DeepWay

1.1 Identity and business model

DeepWay was established in 2020 as a joint venture between Baidu and Lionbridge and now identifies Hefei as its headquarters. The company’s official English pages frame its purpose as electrifying heavy-duty transport, while third-party profiles describe a business centered on designing and selling forward-defined intelligent electric trucks to logistics and supply-chain operators. Public pricing is reported at RMB 470,000–700,000 including VAT. This remains predominantly a vehicle-hardware business: BitAuto reports RMB 3.961 billion of 2025 operating revenue and says new-energy heavy-truck sales consistently supplied more than 99% of revenue. Accordingly, software and autonomous-driving optionality should not be treated as a proven recurring-revenue base.[CO001, CO002, CO005, CO006, CO007]

1.2 Leadership and governance

Founder, chairman, and CEO Wan Jun brings direct logistics-market experience through Lionbridge; public profiles also describe him as DeepWay’s actual controller with a 26.22% direct and indirect holding. The named operating bench covers autonomy, vehicle engineering, commercialization, and finance: Tian Shan is identified as CTO, Tan Changyu as chief engineer, Xu Shangshang as CMO, and Wang Jia as CFO. That mix supports founder-market fit and functional coverage, but public evidence in the assigned pool does not disclose board composition, committee structure, protective provisions, or succession planning. The concentration of founder control and the Baidu technology relationship therefore remain material governance diligence items.[CO003, CO004]

1.3 Capital, scale, and financial risk

DeepWay’s scale accelerated materially: BitAuto reports 8,020 truck deliveries in 2025, up 167% year over year, with a 3.4% share and ninth place in China’s new-energy heavy-truck market. The capital record is less clean. A November 2025 report cited RMB 1.96 billion across 11 rounds, while later sources describe two 2026 pre-IPO financings totaling more than US$310 million; overlap and scope are not reconciled. Valuation indicators also differ by date and method. Financial risk remains substantial: one report calculates RMB 1.70 billion of cumulative losses from 2022 through first-half 2025, and at June 30, 2025 reports liabilities of RMB 5.54 billion against RMB 3.92 billion of assets and RMB 593 million of cash.[CO008, CO009, CO010, CO011, CO012, CO013, CO014]

1.4 Milestones and diligence posture

The operating chronology is coherent even where point estimates are not. DeepWay moved from formation in 2020 to a first truck unveiling in 2021, successive institutional rounds in 2022–2024, first batch delivery in May 2023, and a Hefei headquarters move in November 2023. By 2025 it reported 8,020 annual deliveries, and 2026 pre-IPO financing confirms late-stage capital-market preparation. Two implications follow. First, the company has crossed from prototype to material vehicle shipments. Second, the evidence base is still dominated by company-linked releases, commercial profiles, and secondary reporting rather than a current prospectus or audited cap table, so exact ownership, valuation, cash availability, customer count, and headcount require primary-document verification.[CO001, CO002, CO008, CO011, CO012, CO015, CO016]

1.5 Exhibits

2.1 Market boundary and status quo

DeepWay's core market is purpose-built battery-electric heavy-duty trucks for highway freight fleets, with intelligent-driving capability embedded in the vehicle. That boundary excludes passenger autonomy, light commercial vehicles, stand-alone robotaxi software, and charging infrastructure not bundled with a truck sale. It also prevents the broader autonomous-driving opportunity from being treated as DeepWay's current addressable revenue pool: more than 99% of reported revenue has come from truck sales. The status quo is the diesel heavy truck, and the buyer's economic comparison is anchored in fuel and driver costs rather than software features alone. Public pricing of CNY 470,000-700,000 including VAT confirms an enterprise capital-purchase motion aimed at logistics and bulk-transport operators.[CM001, CM002, CM003, CM004]

2.2 Multiple sizing lenses

The strongest available top-down lens is China's new-energy heavy-truck market, not global autonomous freight. One secondary source reports 231,100 units sold in 2025, up 182%, with 28.89% penetration. Two published 2030 outlooks are directionally consistent but not identical: one points to more than 50% penetration and a market above CNY 250 billion, while a CIC forecast cited by another publisher gives 53.5% penetration and CNY 255.8 billion at a 33.3% CAGR. These are forecasts relayed through secondary outlets, so they define a range rather than an audited TAM. DeepWay's 8,020 deliveries and approximately 3.4% share in 2025 provide a bottom-up company anchor, but do not establish a durable SOM.[CM005, CM006, CM007, CM008, CM009]

2.3 Buyer, user, payer, and adoption path

The economic buyer is generally a fleet owner, logistics enterprise, or industrial transport contractor; drivers and dispatch teams are the users, while procurement or operations leaders control the vehicle and energy budget. Evidence spans express and third-party logistics customers as well as an industrial deployment: DeepWay delivered 100 trucks to Yuanchi New Energy for logistics serving Baowu Steel's Taiyuan operation. Adoption is therefore route-led. Fleets first validate payload, range, charging or swapping access, uptime, safety, and total cost on repeatable lanes, then expand vehicle counts. Reported paid take-up above 30% for intelligent driving functions suggests a software upsell path, but the evidence does not disclose price, cohort retention, or revenue contribution.[CM010, CM011, CM012, CM013]

2.4 Drivers, constraints, and valuation relevance

Electrification penetration, fleet pressure to reduce energy expense, and the possibility of lowering driver-related costs support adoption. Against those drivers, a 49-ton truck configuration with a reported 300 km single-charge range makes route design and battery-swap access material gating factors rather than implementation details. Company-level economics are another constraint: 2025 revenue reached CNY 3.96 billion, but net loss was CNY 649 million. Software monetization is also unresolved. One source says smart-logistics solutions were still in free trial through the first half of 2025, while later reporting says recurring subscription revenue had begun. Without software ARPU, renewal cohorts, contribution margin, or a reconciled launch date, valuation should be tied primarily to truck-market penetration and hardware economics, with software treated as upside rather than a proven market layer.[CM004, CM013, CM014, CM015, CM016]

2.5 Exhibits

3.1 Competitive landscape and routes to market

DeepWay sits between electric-truck manufacturers and autonomous-freight technology vendors: it sells complete trucks today while offering three possible European routes—finished-vehicle export, production with a local manufacturer, or technology licensing to an established brand. That flexibility is strategically useful, but it is not yet equivalent to local presence. Management said DeepWay had no European base and still needed after-sales, charging, financing, and leasing partners. At the same 2026 industry event, BYD described a European heavy-truck launch followed by local manufacturing, Sinotruk and SuperPanther highlighted Austrian assembly with Steyr Automotive, and Sany emphasized cooperation with Putzmeister. DeepWay therefore faces direct products, incumbent industrial networks, and a licensing alternative that could reduce capital intensity but also place customer ownership with partners. Reported sales of more than 13,000 trucks by April 2026 establish commercial scale, although Dealroom's broad valuation range is not a peer benchmark.[CP004, CP005, CP006, CP007, CP011]

3.2 Capability, product economics, and operating model

DeepWay's differentiation rests on forward-designed electric architecture, assisted-driving integration, and greater control of batteries, e-axles, and power electronics. Public specifications indicate meaningful product evolution: the 2021 Xingtu release described 300 kilometers at a full 49-ton load, while a 2026 report described roughly 400 kilometers at the same rated gross weight for the 600 kWh Star. Those figures are different generations and configurations, not a controlled competitor benchmark. A Gasgoo report also relayed a company demonstration claiming 10% lower energy use and 22% better charging efficiency than a converted electric truck; the absence of an independent test limits underwriting value. Pricing is clearer only for DeepWay: two sources place recommended vehicle prices at RMB 470,000–700,000 including VAT. Peer transaction prices, discounts, battery ownership, financing, maintenance, and realized total cost of ownership were not disclosed, so the matrix marks unsupported comparisons as unknown rather than assigning winners.[CP008, CP013, CP014, CP015, CP016]

3.3 Moat durability and displacement risk

The core competitive tension is that DeepWay is integrating vertically from a weaker purchasing position. Its international executive acknowledged that larger Chinese rivals use scale to price lower and that DeepWay has less negotiating power with suppliers, including battery vendors. The company does not own Chinese manufacturing plants and relies on contract manufacturers, while three-electric systems were reported to represent about 60% of manufacturing cost. Bringing those systems in-house may improve control, but it also concentrates execution risk in a low-margin hardware model. Software monetization is not yet proven: a 2025 report said smart-logistics solutions remained in free trial and generated no income. Even basic scale evidence requires reconciliation, with one source reporting about 6,400 vehicles in 2024 while another reports 3,002 deliveries. The defensible near-term moat is therefore integration and accumulated operating data, not demonstrated autonomous-driving superiority, exclusive distribution, or independently validated cost leadership.[CP001, CP002, CP003, CP009, CP010, CP012, CP013]

3.4 Exhibits

4.1 Revenue architecture is concentrated in truck hardware

DeepWay's disclosed revenue base is overwhelmingly transactional rather than recurring. New-energy heavy-truck sales represented more than 99% of revenue in the reported periods, parts contributed less than 1%, and smart logistics solutions were still offered on a free-trial basis without income. Public descriptions also mention vehicle-management, financing-support, and used-vehicle services, but no source allocates revenue to those activities. Reported retail prices of RMB 470,000-RMB 700,000 including VAT establish a list-price envelope, not realized average selling price, net revenue, or gross profit. The revenue quality therefore remains hardware-led and dependent on deliveries, with software and services representing unproven monetization options rather than current recurring revenue.[CI003, CI004, CI005, CI016]

4.2 Pricing and go-to-market evidence is public only at the edges

The available evidence points to direct fleet sales and partnership-led expansion. DeepWay's first delivered vehicles went to XuRi DongSheng for CATL logistics work, showing an enterprise deployment route, while its European plan contemplates licensing technology to established local brands and eventually manufacturing locally. Neither route discloses sales-cycle length, channel margin, customer acquisition cost, payment terms, discounts, or renewal economics. An aggregated source reports an ADAS subscription rate above 30% alongside more than 6,000 L2-equipped vehicle deliveries, but it does not define the subscription-rate denominator or disclose price and recurring revenue. Those limitations make the evidence useful for mapping monetization mechanisms, not for underwriting sales efficiency.[CI013, CI014, CI015]

4.3 Scale is rising faster than demonstrated operating leverage

Reported revenue rose from RMB 426 million in 2023 to RMB 1.97 billion in 2024, and first-half 2025 revenue reached RMB 1.51 billion, up 97.6% year over year. Deliveries also progressed from 509 units in 2023 to 3,002 in 2024 and 8,020 in 2025. That scale has not yet produced profitability: one independent report says DeepWay remained unprofitable after more than 12,000 cumulative deliveries, while another reports RMB 1.70 billion of cumulative accounting losses from 2022 through first-half 2025. R&D remained substantial, and management's planned vertical integration targets three-electric systems that reportedly represent about 60% of manufacturing cost. Public data still omit gross margin, warranty burden, battery procurement economics, and factory capex, so the margin path cannot be quantified.[CI001, CI002, CI006, CI007, CI008, CI011, CI012]

4.4 Capital adequacy cannot be reduced to a public runway estimate

At June 30, 2025, reported liabilities of RMB 5.54 billion exceeded assets of RMB 3.92 billion, while cash and cash equivalents were RMB 593 million. Those figures show balance-sheet strain, but accounting losses are not cash burn and funding raised is not cash available. The public record does not disclose operating cash flow, committed factory capex, working-capital seasonality, debt maturities, project-finance obligations, or restricted cash, so a precise runway would be invented. The financial verdict is therefore mixed: delivery and revenue momentum demonstrate commercial traction, but revenue remains low-diversity hardware income, software monetization is nascent, and the cost of reaching durable margin is unquantified. Underwriting requires management accounts, bank statements, debt schedules, cohort-level pricing, and a reconciled cash plan.[CI003, CI004, CI007, CI008, CI009, CI010, CI012]

4.5 Exhibits

5.1 Product lines and customer workflow

DeepWay's public product map centers on the Xingchen line for long-distance trunk and branch transport and the shorter-wheelbase Xingtu line for mid- and short-distance freight. This is a delivered hardware business rather than a software-only autonomy proposition: the first Xingchen fleet was handed over in May 2023 for logistics supporting CATL, while later sources describe Xingchen II and Xingtu as the current vehicle families. The customer workflow joins truck procurement, route-specific energy configuration, assisted highway driving, and fleet monitoring. The evidence does not establish that every named configuration is available in every geography, and English-language sources render model and system names inconsistently. Therefore, the matrix below treats use-case coverage as reported positioning rather than a complete SKU catalog.[CE001, CE002, CE008, CE013]

5.2 Architecture, control stack, and dependencies

The architecture is vertically integrated at the vehicle layer but externally dependent at the autonomy foundation. DeepWay claims a forward-designed battery-chassis platform, cross-domain electronic/electrical architecture, a 450 kWh pack, and rapid battery swapping; it also reports in-house battery, electric-drive-axle, and electronic-control capabilities. Above the vehicle sits the Tianji driving matrix and Tianshu fleet-operations layer. Baidu is a strategic dependency rather than an independent technology benchmark: an independent report describes an exclusive commercial-vehicle white-box license that lets DeepWay modify Apollo source code. This access may accelerate development, but public evidence does not disclose license durability, change-of-control terms, compute hardware sourcing, or the division of safety responsibility between Baidu and DeepWay.[CE003, CE004, CE005, CE006, CE009, CE010, CE011, CE014, CE016]

5.3 Maturity, safety, and roadmap risk

Product maturity is uneven across layers. Vehicle delivery and L2 assisted driving have the clearest evidence, while platooning and single-vehicle L4 remain a progressive roadmap. The 2021 launch release targeted L4 capability between 2024 and 2026, but newer descriptions still distinguish standard L2 assistance from future platooning and unmanned operation; the older date target should not be treated as proof of current L4 commercialization. A secondary reference reports ISO 21448 process certification, yet the assigned evidence provides no independent safety-case audit, disengagement data, recall history, uptime series, penetration test, privacy assessment, or incident register. Likewise, the absence of a recorded public incident in this pool is not evidence that no incident occurred. Investment diligence should condition any autonomy premium on primary documentation and fleet-level operating data.[CE007, CE008, CE009, CE012, CE015]

5.4 Exhibits

6.1 Customer base and buying surface

DeepWay is principally a business-to-business vehicle vendor serving logistics and supply-chain operators, with Bitauto additionally describing bulk-transport clients as a target segment. Public material frames the core purchase around an intelligent electric heavy truck for freight operations, while Dealroom reports an adjacent service layer spanning vehicle and driver management, financing support, and used-vehicle disposal. That combination suggests a buying process involving fleet operations, finance, and asset management rather than an individual driver alone. However, the allocated sources do not identify the payer, procurement owner, contract term, or revenue contribution by geography, route type, fleet size, or vertical. The segmentation is therefore useful for locating the buyer problem, but it is not a revenue-quality segmentation.[CU001, CU002, CU003, CU012]

6.2 Adoption trajectory and named deployment proof

Reported deliveries rose from 509 vehicles in 2023 to 3,002 in 2024 and 8,020 in 2025, while a separate account says DeepWay delivered about 6,400 vehicles to 311 customers in 2024. Named evidence includes the May 2023 handover to XuRi DongSheng for CATL logistics work, a reported 22-vehicle first tranche from a 72-unit Xinjiang Silk Road order, and 50 intelligent electric trucks delivered to ANE Logistics in June 2026. These records demonstrate physical fleet deployment rather than logo-only marketing. They do not, however, provide customer invoices, utilization by cohort, acceptance criteria, realized savings, or repeat-order attribution. A demonstration journey and company-distributed mileage figures show operational activity, but neither establishes customer-level return on investment.[CU004, CU005, CU006, CU007, CU008, CU010, CU015]

6.3 Durability, expansion, and concentration

The strongest durability-adjacent evidence is indirect. Bitauto reports that the top five customers' revenue share fell from 82.5% in 2023 to 39.5% in 2025, which is consistent with a broader customer base and lower concentration, but the source does not disclose the denominator, customer identities, or whether existing accounts expanded. Baidu Baike lists ANE Logistics, STO, and Maersk as recipients of intelligent products and reports more than 6,000 L2-equipped vehicle deliveries with a subscription rate above 30% by December 2025. Those figures indicate an attach opportunity, not renewal durability. No allocated source discloses gross or net revenue retention, churn, renewal rates, contract duration, cohort behavior, customer satisfaction, or the revenue contribution of repeat purchases.[CU009, CU011, CU014, CU016]

6.4 Commercial quality and diligence implications

Customer scale is currently monetized overwhelmingly through hardware. ChinaBizInsider reports that new-energy heavy-truck sales represented more than 99% of revenue during the periods it reviewed, while smart-logistics solutions remained in a free-trial phase and generated no income. That mix limits the evidence for a recurring software expansion loop even though assisted-driving subscriptions and operational mileage are reported. The next diligence step is not another logo count: it is a customer-level bridge from order to delivery, active utilization, repeat purchase, paid software attachment, renewal, and contribution margin. Without that bridge, the investment case should credit demonstrated vehicle adoption and improving concentration while treating retention, software conversion, and customer economics as material open risks.[CU011, CU012, CU013, CU014, CU015, CU016]

6.5 Exhibits

7.1 Risk ranking and financial-model exposure

The highest residual risk is the economics of scaling a capital-intensive truck manufacturer before software revenue is material. Independent reporting puts 2025 revenue at RMB3.96 billion but the annual loss at RMB649 million, while a separate 2025 account describes cumulative losses of RMB1.70 billion from 2022 through the first half of 2025. At 30 June 2025, reported liabilities of RMB5.54 billion exceeded RMB3.92 billion of assets and cash and cash equivalents were RMB593 million. These are accounting and balance-sheet measures, not cash burn or runway. Gross margin improved from 0.4% in 2023 to 2.9% in the first half of 2025, but more than 99% of revenue still came from truck sales and smart-logistics solutions remained in free trial. The mitigation is disciplined manufacturing cost reduction and conversion of software trials into paid revenue; the kill criterion is sustained low-single-digit gross margin alongside continued losses or renewed dependence on external financing for ordinary operations.[CR002, CR003, CR004, CR005, CR006, CR007, CR008, CR016]

7.2 Operational and dependency concentration

Scale has grown quickly, with multiple sources reporting 8,020 deliveries in 2025, but scale amplifies procurement, warranty, manufacturing, and service execution risk. In 2024 the top five suppliers reportedly represented 92% of procurement and the largest supplier more than 60%; the top five customers represented 51% of revenue and the largest nearly 25%. DeepWay also relied on contract manufacturing by Jianghuai and Shandong Leichi and lacked its own production license in the cited prospectus-based account. Baidu's technology license accelerates development but creates a strategic dependency around a core differentiator. Mitigation should therefore be tested through binding dual-source plans, supplier-change qualification, enforceable OEM capacity and quality terms, software escrow or continuity rights, and customer-concentration reporting. A loss of Baidu rights, a material OEM interruption, or failure to reduce single- supplier exposure before the next major volume step would break the operating thesis.[CR001, CR009, CR010, CR011, CR012]

7.3 Legal, regulatory, safety, and autonomy exposure

Public evidence supports autonomous-driving test permits in four Chinese regions and commercial cargo pilots for platooning, but those facts do not establish unrestricted driverless commercial approval. A 2021 Baidu release said the first truck generation was designed to reach L4 capability on highway routes between 2024 and 2026; later evidence still emphasizes L2 systems, permits, and pilots. That gap is a commercialization and execution risk rather than proof that the technology failed. The assigned pool contains no regulator, court, recall, or incident record sufficient to confirm the absence of litigation, enforcement, safety defects, cyber incidents, or overseas homologation barriers. Diligence must obtain permit-level scope, safety-case evidence, disengagement and incident logs, recall and warranty data, privacy and cybersecurity assessments, and country-by-country approvals. Any serious safety event, license suspension, inability to secure overseas homologation, or repeated slippage without paid autonomous deployments is a thesis-break trigger.[CR013, CR014]

7.4 Monitoring plan and investment implications

The risk program should be milestone-based rather than narrative-based. Quarterly monitoring should separate truck gross margin, warranty cost, working-capital absorption, paid software revenue, concentration by customer and supplier, OEM throughput, and safety events. Autonomy monitoring should distinguish test permits, supervised pilots, paid platooning operations, and driverless commercial authorization; kilometers alone are not proof of safe or profitable deployment. International expansion should be credited only after homologation, local service coverage, and repeat paid deliveries are documented. The market itself is concentrated: one 2025 account said the five largest vendors held 71% of China's new-energy heavy-truck market, which raises the threshold for DeepWay's cost and service differentiation. The investment implication is to retain a high residual risk rating until margin, dependency, regulatory, and recurring-software evidence improves together, rather than treating the planned IPO or pre-IPO financing as proof of adequate runway.[CR008, CR009, CR010, CR011, CR013, CR014, CR015, CR016]

7.5 Exhibits

8.1 Decision frame and recommendation

The investment case combines real commercial scale with weak price discovery. Independent coverage reports roughly 6,400 truck deliveries by June 2025 and approximately RMB 1.5 billion of first-half 2025 revenue, while multiple reports confirm a $310 million pre-IPO financing phase in 2026. Those facts establish operating momentum and access to capital, not investability at any price. DeepWay also remained loss-making, and the retained evidence does not disclose the latest post-money valuation or security terms. The appropriate public-evidence decision is therefore track/research-more rather than invest now. Confidence is medium-low and risk is high because the recommendation is directionally supported but cannot be converted into a risk-adjusted entry price or target return.[CV001, CV003, CV004, CV005, CV006, CV016]

8.2 Financing context, price opacity, and dilution

The second 2026 pre-IPO tranche brought the phase total to $310 million, with Stone Venture or Stone Capital, NGS Super, Xiamen Guosheng Fund, ABC Impact, and Nanjing Ronghe Capital named across coverage. Reporting says the proceeds are intended for operations, autonomous-driving research, and international expansion; funding raised is not the same as unrestricted cash available. A November 2025 article reported an August 2025 share price near RMB 38.94 and an implied valuation above RMB 5.4 billion, but that historical mark does not establish the 2026 round price. Aggregate funding databases also diverge sharply because of different coverage and vintages. Current ownership, fully diluted share count, liquidation preferences, anti-dilution rights, debt, and IPO primary-versus-secondary mix remain open diligence items.[CV001, CV002, CV006, CV007, CV008, CV009, CV010]

8.3 Scenario discipline and comparable evidence

Public comparables provide context but not a valuation answer. Multiples selected a mixed set of passenger-EV, industrial, and private vehicle companies and displayed revenue multiples spanning 0.8x to 13.1x; Tracxn instead emphasized autonomous-trucking and commercial-vehicle peers. That dispersion reflects materially different margin structures, technology exposure, geography, capital intensity, and maturity. Applying any one multiple to DeepWay's revenue would create false precision before normalized revenue, gross margin, net cash, dilution, and recurring software economics are known. The bull, base, and bear cases below are therefore conditional underwriting cases rather than price targets, and no quantitative probability is assigned.[CV004, CV011, CV012, CV013, CV014]

8.4 Exit readiness and thesis-break diligence

A planned Hong Kong Chapter 18C listing is an exit pathway, not evidence of completed liquidity. The 2026 financing broadens the investor base and may bridge the company toward listing, while Baidu's 17.28% reported stake and Apollo relationship remain strategically relevant. However, the public record does not establish IPO timing, offer valuation, lockups, preference conversion, or the split between primary proceeds and shareholder liquidity. The recommendation should improve only after audited current financials show narrowing losses, disclosed software economics demonstrate recurring value beyond truck sales, and round or IPO terms permit a normalized return analysis. The thesis breaks if losses and balance-sheet pressure persist without credible margin expansion, or if financing terms transfer disproportionate downside to new investors.[CV003, CV006, CV008, CV014, CV015, CV016]

8.5 Exhibits