Dongfeng: VOYAH Residuals and AI EV Strategy

Dongfeng and its premium EV brand VOYAH offered a revealing snapshot of where China’s electric vehicle market is heading on September 17, 2026: smarter cars alone are no longer enough. At the 2026 AI-Defined Vehicle Forum, Dongfeng executives and partners argued that the next phase of competition will be defined by system-level intelligence, vehicle-road-cloud collaboration, and AI spanning the full automotive lifecycle. At the same time, VOYAH’s latest residual-value data shows why these capabilities matter commercially: resale value in China’s premium new-energy segment increasingly depends on brand stability, warranty transferability, software relevance, and charging and service infrastructure.

VOYAH’s Residual Values Suggest a Maturing Premium EV Brand

For buyers in China’s high-end new-energy vehicle market, the most important resale metric is not just the first-year residual value, but how slowly that value declines over time. That is where VOYAH is trying to differentiate itself.

According to data cited by D1EV, several 2026 VOYAH models post one-year residual values between 71% and 75%, while their three-year residual values remain at 62% to 65%. Crucially, the drop from year one to year three is limited to roughly 9 to 10 percentage points.

This matters because a flatter depreciation curve typically signals stronger confidence from the used-car market. In practical terms, dealers and second owners are more willing to assign stable pricing when they believe the brand, software platform, after-sales support, and parts supply will remain viable.

VOYAH 2026 residual-value snapshot

One noteworthy takeaway is that the residual-value gap between battery electric vehicles (BEVs) and plug-in hybrids/range-extended models is narrowing. Within the same platform family, VOYAH’s EV and PHEV variants differ by only around 1 percentage point in first-year residual value. That suggests powertrain format alone is becoming less decisive than broader ecosystem factors.

Why Residual Value Now Depends on the Entire Brand System

VOYAH’s numbers are being framed not as an isolated pricing achievement, but as the outcome of four structural supports that increasingly define residual value in China’s EV market.

1. Brand stability and delivery momentum

Residual value is ultimately a market vote on whether a carmaker will still be credible years from now.

VOYAH, founded in April 2019 as Dongfeng’s premium intelligent NEV brand, benefits from the backing of a state-owned automaker with 57 years of manufacturing history. The company also listed on the Hong Kong Stock Exchange main board in March 2026 under stock code 07489.HK.

On the sales side, VOYAH delivered:

• 150,169 vehicles in 2025, up 87% year on year
• 89,453 vehicles from January to July 2026, up 31% year on year
• More than 360,000 global users as of May 2026

Those are the kinds of scale indicators that can reassure used-car dealers and fleet buyers.

2. Transferable warranty coverage

One of the most underrated resale levers in the EV market is whether warranty coverage follows the vehicle rather than the first owner.

VOYAH’s non-commercial vehicles are covered by:

• Vehicle warranty: 60 months or 100,000 km
• Battery/e-drive/EV system warranty:
  • Dreamer, Zhuiguang, Zhiyin, Taishan: 96 months or 160,000 km
  • FREE: 96 months or 120,000 km

Importantly, the remaining warranty does not expire upon ownership transfer, which strengthens confidence in the second-hand market. VOYAH also says spare parts supply will be maintained for at least 10 years after a model is discontinued.

3. Software and intelligent-driving relevance

In EVs, outdated software and driver-assistance hardware can crush resale value faster than traditional mechanical aging.

VOYAH says its technology stack spans:

• High-performance BEV architecture
• Lanhai intelligent hybrid multi-mode system
• Dedicated high-efficiency range-extender system
• Shared safety and smart architecture across multiple powertrains

The company also points to its Kunpeng intelligent driving system and Xiaoyao cockpit architecture, with products such as FREE+, the new Zhiyin, 2026 Dreamer, Zhuiguang L, Taishan, and Taishan X8 already equipped with or slated to receive Huawei Qiankun ADAS and HarmonyOS cockpit technologies.

By March 2026, VOYAH had accumulated:

• 6,028 patent filings
• 2,077 granted patents
• 5,084 invention patents, representing 84.34% of the total patent layout

That IP concentration in new energy, autonomous driving, intelligent connectivity, and body safety supports the argument that software-defined value retention is now real.

4. Charging and service convenience

For long-distance family users, usability directly affects used-car desirability.

VOYAH says its app integrates more than 1 million charging resources nationwide, covering over 2,700 county-level districts and China’s major “ten vertical, ten horizontal” expressway corridors. It is also promoting the VP1000 megawatt-class ultra-fast charger.

The service footprint is equally important:

• 579 stores globally as of June 2026
• Presence in 233 cities
• Direct-sales model combining self-operated + ecosystem retail

A standardized direct-sales and service structure can improve pricing transparency in the second-hand market, especially for premium EVs where condition assessment often depends on software, battery health, and service history.

How Dongfeng Sees the Next AI Battleground

If VOYAH’s resale story explains the commercial side of intelligent EVs, Dongfeng’s forum presentation explains the technical side.

At the AI-Defined Vehicle Forum, Dongfeng R&D Institute senior engineer Liu Zhifeng argued that “single-vehicle intelligence is approaching its long-tail boundary” and that vehicle-road-cloud-chip collaboration will become the industry’s second growth curve.

That is a significant strategic statement. It suggests that as L2 driver assistance becomes mainstream and L3 begins gradual deployment, competitive advantage will shift away from adding isolated in-car features and toward broader system integration involving infrastructure, cloud services, standards, and operational models.

Dongfeng’s intelligent architecture

Dongfeng says it is building its “Tianyuan Intelligence” system around a framework described as “one core, two foundations, two applications”:

• One core: enterprise-level Taiji large model
• Two foundations: Tianyuan electronic/electrical architecture and Tianyuan OS
• Two applications: Tianyuan intelligent driving and Tianyuan smart cockpit

According to the company, this architecture deeply integrates the:

• Power domain
• Body domain
• Cockpit domain
• Intelligent-driving domain

In cockpit-driving integration, Dongfeng disclosed several performance targets:

• 1 millisecond-level cross-domain communication latency
• Around 300 milliseconds intelligent-agent cross-domain scheduling response
• More than 92% cross-domain collaborative task success rate
• 15% to 30% reduction in component count
• More than 15% overall cost reduction

These are meaningful numbers because they show how software-defined vehicles can improve both functionality and manufacturing economics.

From Smart Cars to Vehicle-Road-Cloud Collaboration

Dongfeng’s bigger message, however, is that the next leap in autonomous and connected mobility may not come from the car alone.

The company is already participating in demonstrations and operations in:

• Wuhan
• Shiyan
• Xiangyang
• Beijing Yizhuang
• The Beijing-Tianjin-Hebei highway collaboration project

In Shiyan, more than 230 manned vehicles and 15 unmanned logistics vehicles are already in regular operation. Dongfeng says vehicle-road-cloud collaboration improved overall operating efficiency by more than 20% compared with standalone vehicle intelligence.

On the Rongwu Expressway in Hebei, Dongfeng and partners are exploring two-way data flows between vehicles and the highway itself. That changes the role of the vehicle: it is no longer just a consumer of road data, but also a provider of information to highways, city management, inspections, and safety systems.

In other words, the car is evolving into a mobile intelligent node inside a larger transport network.

Why AI in Cars Must Go Beyond the Cabin and ADAS

A third theme from the forum came from SONATUS co-founder and CTO Fang Yu, who argued that today’s industry conversation about AI-defined vehicles is still incomplete.

His point is important: AI should not be limited to the user-facing layers of autonomous driving and human-machine interaction. The bigger opportunity is to connect the entire automotive lifecycle:

• R&D
• Testing
• Manufacturing
• End-of-line validation
• After-sales service
• Diagnostics

Fang argued that traditional vehicle development is fragmented across teams, software tools, and databases. Valuable knowledge gets trapped in source code, engineering models, Jira tickets, historical test data, and vehicle telemetry. AI can help connect these silos, understand relationships across tools, and generate problem analysis, test cases, and possible solutions.

A practical AI efficiency example

In a case cited from the Nissan Technical Center, combining R&D information with vehicle test data and analyzing it with AI reduced fault-investigation time from roughly two weeks to two days.

That kind of time compression has major implications for quality control, warranty costs, and vehicle software iteration speed.

SONATUS says its Fastlane product line is designed around this lifecycle concept:

• Fastlane Collector for vehicle data collection
• Fastlane Edge for on-vehicle edge computing
• Fastlane Insight for AI-agent analysis in the cloud

Together, they create an “observe-analyze-act” loop.

Just as importantly, Fang stressed that automotive AI needs deterministic correctness in critical scenarios. An AI assistant recommending a restaurant can tolerate ambiguity; vehicle diagnostics and safety systems cannot. That distinction is likely to become increasingly central as generative AI expands into automotive engineering workflows.

What Buyers Should Watch in VOYAH’s Lineup

VOYAH’s portfolio now spans multiple body styles and price bands, giving it broader coverage in China’s premium and upper-mid EV space.

Recommended by usage scenario

• Family long-distance travel / comfort-first buyers:

  • 2026 VOYAH Dreamer series
  • Flagship MPV positioning
  • Residual values from 71% to 73% after one year

• Large families / SUV buyers wanting space and capability:

  • VOYAH Taishan series
  • Large six-seat flagship SUV
  • Best residual result in the set: 75% after one year for Taishan Max+

• Buyers targeting the RMB 200,000-300,000 range:

  • VOYAH FREE+: RMB 219,900-279,900
  • VOYAH Zhiyin: RMB 202,900-242,900
  • VOYAH Taishan X8: RMB 292,900-379,900

This broader product spread matters because residual value tends to improve when brands can sustain volume across multiple categories rather than relying on a single hero model.

Why This Matters Globally

China’s EV market is often discussed in terms of price wars, battery technology, and autonomous-driving features. But the developments highlighted here point to a more mature phase of competition.

Three trends stand out:

1. Residual value is becoming a strategic KPI for premium EV brands, not just a used-car metric.
2. Software-defined competitiveness is shifting from the vehicle to the ecosystem, including cloud architecture, roadside infrastructure, and after-sales intelligence.
3. AI’s highest near-term value may come from industrial efficiency and risk reduction, not only in-car novelty features.

This has implications beyond China. Global automakers and suppliers are grappling with the same questions: how to keep EVs technologically relevant over time, how to control lifecycle costs, and how to integrate AI safely into both the vehicle and the organization behind it.

The Road Ahead

VOYAH’s residual-value data and Dongfeng’s AI roadmap ultimately tell the same story from two different angles. In the next stage of the Chinese EV market, winning will depend less on headline specs alone and more on whether a brand can maintain trust across ownership cycles, software updates, charging access, service consistency, and intelligent-system evolution.

For consumers, that means a premium EV should increasingly be judged like a tech platform with automotive durability. For manufacturers, it means the future is not just about building a smarter car, but a more coherent system around it.