China’s electric-vehicle industry is being reshaped by a powerful new force: AI. Over the past few days, three developments have underscored how quickly the market is evolving. First, a widening AI-driven chip shortage is pushing up the cost of memory and potentially the sticker prices of cars, smartphones, and PCs. Second, automakers and suppliers are accelerating their investment in advanced driver assistance and robotics, from Leapmotor’s promised ADAS upgrades to Mercedes-Benz-backed automated charging. Third, the market conversation around powertrains is maturing, with range-extended EVs gaining broader acceptance as a practical solution for China’s diverse driving conditions.
AI Is Now Affecting EV Prices, Not Just Features
The biggest near-term story for the Chinese EV market may not be a new model launch, but the cost shock coming from AI infrastructure.
According to Bloomberg-cited IDC data referenced by D1EV, major tech companies’ capital expenditure could reach $650 billion in 2026, up about 80% year-on-year. That spending is soaking up memory supply as data centers demand more DRAM and especially HBM (high-bandwidth memory) for AI training and inference.
This matters for automakers because automotive electronics increasingly rely on advanced memory for:
- smart cockpits
- ADAS and autonomous driving compute platforms
- infotainment and connectivity systems
- over-the-air software functions
D1EV reports that data centers’ share of global DRAM consumption has risen from 32% five years ago to around 50% today, and could exceed 60% by 2030. As chipmakers prioritize higher-margin HBM, traditional DRAM supply is tightening.
For the auto sector, the most striking figure is the reported 300% surge in DDR5 automotive memory pricing. That is a serious issue for next-generation EVs, especially those built around powerful domain controllers and AI-heavy software stacks.
Cost Pressure by Sector
| Sector | Reported Impact | Key Detail |
|---|---|---|
| Data centers | Rapid memory demand growth | DRAM share rose to ~50% of global demand |
| PCs | Sharp bill-of-materials pressure | Memory/storage cost share in laptops rose from 15% to ~35% |
| Smartphones | Potential retail price increases | Counterpoint expects memory inflation to lift BOM costs by 15%+ |
| Automobiles | Rising smart-vehicle component costs | Automotive DDR5 prices reportedly up 300% |
D1EV also noted that the updated Zeekr 007 GT could face added cost pressure because it uses scarcer DDR5X memory, with a possible price increase of RMB 5,000 to RMB 8,000. Meanwhile, NIO founder William Li reportedly described memory inflation as the company’s biggest cost pressure, and Xiaomi founder Lei Jun said memory price increases alone could add several thousand yuan per vehicle.
The broader implication is clear: AI is no longer just a selling point in EV marketing. It is becoming a structural cost driver across the supply chain.
Smart Driving Competition Is Entering a New Phase
At the same time, Chinese automakers are stepping up their investment in intelligent driving, even as costs rise.
Leapmotor founder Zhu Jiangming said the company’s ADAS and AI efforts will deliver several surprises in the second half of the year. He openly acknowledged that Leapmotor started later than some rivals in smart driving and AI, partly due to weaker financing capacity, but said the company is now scaling up:
- larger intelligent-driving teams
- higher capital investment
- a shift from rented compute to purchased compute resources
- progress toward first-tier industry capability
That is notable because it reflects a broader trend in China’s EV sector: once cost-focused brands are now moving aggressively into software-defined vehicle competition.
Zhu also laid out ambitious volume goals. After saying the company exceeded expectations with 590,000 vehicles sold in 2025, Leapmotor is targeting 1 million units in 2026, supported by four new models and continued sales growth from existing vehicles.
Whether those targets are fully achievable remains to be seen, but the direction is unmistakable. Volume growth in China’s EV market increasingly depends on offering a stronger mix of:
- competitive pricing
- credible ADAS features
- differentiated in-car software
- product variety across segments
Meanwhile, Xiaomi’s Lei Jun said 2026 will be a breakout year for L3/L4 autonomous driving and embodied AI foundation models. That statement aligns with growing industry optimism that China is approaching a new commercialization cycle for higher-level automation, even if regulatory rollout and real-world deployment will still take time.
Range-Extended EVs Are Gaining Respectability
Another important shift is philosophical rather than technological: the debate over EV powertrains is becoming more pragmatic.
Li Auto product line head Tang Jing argued that criticism of range-extended electric vehicles (EREVs) has faded for two main reasons:
- Some brands that previously criticized the technology have now adopted it themselves.
- More people now understand that managing both range-extension and pure-electric systems is actually complex.
That complexity includes:
- energy management strategy
- NVH tuning during engine start-stop events
- coordinated thermal management
- extreme-condition drivability and reliability
His broader point is difficult to dispute in the Chinese market. Battery-electric vehicles work very well in dense urban clusters such as Beijing, Shanghai, Guangzhou, and Shenzhen, where charging infrastructure is strong. But in lower-tier cities, county-level markets, and long-distance intercity use cases, EREVs can still offer a better ownership fit.
BEV vs EREV in China’s Real-World Market
| Technology | Best-Case Scenario | Main Advantage | Main Challenge |
|---|---|---|---|
| BEV | Large cities with dense charging networks | Lower running costs, simpler drivetrain | Charging convenience varies by region |
| EREV | Broader geographic and mixed-use driving | Better flexibility for long trips and weaker charging coverage | More system complexity |
This is a meaningful sign of market maturity. China’s new-energy vehicle sector is moving away from ideological arguments and toward user-scenario matching. That is especially important as competition intensifies and brands search for growth outside top-tier cities.
Charging Is Becoming a Robotics Use Case
While software and chips dominate headlines, charging technology is also moving into a new phase.
D1EV reports that Nonox and Mercedes-Benz are jointly developing an adaptive robotic charging station that automates the entire charging process. The concept is straightforward but potentially significant: the driver parks, uses an app to initiate charging, and the robot handles the rest.
The workflow includes:
- receiving the charging command
- opening the charge port using 3D AI vision
- picking up the charging gun
- inserting the connector using force-control robotics
- unplugging and closing the charge port after charging completes
The system’s technical highlights include:
- 9-axis high-degree-of-freedom force control
- 3D AI vision for recognizing ports under varied conditions
- tolerance for positioning error through adaptive alignment
- CE and ETL safety certification
- support for 800V architecture and charging power up to 320 kW
Robotic Charging Station Snapshot
| Feature | Details |
|---|---|
| Partner brands | Nonox and Mercedes-Benz |
| User operation | App-based one-click charging request |
| Vision system | 3D AI vision |
| Motion/control | 9-axis robotic force control |
| Charging platform | 800V |
| Peak charging power | Up to 320 kW |
| Safety certification | CE and ETL |
This kind of automated charging could be especially useful in:
- premium EV parking environments
- fleet depots and commercial operations
- accessibility-focused charging scenarios
- autonomous vehicle infrastructure in the future
In other words, charging is no longer just about faster power delivery. It is becoming part of the broader machine-to-machine mobility ecosystem.
AI’s Broader Momentum Is Feeding the Auto Industry
Although not directly automotive, the AI backdrop matters. D1EV highlighted that Chinese large models logged 4.19 trillion tokens in weekly calls from March 2 to March 8, up 34.9% week-on-week, surpassing US models at 3.63 trillion tokens, which fell 8.5% over the same period.
Among the global top five by weekly usage:
- MiniMax M2.5 ranked first with 1.87 trillion tokens
- DeepSeek V3.2 ranked third with 0.83 trillion tokens
- Step 3.5 Flash entered fifth with 0.75 trillion tokens, up 69%
Why does that matter for EVs? Because China’s auto industry is deeply intertwined with the country’s AI stack. Foundation models, cloud services, edge compute, and agent-based systems are increasingly relevant to:
- in-car assistants
- cockpit UX
- voice interaction
- smart manufacturing
- enterprise workflow automation across OEMs and suppliers
Tencent’s launch of the WorkBuddy AI agent and testing of the QClaw one-click OpenClaw deployment package further signals how quickly Chinese technology companies are pushing AI agents into mainstream workflows. For automakers, that means AI competition will extend beyond the vehicle itself and into design, engineering, aftersales, and enterprise productivity.
Why This Matters Globally
China’s EV industry is often discussed in terms of price wars and export momentum, but this week’s developments show a more complex reality.
Three structural themes are converging:
- AI is raising capability ceilings for smart driving, robotics, and vehicle software.
- AI is also raising costs, especially through memory and semiconductor shortages.
- Technology pluralism is winning, with BEVs, EREVs, automation, and robotics all finding valid use cases.
For global automakers and suppliers, the lesson is that China is not just scaling EV production. It is becoming a live test bed for the integration of:
- automotive AI
- embodied intelligence
- robotics-enabled charging
- multi-path electrification strategies
If memory shortages persist into 2027-2028, as industry expectations cited by D1EV suggest, the winners may be the companies that combine software ambition with supply-chain discipline.
What to Watch Next
Several questions will shape the next phase of the Chinese EV market:
- Will memory inflation materially push EV retail prices higher in 2026?
- Can brands like Leapmotor translate heavier AI investment into meaningful ADAS differentiation?
- Will robotic charging move beyond pilot deployments into real commercial scale?
- How fast will L3/L4 deployment progress under Chinese regulation?
- Will EREV acceptance continue expanding as automakers chase lower-tier city demand?
The common thread is simple: AI is no longer adjacent to the EV business. In China, it is becoming central to product planning, manufacturing, ownership experience, and even pricing. That shift may define the next chapter of the world’s most competitive electric-car market.
