Changxin Technology Group surged nearly 466% on its Shanghai debut to become China's most valuable listed company3, CaoCao launched driverless robotaxi testing on public roads in Hangzhou1,2, Japanese engineers conceded they cannot close the humanoid-robotics gap with China in the short term6,7, and Chinese cleaning-robot makers secured 70% of the global market8. ANALYSIS Taken together, these developments trace a single strategic arc: China is compounding advantages across every layer of the physical-AI stack — silicon, autonomy software, and embodied robotics — at a pace that is reshaping competitive benchmarks worldwide.
Why it matters
The conventional framing of China's AI ambitions centers on large language models and cloud compute. ◆ But the week's evidence points to a less-discussed and arguably more consequential buildup: the hardware-and-autonomy layer where chips power vehicles that drive themselves and robots that navigate homes and factories. Each strand reinforces the others — DRAM fuels the edge compute that robotaxis and robots require, while scaled autonomous fleets generate the real-world data that sharpens AI models.
The big picture
Changxin Technology Group surged nearly 466% on its debut on Shanghai's STAR Market, closing at 49 yuan per share and reaching a market capitalization of approximately 3.3 trillion yuan, overtaking Industrial and Commercial Bank of China's 2.6 trillion yuan to become the most valuable China-listed company. The Hefei-based chipmaker had raised 57.92 billion yuan ($8.6 billion) in Asia's biggest IPO so far this year, with proceeds earmarked mainly for mass-producing memory wafers. CXMT held a 7.67% share of the global DRAM market in 2025, according to its IPO prospectus. The company swung to an operating profit of 35.43 billion yuan in the first quarter from a loss of 2.83 billion yuan a year earlier.
ANALYSIS That profit reversal and the market's explosive reception signal investor conviction that indigenous memory production is a strategic asset — one that feeds directly into the compute demands of autonomous systems.
On the autonomy front, CaoCao — originally incubated by Geely — launched robotaxi testing without an in-vehicle safety operator on public roads in Hangzhou's Binjiang District on July 27. Vehicles receive ride assignments only when designated stations trigger a request, and every test vehicle is connected to CaoCao's RAS remote safety platform for continuous monitoring, emergency response, and real-time remote assistance. CaoCao received regulatory approval in April to become Hangzhou's operator for such testing.
CaoCao is not alone. China's robotaxi industry is "rapidly pivoting from small-scale pilot programs to mass-produced commercial fleets," as China Daily reported4. SAIC Motor's mobility arm plans to launch a mass-produced robotaxi in 2027, built on an IM Motors electric platform and powered by Momenta's autonomous driving system5. XPeng completed its first internal robotaxi ride by CEO He Xiaopeng in July, just eight months after announcing the project in November 2025, and plans to launch a pilot service in Guangzhou in the third quarter of 2026. Baidu's Apollo Go logged 3.2 million fully autonomous orders in the first quarter, with cumulative global orders surpassing 22 million across 27 cities as of April 2026. Pony.ai's first-quarter revenue jumped 395.4% year-on-year, and its global fleet has exceeded 1,700 vehicles. Goldman Sachs upgraded its 2030 China robotaxi fleet forecast from 632,000 to 705,000 units in April 2026 and projects the fleet will expand to 3.1 million vehicles by 2035.
Between the lines
The robotics picture extends well beyond vehicles. Chinese manufacturers of household cleaning robots now hold 70% of the global market, competing on unique features rather than price alone. Meanwhile, Nikkei xTECH's teardown of Unitree Technology's G1 humanoid robot led Japanese engineers to conclude that it is "probably unrealistic" for Japan to close the gap with China in humanoid robotics in the short term. Engineers expressed amazement at the G1's highly integrated battery, biomimetic ankle linkage structure, and high-performance motors.
ANALYSIS Japan's own data underscores the shift: China accounted for 295,000 industrial robot installations in 2024, or 54% of the global total, and China's domestic manufacturers outsold foreign brands in China's home industrial robot market for the first time in 2024, reaching 57% market share. Japan installed 44,500 industrial robots in 2024, down 4% year over year. The combination of manufacturing scale, domestic market dominance, and now a conceded technical lead in humanoid design marks a structural realignment, not a cyclical fluctuation.
CaoCao's ambitions illustrate how these layers connect: the company launched its RoboX strategy in June 2026 and revealed the Eva Cab, a custom robotaxi built on Geely's SEA architecture, with a target of deploying 100,000 robotaxis and 100,000 autonomous delivery vans across China by 2030. That vision requires edge-compute chips, autonomous software, and robotic-grade hardware — precisely the stack China is now building domestically.
What's next
XPeng's Guangzhou robotaxi pilot is scheduled for the third quarter of 2026. SAIC Motor's mass-produced robotaxi targets a 2027 launch. CXMT's IPO proceeds will flow into memory-wafer production lines that serve the same autonomous and AI workloads. ANALYSIS The question is no longer whether China can build a vertically integrated physical-AI ecosystem, but how quickly the compounding across chips, vehicles, and robots translates into global market-share shifts that competitors find irreversible.