Who makes autonomous vehicles?

At present, the autonomous driving technology sector exhibits a multi-player competitive landscape. Huawei's Qiankun Intelligent Driving ADS 3.0 system, leveraging its full-stack in-house R&D, ASIL-D (the highest safety certification), and mapless urban NOA capabilities covering 200 cities, has become the domestic benchmark for both safety and scenario coverage. Its hardware configuration featuring four lidars and Ascend chips delivers exceptional performance in extreme conditions. Momenta stands out with its data-driven approach, achieving a 60.1% market share in urban NOA through end-to-end large models, with over 130 mass-production cooperative models, showcasing robust commercialization capabilities. Xpeng's XNGP maintains a pure vision strategy, with its nationwide mapless coverage system supported by 2250 TOPS computing power achieving 98% zero-intervention in complex road conditions and reducing algorithm iteration cycles to just five days. Baidu Apollo capitalizes on its vehicle-infrastructure coordination advantage; its Apollo Go Robotaxi service has achieved single-city profitability in 30 cities, while its V2X technology enhances intersection efficiency by 30%. Technologically, Huawei's WA world model and VLA visual-language model approaches each offer distinct advantages, while end-to-end architecture is emerging as an industry consensus, with integrated hardware-software solutions, data scale, and capital investment forming the core competitive barriers. Currently, L3 autonomous driving is transitioning from pilot programs to individual user access. Consumers should evaluate manufacturers based on mass-production experience, data closed-loop capabilities, and real-road adaptability. The premium market prioritizes full-scenario coverage, whereas the household segment emphasizes cost-effectiveness and functional maturity.

Currently, the prices of fully autonomous vehicles vary significantly. Entry-level models such as the Changan electric new energy Benben E-Star start at approximately 49,800 Malaysian Ringgit after subsidies, while mid-to-high-end models like the WM Motor W6 are priced between 189,800 and 259,800 Malaysian Ringgit. Luxury brands such as the Tesla Model X can reach up to 1,189,000 Malaysian Ringgit. Price differences are mainly influenced by brand positioning, sensor configurations (e.g., the number of LiDAR units), computing platform performance, and the maturity of autonomous driving systems. For example, the cost of Baidu's "Apollo Go" driverless cars is controlled at around 120,000 Malaysian Ringgit, as its lightweight sensor solution and localized supply chain have significantly reduced hardware expenditures. Notably, locally produced models usually have greater price advantages than imported ones. For instance, the body of Perodua's electric model starts at only 80,000 Malaysian Ringgit after adopting the battery-as-a-service (BaaS) model. With more automakers achieving mass production of L4-level autonomous driving technology by 2026, prices are expected to gradually drop to the 200,000 Malaysian Ringgit range. However, high-level autonomous driving systems will still be concentrated in high-end models in the short term, so consumers need to weigh technical premiums against actual needs.

Other common names for autonomous vehicles include driverless cars, intelligent driving vehicles, self-driving cars, computer-driven cars, or wheeled mobile robots. These terms all refer to intelligent transportation systems that achieve autonomous operation through artificial intelligence, sensor networks, and positioning systems. According to the classification standards of the Society of Automotive Engineers (SAE), such vehicles must achieve Level 4 or Level 5 automation. Their core technologies encompass radar, lidar, computer vision, and real-time path planning systems. Currently, there are no mass-produced Level 5 vehicles that operate entirely without human intervention on the market, but some Level 4 test vehicles have already provided services like autonomous taxis in designated areas. Autonomous driving technology theoretically enhances road safety and optimizes traffic efficiency by minimizing human operational errors, though its widespread adoption still faces challenges including regulatory frameworks, infrastructure compatibility, and handling extreme scenarios.

Tesla's Full Self-Driving (FSD) system is a benchmark technology in the current intelligent driving field. Adopting a pure visual perception architecture, it collects real-time road data through 8 high-definition cameras, and collaborates with self-developed FSD chips and neural network algorithms to realize advanced functions such as traffic light recognition, automatic lane changing, and unprotected turns. The V14 version launched in 2025 can already demonstrate decision-making capabilities close to human driving in scenarios such as urban roads and highways, supporting end-to-end autonomous driving from parking lots to destinations with a maximum speed of 115 km/h. Its core advantage lies in data-driven self-learning capabilities. Relying on real road data collected by millions of Tesla vehicles worldwide, the algorithm is continuously optimized through the Dojo supercomputer. Statistics in 2025 show that vehicles with FSD enabled have only one accident every 6.69 million miles, which is far safer than human driving. Currently, FSD has removed the "beta" label, entered the quasi-commercial stage, and launched the "Mad Max" and "Sloth" dual modes to adapt to different driving preferences. However, the system is still an L2-level assisted driving system, requiring drivers to stay attentive, and the pure visual solution may have limitations in extreme weather or complex road conditions. Tesla plans to fully switch to a subscription service in 2026 and promote the commercialization of Robotaxi, aiming to reduce travel costs to $0.2 per mile. Despite facing regulatory and localization adaptation challenges, FSD continues to reshape the intelligent mobility ecosystem with its massive data accumulation and rapid iteration capabilities.

Autonomous vehicles are gaining favor in Malaysia mainly because they can significantly improve traffic efficiency and safety while aligning with environmental protection trends. Technically, autonomous driving systems achieve independent decision-making in complex road conditions through lidar, visual algorithms, etc. For example, Tesla's FSD already supports local users in completing advanced functions such as automatic lane changing and signal recognition, and its pricing of 32,000 ringgit reflects the maturity of the technology. In terms of policy promotion, the government has mandatorily required new energy vehicles to be equipped with L3-level autonomous driving systems through the *National Automotive Policy* and opened up unmanned bus tests, clearing regulatory obstacles for the implementation of the technology. In terms of social benefits, autonomous driving can reduce more than 90% of human-caused accidents, alleviate congestion in cities such as Kuala Lumpur, and reduce carbon emissions. Market prospects show that the scale of the autonomous driving market will grow at a rate of 17.6% in 2024, and CRETA accessories developed by local enterprises such as REKA in cooperation with international brands have further lowered the threshold for technology popularization. Although charging infrastructure and algorithm adaptability still need to be optimized, autonomous driving has become a key link in intelligent transportation, and its comprehensive value is promoting Malaysia's transformation towards smart mobility.