After years of anticipation, Musk's vision of a "self-driving taxi" has finally emerged from concept to reality—Tesla's Robotaxi has quietly embarked on small-scale road tests in Austin, Texas, ushering L4 autonomous driving from PowerPoint presentations onto real streets for the first time.

This bold move, hailed as a "technological leap milestone" by the industry, underscores a cautious approach during the engineering validation phase: limited areas, small-scale fleets, and debates over the reliability of the pure vision FSD system make the "disruptive moment" anticipated by the capital market seem more like a controlled risk stress test.

For A-share investors, Tesla's breakthrough serves as both a catalyst for the intelligent driving industry chain and a mirror reflecting the truth—the three major challenges of technological maturity, regulatory gaps, and ethical controversies may expose funds blindly chasing the "Robotaxi concept" to valuation traps. As global giants surge ahead in autonomous driving, can China's supply chain capitalize on this momentum to breakthrough?

01. The Future of Autonomous Driving

The autonomous driving industry stands at the cusp of large-scale commercialization, with technological iteration and market validation creating a symbiotic effect.

Industry monitoring data reveals that the operating metrics of global Robotaxi leaders have shown exponential growth since the second half of 2024: Waymo's weekly order volume in Los Angeles has surpassed 50,000, a 320% increase from the beginning of the year; Cruise has partnered with the world's largest taxi company, aiming to deploy 25,000 autonomous driving fleets across 15 cities by 2025. This transition from quantitative to qualitative change signifies that the industry has moved beyond the technological verification phase and officially entered a new era of commercial scale expansion.

However, Tesla's initial Robotaxi test in Austin has tempered this optimism—what Musk touts as a "transformation engine" is, in reality, a technologically constrained experiment.

The divergence and integration of technological evolution paths are reshaping the industry landscape. Unlike L2+ assisted driving, which relies on human backup, L4 Robotaxi must establish a fully autonomous decision-making system.

Tesla's test uses the Model Y, equipped with the FSD V12 version and HW4.0 hardware, ostensibly demonstrating urban driving capabilities but revealing significant limitations in actual operations: the geofenced area is strictly limited to a 50-square-kilometer region in southern Austin, operating hours are confined to 6 am to midnight, and passengers must participate via an invitation-only mechanism. Notably, despite eliminating the traditional driver, a safety officer remains in the front passenger seat, and the backend monitoring system has remote takeover capabilities, indicating that its level of automation still falls short of L4 standards.

In terms of perception solutions, the industry presents a differentiated layout amidst converging technological paths. Tesla adheres to a pure vision approach, achieving spatial semantic understanding through the BEV+Transformer architecture, with the FSD V12 version boasting a 98% processing rate in normal road conditions but still experiencing a 12% recognition error rate in extreme scenarios like heavy rain and night vision.

In contrast, Huawei's 96-line hybrid solid-state radar maintains a ranging accuracy of 3 cm at a distance of 200 meters, but multi-sensor fusion introduces a system response delay of 35 milliseconds. DJI Automotive's binocular vision + 4D millimeter-wave radar solution offers 95% L4 scenario coverage while reducing costs by 60%, making this compromise solution a potential mainstream choice for mid-range models. Tesla's test underscores the practical dilemma of the debate over technological paths—pure vision solutions excel on structured roads but their reliability in complex scenarios still requires real-world testing.

Technological cost reduction and scenario breakthroughs form a dual-wheel drive for large-scale implementation. Baidu Apollo's sixth-generation system has reduced hardware costs by 82% compared to 2020 through chip localization, reaching the commercialization tipping point of 480,000 yuan per set. WeRide's test data from Guangzhou's Bio Island indicates that when the daily operating mileage per vehicle exceeds 350 kilometers, the overall cost is on par with traditional ride-hailing services.

However, Tesla Robotaxi's flat-rate pricing strategy of $4.2 per order still faces the risk of cost inversion under the current safety officer configuration. More pressing is the regulatory challenge: Texas' new regulations require autonomous vehicles to possess risk minimization capabilities, recording devices, and state approval, and Tesla's insistence on its original launch plan essentially pits technological verification against policy compliance.

Looking ahead from the 2025 perspective, the large-scale implementation of Robotaxi is not a breakthrough in a single technological dimension but a systematic project encompassing data closed loops, algorithm evolution, hardware cost reduction, and scenario adaptation. As industry leaders begin validating technical feasibility with real orders and using operational data to inform system iterations, this technology-driven transportation revolution is transitioning from the laboratory to the streets.

Tesla's "cautious test run" stands in stark contrast to Waymo's "heavy-duty push," with the former prioritizing generalization capabilities and marginal cost advantages, and the latter constructing safety redundancies through lidar and high-precision maps. The competition and cooperation between these two paths may determine the ultimate form of autonomous driving commercialization.

02. China's Supply Chain Poised for Breakthrough

Compared to the U.S. market, China's autonomous driving industry chain is also embracing a historic window of opportunity for breakthroughs. Amidst Tesla's global Robotaxi rollout being constrained by legislative thresholds and safety costs, Chinese supply chain enterprises have demonstrated unique competitiveness through technological accumulation and policy adaptation advantages.

The Ministry of Industry and Information Technology's new regulations on intelligent driving supervision issued in April mandate automakers to clarify functional boundaries and eliminate false advertising, compelling the industry chain to accelerate technological standardization. Tesla's China official website has renamed the FSD system from "Full Self-Driving" to "Intelligent Assisted Driving," a change reflecting the realistic path for technology implementation amidst tightening regulations—accumulating data through progressive functional rollouts rather than overpromising full self-driving capabilities.

In the realm of core components, local enterprises have achieved multi-dimensional breakthroughs. Top Group, Yinlun Co., Ltd., and other T-chain suppliers participate in Tesla's supply chain, with their lightweight chassis and thermal management systems passing market verification. NavInfo, as the sole domestic solution provider integrating "maps + algorithms + chips," boasts high-precision maps covering 320,000 kilometers of highways nationwide, with an hourly update frequency, meeting the demand for dynamic geographic information for L4 autonomous driving. RoboSense pioneered mass production of automotive-grade lidar, with its M1 platform products achieving a point cloud density of 300,000 points per second, triple that of the previous generation, and has secured designated orders from automakers like XPeng and Xiaomi, with costs controlled within $500, breaking the price monopoly of foreign manufacturers.

In the domain of algorithms and computing platforms, Horizon Robotics' Journey 6 chip, based on the BPU architecture, boasts a computing power of 560 TOPS, supporting the BEV+Transformer spatial perception algorithm, achieving a scenario coverage rate of 98% in actual tests on Lixiang One's L9 model. HiRain Technologies' 4D millimeter-wave radar has surpassed traditional technical barriers, enabling commercial operation without safety officers in Qingdao Port. Zhixing Technology's front-fusion perception algorithm has achieved the industry's first mass production on Geely's Zeekr 001, meeting L4 safety redundancy requirements.

The collaborative innovation between vehicle enterprises and supply chains is reshaping the industrial landscape. The SAIC Motor and Momenta joint venture Zhiji Automobile is equipped with an L3 autonomous driving system, enabling point-to-point automatic pickups in specific areas of Shanghai. The user penetration rate of XPeng's XNGP system reaches 93.3%, and its AI valet parking function has completed over 100,000 unassisted stops in Guangzhou's Tianhe business district, verifying the reliability of high-precision positioning algorithms in complex scenarios. This "OEM defining requirements + supply chain rapid iteration" model enables Chinese enterprises to gain a first-mover advantage in the commercialization of Robotaxi.

While the global autonomous driving industry grapples with legislative compliance and safety costs, China's supply chain has forged a closed-loop ecosystem of "policy norms - technological innovation - scenario verification." From the concurrent surge in lidar quantity and price to the computing power breakthroughs of AI chips, to the real-time updates of high-precision maps, local enterprises are transforming policy dividends into market advantages through technological cost reduction and scenario deep diving. This industrial revolution sparked by Robotaxi may witness the transformation of China's supply chain from follower to leader.

- End -