Tesla and Huawei Signal EV Autonomy’s Next Phase

Tesla’s first week of Cybercab operations in Austin has not produced the public meltdown many critics expected. Early rides reported by outlets including MotorTrend and The Verge suggest the steering-wheel-free robotaxi is mostly stable, with complaints centered on pickup/drop-off accuracy, occasional abrupt braking, long wait times, and a limited operating area. At the same time, fresh U.S. sales data show Tesla’s EV market share rebounding to 52% in the first eight months of 2026, even as its own volume fell 16% year on year to 325,351 units. Taken together with China’s latest autonomous-driving leaderboard and real-world safety anecdotes around Huawei’s driver-assistance system, the picture is becoming clearer: the next EV battle is no longer just about selling cars, but proving that autonomy can be safe, scalable, and commercially viable.

Tesla’s Cybercab Starts Quietly, but the Stakes Are Huge

The biggest development in Austin is not that Tesla’s Cybercab is flawless. It is that a vehicle with no steering wheel and no pedals has entered real operation without immediately collapsing under public scrutiny.

According to the D1EV synthesis of early media impressions:

• MotorTrend reportedly completed five consecutive ride experiences
• The Verge also spent extended time testing the service
• The broad conclusion: the system performs more steadily than many expected

The issues raised so far are operational rather than existential:

• Imprecise drop-off points
• Occasional sudden braking
• Longer-than-ideal wait times
• Geofenced service limitations
• Regulatory constraints

That distinction matters. In earlier autonomous-driving trials, debate focused on whether the software could drive at all. In Austin, the question has shifted toward whether Tesla can make driverless mobility reliable enough, safe enough, and cheap enough to scale.

Why MPCI Is Becoming the Key Autonomy Metric

One of the most important ideas emerging from the Chinese coverage is MPCI: the average distance traveled per critical safety intervention. It is a more useful metric than viral clips or one-off test drives because it speaks directly to real-world robustness.

D1EV argues that:

• Much of China’s intelligent-driving sector is still operating at roughly the 100-km level in urban conditions
• Tesla’s FSD V14 has reportedly moved into the 1,000-km level
• The next meaningful targets are 1,000 km and then 10,000 km in city driving

Autonomy Milestones at a Glance

If an autonomous vehicle can safely operate for thousands of kilometers in dense urban traffic without critical intervention, the business equation changes dramatically. At that point, robotaxi economics begin to look less theoretical and more industrial.

Tesla’s U.S. EV Share Is Rising Again — for an Uncomfortable Reason

Tesla’s regained U.S. market share is not being driven by a sales boom. It is being driven by a broader retreat from EVs among legacy automakers.

According to Motor Intelligence data cited by D1EV:

In other words, Tesla lost volume, but rivals lost ground faster.

This follows a difficult period for the company:

• Tesla’s U.S. sales reportedly fell to 589,000 units in 2025
• Market share dropped to roughly 41%, a record low by recent standards
• Brand damage intensified amid political controversy and public protests

Yet in 2026, several traditional automakers have scaled back EV ambitions:

• Ford announced a $19.5 billion EV asset write-down in December 2025 and ended production of the all-electric F-150 Lightning, pivoting toward hybrids and EREVs
• GM took about a $6 billion special impairment in January 2026 and cut back EV capacity expansion plans
• Honda canceled three U.S.-bound EV projects
• Nissan reportedly halted the 2026 Ariya EV SUV
• AutoNews data cited in the source suggest combined write-downs tied to canceled or delayed EV programs have surpassed $70 billion

Tesla is benefiting from a weaker competitive field, but that does not necessarily strengthen the long-term case for its car business. Elon Musk’s strategic focus is increasingly centered on:

• Robotaxi
• Optimus humanoid robotics
• Artificial intelligence infrastructure

That makes Austin more than a product test. It is a public test of Tesla’s post-car identity.

China’s Autonomous Driving Race Is Splitting Into Two Paths

While Tesla tests a direct leap toward driverless operation, China’s leading players are taking two overlapping but distinct routes.

Path 1: Huawei’s L3-first approach

Huawei is prioritizing a rules-based, legally structured transition to higher autonomy. That strategy gained weight after China released its first mandatory national standard for L3/L4 autonomous driving safety in July 2026: “Safety Requirements for Autonomous Driving Systems of Intelligent Connected Vehicles.”

The standard will take effect on July 1, 2027, and Huawei was described in the source as:

• A core drafting participant
• A lead contributor on dynamic driving task safety requirements
• A participant in China’s first batch of L3 vehicle admission pilots

Huawei’s logic is pragmatic:

• Define when the machine is responsible
• Define when the human is responsible
• Bring automakers, regulators, insurers, and consumers into the same operating framework

That may prove crucial in China, where legal accountability and public trust could be just as important as raw software capability.

Path 2: XPeng, Li Auto, Momenta and others push toward L4

Other players are leaning harder into the Tesla-style trajectory:

• XPeng
• Li Auto
• Momenta
• WeRide
• Additional stack developers such as DeepRoute and Zhuoyu

The strategic bet here is that larger multimodal models, stronger perception, and richer long-sequence reasoning will ultimately outperform more incremental L3 deployment paths.

China’s robotaxi ecosystem is also broadening beyond carmakers, with mobility platforms such as:

• Didi
• Gaode (Amap)
• CaoCao Mobility
• T3 Mobility

all moving into pilot robotaxi operations.

China’s Latest Intelligent-Driving Leaderboard Shows the Field Tightening

D1EV’s August 2026 smart-driving leaderboard is based on a six-month weighted methodology:

• 50% competition data
• 30% crowd-sourced public testing
• 20% editorial evaluation

In August alone, the database absorbed:

• 11 editorial tests in Beijing
• 10 public tests nationwide
• 23 competition runs in Guiyang
• 44 total new data entries

The headline changes versus July were notable:

• NIO improved both score and rank, rising to No. 4
• Zhuoyu officially entered the ranking at No. 8 after meeting data thresholds
• Momenta, Zeekr Qianli Haohan, DeepRoute, Xiaomi, and Li Auto each slipped one place
• Horizon Robotics, Huawei, Momenta, Zeekr, XPeng, Xiaomi, and Li Auto all saw average scores decline

The reason, according to the source, was straightforward: the 23 new Guiyang competition runs averaged 78.8 points, nearly 7 points lower than the existing competition data pool.

Despite that, the overall hierarchy did not fundamentally change:

• Bosch-WeRide claimed a seven-month winning streak
• Horizon Robotics stayed No. 2 for the fourth consecutive month
• Huawei remained No. 3 for the fourth straight month

Bosch-WeRide was again the only system above the symbolic 100-point mark, helped by strong performances from the Aion N60:

• 117.81 points in the Guiyang preliminary round
• 91.96 points in the final, highest among 11 vehicles that day
• 120 points in Dongguan public testing, the first full score since the public-test program began

A Viral Huawei Safety Case Shows Why Consumer Trust Is Built Differently

Not every autonomy story is about robotaxis. Some of the most commercially important moments happen in ordinary low-speed scenarios.

A widely discussed user video highlighted Huawei’s in-car assistant “Xiaoyi” repeatedly warning a driver to lift off the accelerator after he accidentally pressed it while stopped at a red light. The driver had leaned back to retrieve earphones, mistakenly hit the throttle, and the system reportedly prevented the vehicle from moving into the Rolls-Royce Cullinan in front.

According to the account:

• The car was in D gear
• The accelerator input was detected
• The system triggered an anti-misapplication protection function
• Voice prompts repeatedly said, “Please lift your foot”
• The system later shifted to P automatically

Anecdotes are not a substitute for validated safety data, but they matter because they show how advanced driver assistance earns trust: one avoided low-speed crash at a time.

For Huawei, this kind of user experience supports its broader L3-first narrative. Before consumers accept hands-off or driver-out systems, they need to believe the software can reliably handle the mundane mistakes humans make every day.

Why This Matters Globally

The EV industry is entering a new phase in which market share, autonomous-driving capability, regulation, and AI capital spending are becoming tightly linked.

Three global implications stand out:

1. The EV race is no longer just about electrification

Battery-electric powertrains are increasingly commoditized. The bigger differentiator is now software capability:

• Urban NOA performance
• End-to-end decision-making
• Safety intervention rates
• Fleet operations economics

2. China and Tesla are converging on the same endgame from different angles

Tesla is attempting to validate pure driverless operation in public service quickly. China’s leaders are advancing through a combination of L2+/L3 commercialization, robotaxi pilots, and national standards.

The route differs, but the destination is similar: scalable autonomous mobility.

3. Capital intensity will keep rising

The source frames this as part of a wider AGI race, noting that several global tech giants have seen free cash flow swing sharply as AI investment accelerates. Even if one discounts the most aggressive AGI rhetoric, the spending trend is undeniable.

For automakers and autonomy suppliers, that means:

• Annual investment in the tens of billions of yuan may become normal
• Falling behind in model capability could become existential
• Not every company currently in the race will survive it

The Next Test: Can Anyone Reach 10,000 km MPCI in Cities?

That is the real question now. Tesla’s Austin launch suggests the conversation has moved beyond “can it drive at all?” China’s latest leaderboard suggests domestic players are improving, but still face a large gap between competent city NOA and truly scalable L4 autonomy.

In the near term, expect two things to happen at once:

• More practical safety features like anti-misapplication braking and better urban NOA will spread quickly into consumer vehicles
• More aggressive robotaxi and L4 pilots will test whether large-model autonomy can sustain intervention-free operation over much longer distances

If urban MPCI can move from the 100-km range toward 1,000 km and eventually 10,000 km, the implications will extend far beyond EVs. That would mark a turning point not just for transportation, but for how AI proves itself in the physical world.