China’s EV and autonomous-driving story moved several steps forward this week, with Tesla outlining a three-month timeline for Grok-powered FSD voice control, Li Auto intensifying its push to be seen as an AI company, Huawei preparing an ADS 5 rollout, and Chinese autonomous-driving startup Momenta reportedly nearing a Hong Kong IPO at a valuation of about $9 billion. At the same time, China’s auto supplier sector is accelerating M&A, with 67 listed auto companies announcing deals in the first half of 2026 through June 12, underscoring how software, chips, autonomy and supply-chain control are converging into one high-stakes competition.
Tesla Pushes FSD Toward Natural-Language Driving
The biggest global headline came from Tesla. According to D1EV, CEO Elon Musk said on X that Grok voice control for Tesla FSD (Supervised) could arrive in about three months. If the timetable holds, the feature would begin rolling out in autumn 2026 across Tesla’s current lineup.
This is more significant than a routine voice-assistant update. The reported goal is a deeper connection between the Grok large language model and Tesla’s FSD decision chain, allowing drivers to use natural language for tasks such as:
- lane changes
- route adjustments
- parking instructions
- destination-specific maneuvers
A practical example cited in the report is a driver saying: navigate to a hardware store and back into a parking space near the storefront on arrival. If Tesla can make that work reliably, it would mark a meaningful shift from menu-based commands to intent-based driving interaction.
Tesla is also expanding Cybercab testing beyond North America. Videos on social media reportedly showed two gold Cybercab prototypes headed to Wellington, New Zealand, where winter conditions in the Southern Hemisphere can help validate performance on icy, low-temperature roads. That matters because camera-based perception systems and supporting hardware must prove robust in poor weather if robotaxi services are to scale internationally.
Tesla has already pushed FSD V14 version 14.3.3 to Australia and New Zealand, according to the source, raising the possibility that Oceania could become Tesla’s first overseas robotaxi market. In the US, Cybercab mass production reportedly began at the Texas Gigafactory in late April, with more than 100 vehicles awaiting delivery and both steering-wheel-free and manually controlled variants being built for different regulatory markets.
Li Auto’s AI Reinvention Is Ambitious — but Not Yet Fully Convincing
While Tesla is shipping products, Li Auto is still trying to win the narrative battle.
A D1EV analysis argues that Li Auto founder and CEO Li Xiang has spent the past year repositioning the company from a successful carmaker into an artificial intelligence company. That effort has included:
- reducing dependence on Huawei-related software architecture
- open-sourcing its Xinghuan OS
- advancing in-house chip development
- launching its first AI glasses product
- reframing the company’s long-term identity around embodied intelligence
At Li Auto’s June 15 software and embodied-intelligence event in Beijing, Li Xiang laid out the company’s “next decade” vision: moving from building a “mobile home” to giving both cars and homes “life.”
The event showcased several core technologies:
- Mach M100 chip
- Mach Mind large model
- Mach VLA intelligent-driving system
- a full-year OTA roadmap
- a claim that its future system could exceed human performance in safety and efficiency
The standout hardware claim is the Mach M100. Li Auto says it is the world’s first mass-produced dynamic dataflow AI chip for production vehicles, built on a 5 nm automotive-grade process, with 1,280 TOPS per chip and more than 82% real-world operating efficiency. A dual-chip setup in the new Li L9 Livis would deliver 2,560 TOPS.
Li Auto’s AI Challenge: Compute Is Not the Same as Data
On paper, those figures are impressive. But the harder question is whether compute can compensate for weaker real-world data scale.
Tesla’s FSD stack benefits from millions of vehicles generating driving data. Huawei’s ADS platform draws from a broad ecosystem of partner brands. Li Auto, by contrast, is still seen as being behind the largest players in accumulated assisted-driving mileage and real-world edge-case exposure.
That creates a strategic tension: Li Auto is trying to use proprietary silicon and simulation to narrow the gap.
The company’s logic appears to be:
- build the chip stack in-house
- control the full software-to-hardware toolchain
- train more aggressively with simulation and synthetic data
- use OTA updates to close capability gaps faster
That is a credible strategy, but it is also capital intensive and time sensitive. In the near term, Li Auto’s AI push may matter as much for talent retention and investor confidence as it does for immediate competitive advantage in assisted driving.
Huawei and XPeng Add Fuel to China’s Smart Driving Debate
Tesla and Li Auto are not operating in a vacuum. Chinese competitors are also moving quickly.
Huawei’s smart driving arm is reportedly preparing to push ADS 5, initially for flagship HarmonyOS-enabled models. The system was officially launched on April 23 and features the WEWA 2.0 architecture, which introduces:
- multi-agent game-theory methods in the cloud
- online reinforcement learning
- real-time safety risk field assessment in the vehicle
- a DrivingAgent module for strategy optimization
Huawei also published fresh usage figures. As of May 31, Qiankun ADS had accumulated 11.47 billion km of real-world user driving mileage. The company’s website now shows assisted-driving mileage surpassing 12 billion km, while total mileage of vehicles equipped with the system has reached 35.4 billion km.
Those are not just vanity metrics. In intelligent driving, fleet miles help improve corner-case learning, validation depth and OEM confidence.
Meanwhile, XPeng moved to debunk rumors around upcoming autonomous-driving regulations. A company executive rejected claims that China’s proposed standards would require:
- dual redundancy for all L3 systems
- LiDAR for all L4 systems
XPeng’s position is that the draft safety requirements are outcome-based rather than prescribing a specific sensor stack. That is an important distinction.
Sensor Strategy Is Becoming a Defining Fault Line
The XPeng clarification highlights one of the industry’s most important technical divides: whether future advanced driver-assistance and autonomous systems will require LiDAR as standard, or whether camera-heavy and AI-first approaches can meet regulatory and safety targets.
For investors and consumers, this is more than a component debate. It affects:
- bill of materials cost
- software complexity
- validation requirements
- scalability across vehicle segments
- long-term gross margins
Momenta’s $9 Billion IPO Plan Shows Capital Still Believes in Autonomy
If public-market enthusiasm for EV makers has become more selective, capital is still willing to back autonomy specialists with OEM relationships and international partnerships.
According to The Wall Street Journal, as cited by D1EV, Momenta has reached an approximate valuation of $9 billion and is preparing for a Hong Kong IPO that could raise around $1 billion. The company is based in Suzhou and supplies intelligent-driving solutions to automakers.
Its backers reportedly include:
- General Motors
- Toyota
- SAIC Motor
- Mercedes-Benz
- Bosch
- Temasek
- Tencent
- Yunfeng Capital
Commercially, Momenta has already been linked with robotaxi and mobility collaborations in:
- Munich with Uber
- Abu Dhabi with Mercedes-Benz
- Southeast Asia with Grab
Regulatory filings cited in the report show Momenta plans to issue up to 43.75 million shares in Hong Kong. If completed, it would become another major Chinese autonomous-driving company to tap Hong Kong after peers such as WeRide and Pony.ai.
The timing is notable. Hong Kong Exchanges and Clearing data cited in the source show companies raised about HK$166.8 billion in the first five months of 2026, more than double the HK$79.2 billion raised in the same period a year earlier.
China’s Auto Suppliers Are Racing for Control, Not Just Exposure
Beyond headline-grabbing EV brands, a quieter but equally important shift is unfolding in China’s parts industry.
According to data cited by Economic Information Daily and iFinD, 67 listed auto companies announced M&A deals in the first half of 2026 through June 12, with parts suppliers leading the charge.
The most important pattern is that buyers increasingly want control, not minority stakes. The logic is straightforward: when industry growth slows and technology shifts accelerate, passive financial investments are less useful than directly controlling product roadmaps, customer relationships and manufacturing capacity.
Key M&A Themes in China’s Auto Supply Chain
- Vertical integration: buyers are filling product gaps upstream and downstream
- System-level capability: OEMs increasingly want integrated solutions, not single parts
- Global expansion: overseas acquisitions help secure customers and channels in Europe and North America
- Chassis intelligence: steering, brake-by-wire and electronic control assets are becoming more strategic
Key Numbers at a Glance
| Topic | Company | Key Figure | Why It Matters |
|---|---|---|---|
| FSD voice control rollout | Tesla | ~3 months | Signals LLM integration into driving interaction |
| Cybercab production | Tesla | 100+ vehicles reportedly awaiting delivery | Suggests robotaxi program is moving from prototype to fleet stage |
| AI chip compute | Li Auto | 1,280 TOPS per Mach M100 | Shows aggressive vertical integration in automotive AI silicon |
| Dual-chip compute | Li Auto L9 Livis | 2,560 TOPS | Positions Li Auto for simulation-heavy ADAS training |
| Real-world smart driving mileage | Huawei ADS | 11.47 billion km by May 31 | Indicates large-scale fleet learning |
| Assisted-driving mileage | Huawei ADS | 12+ billion km | Reinforces data advantage in training and validation |
| IPO valuation | Momenta | ~$9 billion | Highlights investor confidence in Chinese autonomous driving |
| Hong Kong IPO target | Momenta | ~$1 billion | Could become one of 2026’s key mobility listings |
| Auto-sector M&A announcements | China listed auto firms | 67 deals | Shows rapid restructuring of the EV supply chain |
Why This Matters Globally
What is happening in China’s EV industry is no longer just about vehicle sales volumes. It is about the industrial stack beneath the car.
Tesla, Li Auto, Huawei, XPeng and Momenta are all competing across overlapping layers:
- AI models
- autonomous driving software
- chips and compute architecture
- OTA-defined product cycles
- robotaxi deployment
- data accumulation at fleet scale
China’s domestic market gives these players a rare combination of scale, competition and iteration speed. That matters globally because the winners in China are likely to shape:
- future autonomous-driving cost curves
- software-defined vehicle architectures
- chip localization strategies
- smart-cockpit and voice-AI standards
- export competitiveness in Europe, Southeast Asia and beyond
The supplier M&A wave adds another dimension. As automakers demand system-level integration, the old parts business is becoming a software-electronics-chassis business. The suppliers that secure control of steering, braking, BMS, electronics and manufacturing assets will be better placed to serve next-generation EV platforms.
The Big Picture: China’s EV War Is Becoming an AI and Supply-Chain War
This week’s developments point to a broader conclusion: China’s EV contest is evolving from a product race into a full-stack technology race.
Tesla is trying to make vehicle interaction more conversational while scaling Cybercab internationally. Li Auto is investing heavily to prove it deserves a place in the top tier of automotive AI. Huawei is leveraging fleet data and partner scale to strengthen ADS. XPeng is defending a flexible, software-led view of autonomous-driving architecture. Momenta is turning technical credibility into capital-market momentum. And suppliers across the country are buying control to avoid being commoditized.
The next milestones to watch are clear:
- whether Tesla delivers Grok-FSD voice control on schedule
- whether Li Auto’s Mach stack translates into measurable real-world ADAS gains
- how quickly Huawei ADS 5 reaches production scale
- whether Momenta’s Hong Kong IPO prices strongly
- which supplier acquisitions produce genuine system-level advantages
In short, the Chinese EV market is no longer just building electric cars. It is building the AI-defined automotive industry that much of the world may soon have to compete with.
