Deepal S07 Leads Smarter EV Shift From China

China’s mobility sector delivered a revealing snapshot of where the industry is headed this week. On June 13, Deepal launched the new S07 EREV starting at RMB 159,900, bringing Huawei’s ADS Pro intelligent driving suite, a 27-sensor perception stack, and CAS 4.0 active safety into the mass-market SUV space. At the same time, two other developments underscored the broader direction of travel: Renault is pushing automotive electrification know-how into military mobility with a new hybrid tactical vehicle, while AgiBot’s humanoid robot reached a new milestone by autonomously playing table tennis in real time.

Taken together, these stories point to a bigger trend across the EV industry: software-defined vehicles, sensor fusion, high-speed perception, and embodied AI are no longer isolated technology tracks. They are increasingly converging across passenger cars, robotics, and even defense applications.

Deepal S07 EREV Launch: Huawei ADS Pro Goes More Mainstream

Deepal officially launched the updated S07 extended-range EV on June 13 with a starting price of RMB 159,900. The headline feature is Huawei Qiankun ADS Pro, paired with intelligent parking assistance and the CAS 4.0 omnidirectional collision avoidance system.

This matters because it pushes advanced driver assistance technology further into a more accessible price band in China’s fiercely competitive EV market. Rather than positioning smart driving as a premium-only feature, Deepal is clearly leaning into the idea of mass adoption.

Key launch highlights

• Starting price: RMB 159,900
• Powertrain: EREV (extended-range electric vehicle)
• Intelligent driving system: Huawei Qiankun ADS Pro
• Safety suite: CAS 4.0
• Sensor count: 27 high-precision sensors
• Cloud computing support: 45 EFLOPS
• Cloud training mileage: 1 billion km
• Cloud algorithm iteration cycle: every 4 days
• Average takeover mileage on highway assist: 1,500 km
• Limited-time subsidy for ADS package: RMB 6,000

A 27-Sensor Stack Signals How Fast ADAS Is Maturing

One of the most notable aspects of the new Deepal S07 is the sophistication of its perception hardware. The SUV uses 27 sensing units, including:

• 1 LiDAR-vision integrated sensor
• 3 4D millimeter-wave radars
• 12 ultrasonic radars
• 11 high-definition cameras

Huawei says its Limera laser-vision solution integrates LiDAR and camera functions in a way that improves spatiotemporal synchronization while reducing the need for traditional fusion alignment. In practical terms, the promise is better obstacle detection accuracy, especially for small or irregular objects in low-light conditions.

According to the launch material, the system can detect unusual roadside or roadway obstacles such as:

• traffic cones
• crash barrels
• cardboard boxes
• water barriers
• warning triangles

Detection is claimed at distances of up to 115 meters for such non-standard objects. That is especially relevant in China, where urban construction zones, temporary lane changes, and mixed road users create highly variable driving environments.

City, Highway, and Parking: The Real Battleground for Smart Driving

Huawei ADS Pro on the Deepal S07 is built around a new WEWA architecture, with cloud and vehicle-side upgrades designed to improve both training and deployment. This is important because China’s ADAS race is no longer just about adding hardware; it is about how quickly systems can learn, iterate, and expand scenario coverage.

Claimed smart driving capabilities

Urban navigation assist

The system is designed to handle complex city scenarios such as:

• narrow-road passing
• roundabouts
• temporary construction detours
• lane changes and U-turns
• mixed traffic with pedestrians, bicycles, and tricycles

Deepal and Huawei emphasize smoother, more conservative city behavior, with the vehicle able to judge whether to slow, brake, or reroute around sudden cut-ins or vulnerable road users.

Highway navigation assist

On highways, the system supports:

• smart on-ramp and off-ramp transitions
• lane changes and overtaking
• curve handling
• obstacle avoidance
• construction zone recognition
• adaptive speed adjustment based on traffic flow

The claimed 1,500 km average takeover mileage is a notable marketing figure, though as always with ADAS metrics, real-world performance will vary widely by route, weather, traffic density, and driver behavior.

Intelligent parking

Parking remains one of the most commercially valuable ADAS features in China, especially in dense cities. The Deepal S07 supports:

• remote exit parking
• track-based reverse assist
• custom parking space parking
• narrow-space parking
• payment integration for partner parking lots

Deepal says the parking payment feature works across more than 200,000 cooperative parking lots, with a QR code appearing automatically on the center screen when exiting.

CAS 4.0: Active Safety Is Becoming a Selling Point, Not Just a Compliance Box

If Chinese automakers spent the last few years selling screen size and acceleration, the next phase may be defined by how convincingly they sell safety. Deepal’s new S07 makes CAS 4.0 a core part of its pitch.

Huawei describes the system as covering five dimensions:

• all speeds
• all directions
• all targets
• all weather
• all scenarios

Notable safety functions

• Forward AEB activation from 4 km/h to 130 km/h
• Rear-end collision mitigation
• Pedestrian crossing response
• Oncoming traffic risk response
• Side blind-spot active avoidance
• Door opening warning
• Rear cross-traffic braking
• Misapplication of accelerator prevention
• Driver incapacitation assistance
• Detection of irregular obstacles and dropped cargo

The mention of “ghost probe” pedestrian crossings, nighttime stationary vehicle detection on highways, and incapacitated-driver scenarios reflects a broader shift in China’s EV market. Brands increasingly know that buyers are paying attention not just to whether a car can drive itself in ideal conditions, but whether it can avoid the edge-case accidents that dominate social-media scrutiny.

Deepal S07 EREV at a Glance

Beyond EVs: Why Renault’s Defense Push Is Relevant

A seemingly unrelated story from Europe actually fits the same technology narrative. On June 15, Renault Group said it would work with Thales to develop military vehicles, broadening its role in defense manufacturing as Europe accelerates rearmament.

The prototype, called the 4 TROOP, debuted at the Eurosatory defense exhibition near Paris. Renault describes it as a four-wheel-drive hybrid vehicle designed for:

• reconnaissance
• troop coordination
• area patrol
• drone deployment
• unmanned ground vehicle deployment

It also includes V2L (vehicle-to-load) capability, allowing it to power military electrical equipment in the field.

Why this matters for EV watchers

This is not just a defense industry headline. It shows how automotive capabilities developed for the EV era are becoming transferable across adjacent sectors:

• electrified powertrains for lower fuel dependence and flexible energy use
• software integration for mission-specific functionality
• scalable manufacturing for faster deployment
• onboard power export through V2L

Renault CEO Francois Provost reportedly said the advantage of the auto industry lies in speed, contrasting defense programs that can take decades with automotive product cycles that can deliver in 12 months. That is a telling remark in an era when vehicle platforms are increasingly judged not only by consumer value, but by modularity, rapid iteration, and production efficiency.

AgiBot’s Table Tennis Robot Shows the Other Side of the Same Tech Race

The third development came from China’s robotics sector. On June 15, AgiBot announced that its full-size bipedal humanoid robot Expedition A3 became the first of its kind to autonomously play table tennis through full-process self-decision-making.

The feat is more than a gimmick. Table tennis is a brutal test of machine perception and motion control because the ball can exceed 5 meters per second, while spin, bounce, and trajectory change constantly.

According to the company, the robot completed the task without:

• remote control
• scripts
• human intervention

The system reportedly relies on:

• real-time visual perception
• trajectory prediction
• whole-body motion planning
• precise strike execution

A key enabler was a collaboration with Peking University teams, including the SpikePingpong control algorithm and a 20 kHz high-frequency event camera, which reportedly improved visual response speed by 10x versus traditional approaches and enabled millimeter-level prediction of paddle contact points.

What Humanoid Robots Have to Do With Chinese EVs

Quite a lot, actually. China’s smart EV and robotics ecosystems increasingly overlap in several critical areas:

The broader takeaway is that China’s next industrial champions may not be confined to one category. The firms building advanced driver assistance systems, embodied AI, robotics controllers, and high-speed perception stacks are often pulling from the same talent pools, supply chains, and algorithmic foundations.

Why This Matters Globally

The Deepal S07 launch is the most directly relevant news for consumers, but all three developments reinforce several important market themes.

1. Advanced ADAS is moving downmarket

A starting price of RMB 159,900 for an EREV with a high-profile Huawei smart driving stack shows how quickly advanced driver assistance is moving toward the mainstream in China.

2. Safety is becoming central to EV competition

Features like AEB across 4-130 km/h, side avoidance, rear cross-traffic braking, and incapacitated-driver assistance are increasingly part of the brand story, not hidden in spec sheets.

3. Automotive technology is becoming multi-sector infrastructure

Renault’s defense collaboration and AgiBot’s robotics breakthrough both demonstrate that what the EV industry builds today, from power systems to perception stacks, may underpin entirely different markets tomorrow.

4. China’s ecosystem advantage is widening

From Huawei’s cloud-trained ADAS to full-size humanoid robotics, China is demonstrating a level of vertical integration and iteration speed that global competitors will find difficult to ignore.

Competitive Context: Deepal’s Position in China’s EV Market

Deepal, backed by Changan, is operating in one of the toughest segments in the Chinese market: affordable-to-midrange SUVs where buyers now expect electrification, strong digital features, and credible intelligent driving.

The S07’s strategy is straightforward:

• use EREV to reduce range anxiety
• use Huawei branding to boost smart-driving credibility
• use active safety to build trust
• use pricing and subsidies to lower adoption barriers

The RMB 6,000 subsidy for the higher-level ADS package is especially important. In China’s EV market, adoption often depends less on whether a feature exists and more on whether customers feel the upgrade cost is justified in a crowded field of alternatives.

The Next Step

The key question now is not whether Chinese automakers can launch increasingly capable ADAS systems. It is whether they can deliver consistently safe, low-friction real-world performance at scale, while managing customer expectations and regulatory scrutiny.

For Deepal, the new S07 EREV looks like a serious attempt to bring Huawei-backed intelligent driving and active safety to a broader audience. For the wider industry, the week’s headlines suggest something even bigger: the future of EV competition will be defined not only by batteries and range, but by who can best combine sensing, software, safety, and AI across multiple mobility domains.