China’s EV industry delivered a revealing set of signals this week: on March 22, the Jinhua final of the Second Intelligent Driving Competition crowned the Xingjiyuan ET as the standout performer in a demanding 27-kilometer urban NOA test, while separate UK market data showed nearly-new cars losing value at an alarming pace as Chinese entrants intensify competition and automakers cut prices. At the same time, BAIC is deepening its smart-tech push through supplier engagement around AI, multimodal perception, and advanced vehicle intelligence. Put together, the stories show the Chinese EV market entering a new phase—where software capability, ecosystem depth, and global pricing pressure matter as much as hardware.
Jinhua Smart Driving Competition: A New Benchmark Emerges
The most immediate headline came from Jinhua, where 10 vehicles faced off in a "closed-book" intelligent driving final with no prior route disclosure. The course was tough by any standard:
- 27 km total distance
- 64 traffic lights
- 5 waypoints
- 7 major test points
- ~70 minutes expected drive time
- Mixed scenarios including:
- village roads
- artificial obstacles
- community internal roads
- commercial districts
- mall internal roads
- U-turns
In that environment, the Xingjiyuan ET—a first-time entrant—took first place with a record-setting 112.05 points, setting what D1EV described as the highest score in the history of the competition.
Final standings in Jinhua
| Rank | Model | Score | Key notes |
|---|---|---|---|
| 1 | Xingjiyuan ET | 112.05 | No deductions in NOA scenarios or safety; 1 takeover |
| 2 | Exeed ET5 | 110.43 | Strong NOA scenario performance; 2 violations, 1 takeover |
| 3 | Wey Lanshan | 104.49 | Best result yet for the model; 2 takeovers |
| 4 | NIO ES6 | 97.00 | Hurt by one traffic accident deduction |
| 5 | Xiaomi SU7 | 91.27 | Only car to clear the toughest community-road test point |
| 6 | Zeekr 7X | N/A | Older software version; efficiency deductions weighed on score |
| 7-10 | Li Auto L6 / XPeng P7 / Luxeed S7 / Buick Zhijing L7 | N/A | Major deductions, especially in unfamiliar internal-road scenarios |
A particularly notable result was that the top three all scored above 100 points, suggesting a widening gap between the strongest assisted-driving stacks and the rest of the field under complex real-world conditions.
Why the Winners Stood Out
The Jinhua event was less about flashy lane changes and more about whether a vehicle could maintain composure in confusing urban edge cases. That matters because many Chinese automakers now market urban NOA as a defining product feature.
According to the event report, the winning trio relied on different supplier ecosystems:
- Xingjiyuan ET: Bosch + WeRide solution
- Exeed ET5: Horizon solution
- Wey Lanshan: DeepRoute.ai solution
That is important for two reasons.
First, it reinforces that China’s intelligent-driving race is increasingly an ecosystem competition, not just a brand competition. Second, Bosch + WeRide reportedly secured a third consecutive win, underlining the growing role of Tier 1 and autonomous-driving partners in shaping real-world performance.
The hardest scenarios exposed real weaknesses
The easiest checkpoints were village roads, parts of the business-district section, and U-turn tests. But three sections proved decisive:
- Artificial obstacle handling
- Mall internal roads
- Community internal roads
The toughest was the community internal-road scenario, where only the Xiaomi SU7 passed without issue. Several well-known names stumbled badly on unfamiliar park or campus-style internal roads, especially in road recognition and route continuity.
That is arguably the most valuable takeaway from the event: many systems perform well on standard mapped roads, but unstructured or semi-structured internal environments remain a weak spot.
NIO, Xiaomi, XPeng, Li Auto and Others: Mixed Results
For China’s best-known EV brands, the Jinhua final was a reminder that software maturity still varies sharply.
Key performances
- NIO ES6 finished fourth with 97 points. Without a major accident deduction of 10 points, it could have challenged for the podium.
- Xiaomi SU7 finished fifth with 91.27 points. That may not sound dominant, but its strong showing in the hardest test section suggests Xiaomi’s assisted-driving team is making meaningful progress.
- Zeekr 7X placed sixth despite using the only older-version smart-driving system among the field, making the result more respectable than it first appears.
- XPeng P7 had updated XOS 6.0.0.8908, but the much-anticipated XPeng VLA 2.0 was not pushed before the event.
- Li Auto L6 delivered one of the biggest surprises on the downside, with heavy deductions for wrong routing, avoidance behavior, and efficiency.
- Luxeed S7 and Buick Zhijing L7 both suffered from road-recognition failures in unfamiliar internal-road scenarios and finished near the bottom.
What the competition revealed
The most striking pattern was that several brands lost most of their points in unfamiliar internal-road environments, not on standard arterial roads. In other words, the challenge is no longer basic lane keeping or traffic-light response; it is scene generalization.
For brands like XPeng, Li Auto, Huawei-backed Luxeed, and Buick’s Momenta-supported model, that raises difficult but familiar questions:
- How robust is the perception stack in low-standardized road layouts?
- Can the planner recover when lane boundaries are ambiguous?
- How much does HD map dependence still matter?
- Are route-following and detour strategies mature enough for “human-messy” environments?
Those are exactly the questions that will define the next round of competition in Chinese autonomous-driving and advanced driver assistance systems.
BAIC’s Next Move: Building the Smart EV Supply Chain
While the Jinhua competition showed what current systems can do on the road, BAIC is focusing on how the next generation gets built.
On March 26, Gasgoo, together with BAIC Research Institute and the Beijing Society of Automotive Engineers, will host a technical exchange centered on:
- AI technologies
- Multimodal perception
- Smart vehicle systems
- Supplier-R&D collaboration
This is more than a networking event. It reflects BAIC’s broader industrial strategy as it tries to accelerate both electrification and intelligence.
BAIC’s strategic backdrop
BAIC says its “Three-Year Leap Action” will run from 2025 to 2027, targeting:
- More than 3 million vehicle sales
- A place in China’s top six automakers
The group has already been reshaping its brand structure with:
- BEIJING brand
- Kaiven/Cavan new-energy commercial vehicles
- Stelato (Xiangjie), developed with Huawei for the premium segment
- Continued development around Arcfox and intelligent connected vehicle technologies
On the technology side, BAIC highlights progress in:
- Advanced driver assistance and intelligent connected systems
- Battery and e-powertrain technologies
- Smart cockpit integration
- Off-road technologies
- Hydrogen commercial vehicles
The bigger point is that Chinese automakers are no longer just buying components—they are trying to orchestrate full-stack ecosystems across software, sensors, compute, batteries, and vehicle architecture.
UK Residual Values Show the Global Pressure Building
If Jinhua showed the performance race inside China, the UK data showed the market consequences outside it.
According to Bloomberg, citing Cox Automotive, nearly-new cars in the UK are losing value unusually quickly. In March, cars less than a year old retained only 63% of their original purchase price on average. That means a depreciation of roughly 37% in under a year.
The situation is even harsher for EVs.
UK depreciation snapshot
| Vehicle type | Residual value after <1 year | Approx. depreciation |
|---|---|---|
| Average nearly-new car | 63% | 37% |
| Nearly-new EV | Just over 50% | Nearly 50% |
| Nearly-new ICE/hybrid vehicles | 65% | 35% |
Cox Automotive said the weakness reflects several overlapping factors:
- More competition from Chinese brands
- Price cuts by automakers trying to meet UK EV sales targets
- Oversupply in the broader market
- Consumer concerns over charging infrastructure
- Rising raw-material costs linked to geopolitical tensions
For Chinese EV brands expanding into Europe, this is a double-edged sword.
On one hand, aggressive pricing and strong product value are helping them gain visibility. On the other, falling residual values can undermine leasing economics, hurt brand perception, and make customers more cautious about purchase timing.
Comparison: What These Three Stories Tell Us
Taken together, the week’s developments reveal a Chinese EV industry balancing three fronts at once.
| Theme | What happened | Why it matters |
|---|---|---|
| Smart driving validation | Xingjiyuan ET won the Jinhua final with 112.05 points | Real-world NOA performance is becoming a core competitive metric |
| Ecosystem building | BAIC deepened supplier engagement around AI and perception | Future winners will rely on supply-chain depth, not just branding |
| Global market pressure | UK nearly-new cars lost 37% of value in under a year | Chinese EV expansion is reshaping pricing and residual values abroad |
Why This Matters Globally
The Chinese EV story is no longer just about selling more battery-electric cars. It is increasingly about exporting a complete competitive model:
- Fast software iteration
- Supplier-led innovation
- Aggressive pricing
- Rapid model refresh cycles
- Real-world ADAS differentiation
That model is incredibly effective at pushing technology forward. But it also creates side effects:
- sharper depreciation
- consumer hesitation over when to buy
- pressure on used-car pricing
- harder profitability for both legacy and emerging brands
For Europe in particular, the UK residual-value trend is a warning. Chinese EVs are not just adding choice; they are changing the economics of ownership and resale.
Meanwhile, for Chinese automakers, competitions like Jinhua serve a second purpose beyond publicity: they provide public evidence of who is actually ready for wider deployment of advanced urban driver assistance.
The Road Ahead
The next stop in D1EV’s intelligent driving competition is Wuhu on April 18, and that should offer another useful benchmark for how quickly brands can improve after Jinhua’s exposed weaknesses.
The biggest question now is whether leading Chinese EV makers can do three things at once:
- Improve urban NOA reliability in edge-case environments such as internal roads and semi-structured spaces
- Scale intelligent-driving ecosystems through deeper cooperation with suppliers and AI partners
- Protect residual values as they expand globally into increasingly price-sensitive markets
For now, the week’s message is clear: in the Chinese EV market, the winners will not just be those with the biggest batteries or the cheapest sticker prices. They will be the brands that can pair smart-driving competence with disciplined market strategy—at home and abroad.
