creating demand pockets and OEM inventory risk for DeepWay. Platooning and L4 pilot operations in Inner Mongolia and Xinjiang create safety and liability exposure in the absence of a recognized insurance or incident-response framework.[CR007, CR013, CR014, CR015, CR017, CR018]

7.3 Competitive and Financial Risks

DeepWay faces intensifying competitive pressure from multiple vectors. Inceptio Technology has accumulated over 400 million cumulative autonomous km across its fleet, which represents a 4× data-scale advantage over DeepWay's approximately 100 million reported km. Inceptio's software-only OEM-integration model allows it to embed into multiple truck brands without the capital requirements of a vertically-integrated OEM, giving it a structurally lower cost to scale. KargoBot, spun out of Didi in 2023 and having raised RMB 600 million (~$83M) in a 2024 Series C, has demonstrated commercial L4 supply-chain operations and operates a purpose-built cabless autonomous truck — competing on the same port-logistics corridor as DeepWay's Platypus deployment. Zeron, backed by CATL, Nio Capital, and Momenta, raised $175M in 2024 and represents a future competitive threat with CATL supply-chain access that may match DeepWay's battery partnership advantage. On the financial side, DeepWay reported a net loss of RMB 649 million in 2025 despite revenue of RMB 3.96 billion (+101% YoY), implying an EBIT margin of approximately -16%. Hardware (truck sales) represents over 90% of revenue, meaning the business is effectively a capital-intensive manufacturing operation, not a software business. R&D investment required to bridge the L2-to-L4 gap must be sustained alongside production scale-up, creating dual capital demands with limited recurring-software revenue to offset. The HKEX Chapter 18C IPO is targeted for a market environment where loss-making autonomous-vehicle companies have faced heightened investor scrutiny globally. Baidu holds approximately 13–17% of DeepWay (varying by source), acting both as a strategic anchor and a potential overhang if Baidu reduces its autonomous-vehicle strategic priority. A discrepancy between a $310M pre-IPO funding announcement and robotics.press's tracked first tranche of ~$173M raises data-quality questions about disclosed metrics.[CR016, CR023, CR024, CR025, CR026, CR027]

7.4 Partner Dependencies, People Risks, and Mitigations

DeepWay's business model rests on three concentrated partner dependencies: CATL for batteries, Baidu Apollo for the ADS software stack, and JAC for vehicle manufacturing. The CATL battery relationship is the most structurally embedded: CATL co-delivered the Xinchenghui battery-swap fleet in October 2024 and has built charging-and-swap infrastructure alongside DeepWay deployments. This creates mutual switching costs but also consolidates critical components under a single supplier whose strategic priorities (including CATL's own EV truck brand Zeron) may not permanently align with DeepWay's. Baidu Apollo provides the autonomous-driving perception and planning stack; the exclusive licensing relationship underpins DeepWay's L4 development roadmap. Any strategic realignment by Baidu — including exit, reduced investment, or competitive repositioning in the autonomous freight space — would simultaneously remove DeepWay's primary technical partner and a key marketing credibility anchor. On people risk, CEO Wan Jun is the co-founder of Lionbridge (the logistics-leasing parent) and is the primary commercial relationship architect; CTO Tian Shan, a 13-year Baidu veteran, is the technical anchor of the ADS development team. Departure of either executive would create a material continuity gap. Mitigations that partially address the risk profile include: growing the autonomous-km data flywheel, diversifying the international customer base to reduce China-only concentration, expanding the service-center network (400+ centers already operational), and pursuing the HKEX IPO to establish a listed entity with improved public-market accountability and capital access. The unit-economics breakthrough observed in Ordos L4 pilots in H1 2025 (reported as positive at that scale) represents a potential inflection indicator if replicated at national scale. Kill criteria for the investment thesis include: failure to obtain national L4 commercial deployment authorization post July 2027 standard, departure of the Baidu Apollo partnership, or inability to sustain the IPO financing runway through the L4 ramp.[CR034, CR035, CR036, CR037, CR038, CR039]

7.5 Exhibits

8.1 Comparable Company Universe and Revenue Quality Analysis

DeepWay enters valuation with an unusual public profile for an autonomous-trucking company. Its FY2025 revenue of RMB 3961.1 million, roughly $582 million at the user's reference exchange rate, is far larger than the revenue currently disclosed by Aurora, WeRide, or Pony.ai. That scale matters because it proves real truck commercialization and not just pilot deployments. But the quality of the revenue matters just as much as the quantity. The prospectus shows that 99.2% of FY2025 revenue still came from truck hardware and only 0.8% from software and other streams, leaving DeepWay much closer to an EV truck OEM with autonomy upside than to a software-first autonomy platform. That mix explains why public comparables split into two camps: DeepWay deserves a premium to plain truck manufacturers for growth and software optionality, but a discount to pure-AV or ADAS names whose gross margins and recurring revenue are already much stronger.[CV001, CV002, CV014, CV016, CV017, CV018]

8.2 Market Context and Growth Addressability

The market backdrop is large enough to justify a meaningful valuation range even before DeepWay proves software monetization. Mordor estimates China's road-freight market at $472.77 billion in 2025 and $668.55 billion by 2031, while ARK frames autonomous transportation as a multi-trillion-dollar global opportunity by 2030. Those figures do not mean DeepWay deserves software multiples today, but they do explain why investors continue to fund the company aggressively. DeepWay is still early in share terms: the filing places it ninth in China's 2025 NEV heavy-duty truck market, leaving a long runway before any claim of category leadership. The competitive context is also important. China is emerging as a major commercialization venue for assisted-driving trucks, and rivals such as Inceptio and KargoBot are each building their own valuation stories around commercial mileage, corridor economics, and OEM reach. DeepWay's addressable upside therefore depends on converting its current truck scale into durable market share and software attach rather than on TAM alone.[CV029, CV030, CV031, CV032, CV049, CV055]

8.3 Valuation Methodology and Multiple Analysis

Revenue multiples are the most practical framework because DeepWay remains loss-making and its exact IPO share count and price are still redacted in the draft prospectus. A DCF would create false precision, while EBITDA or earnings multiples are premature. The cleanest method is therefore a blended EV-to-revenue range anchored on three layers. First, a hardware floor of roughly 1.5x to 3x sales captures the reality that most revenue is still generated by truck hardware with only 4.9% gross margin. Second, a mixed-model middle band of roughly 5x to 8x sales reflects DeepWay's very fast growth, positive operating cash flow, and credible but still unproven software upside. Third, software-style upside of 10x to 15x sales is only rational if Tianji subscriptions begin to matter and gross margins step materially higher. The latest private financing round supports the idea that investors see upside optionality, but public evidence still leaves exact pre-IPO valuation terms unresolved.[CV022, CV023, CV024, CV025, CV026, CV027]

8.4 Scenario Analysis and IPO Pricing Context

Scenario analysis is more informative than a single target value because DeepWay's next twelve to twenty-four months could push the company toward very different peer sets. In the bull case, software revenue becomes visible, gross margins expand above 15%, and corridor or platooning economics create a service-like revenue stream that justifies a $6 billion to $8 billion range. In the base case, DeepWay keeps growing rapidly, but remains mainly a truck-plus-ADAS business, supporting a $3 billion to $4.5 billion band similar to a discounted version of Pony.ai or WeRide valuation logic. In the bear case, hardware margins stay thin, established OEMs pressure pricing, and public-market skepticism toward China-origin autonomy names intensifies, pulling the company toward a $0.8 billion to $1.8 billion range. TuSimple is the essential adverse precedent because it demonstrates how quickly an autonomous-trucking equity story can implode after listing if governance, regulation, or commercialization disappoints.[CV034, CV035, CV036, CV037, CV041, CV042]

8.5 Key Catalysts, Risks, and Investor Considerations

The final investment question is not whether DeepWay has built a real business; it has. The question is whether the IPO price converts that proof into an attractive return. Near-term rerating upside would come from disciplined IPO pricing, evidence that gross margin keeps climbing in H1 2026, regulatory permission for more advanced platooning or L4 monetization, and proof that overseas expansion adds demand rather than distraction. The main reasons for caution are equally clear: negative equity remains a real overhang, software revenue is still negligible, and both Chinese autonomy specialists and established truck OEMs can attack DeepWay's narrative from different angles. As a result, the public evidence today supports a price-sensitive stance rather than an unconditional buy call. Valuation looks fair in a mid-single-digit-sales framework, stretched if IPO pricing pushes into software-style territory, and attractive only if the final deal leaves room for execution risk that remains unresolved in public data.[CV010, CV011, CV012, CV015, CV038, CV047]

8.6 Exhibits