NextFin News - As the 2025 fiscal year concludes, the competitive landscape of China’s intelligent driving sector has reached a definitive turning point, characterized by a deepening "Matthew Effect" where market leaders are consolidating their grip on both high-end and mass-market segments. According to the latest data from the Gaogong Industry Institute (GGII), a prominent Chinese market research firm specializing in artificial intelligence and robotics, the market for advanced driving assistance systems (ADAS) and urban Navigate on Autopilot (NOA) has become a battleground for two primary titans: the American semiconductor giant Nvidia and the Chinese homegrown leader Horizon Robotics.
The GGII report reveals that in 2025, the total number of vehicles delivered in the Chinese market equipped with factory-installed front cameras for ADAS reached 8.92 million units, representing a penetration rate of 38.83%. In this high-volume segment, Horizon Robotics emerged as the dominant force, securing a 47.66% market share. This performance places the company significantly ahead of its nearest competitor, Mobileye, which held 27.81%. Together, these two entities control over 75% of the front-view integrated camera and small-domain control computing chip market, leaving legacy players like Renesas and Xilinx with shares of 10.90% and 4.28%, respectively.
Simultaneously, the more technologically demanding urban NOA segment witnessed an explosive growth of 155.83% year-on-year, with 2.07 million units delivered. In this arena, Nvidia remains the undisputed leader with a 49.36% market share, followed by Huawei at 23.07% and Horizon Robotics at 17.88%. The concentration in this high-end tier is even more pronounced, with the top three players collectively commanding 90% of the market. This data underscores a critical shift: while automakers like Nio and BYD are exploring self-developed platforms, the actual adoption rate of in-house computing hardware in mass-produced vehicles remained below 5% in 2025, reinforcing the industry's reliance on third-party chip specialists.
The divergence in market leadership between Nvidia and Horizon Robotics reflects a broader "K-shaped" differentiation in the Chinese automotive market. Nvidia’s dominance in the urban NOA segment is a direct result of its high-performance Orin-X and the newly introduced Thor platforms. These chips provide the massive computing power—often exceeding 500 TOPS—required for the "end-to-end" neural network architectures that became the industry standard in 2025. As U.S. President Trump’s administration continues to navigate complex trade dynamics, the reliance of Chinese premium EV brands on Nvidia’s high-end silicon remains a strategic vulnerability, yet a technical necessity for maintaining global competitiveness in autonomous driving software.
Conversely, Horizon Robotics has successfully executed a "democratization" strategy. By focusing on the Journey 6 series, which offers a scalable range of computing power from 10 to 560 TOPS, the company has enabled automakers to bring advanced driving features to the 100,000 to 150,000 RMB price segment. Founder Yu Kai has emphasized that the true value of technology lies in widespread accessibility. This approach has allowed Horizon to capture the vast middle-market volume that Nvidia’s high-cost solutions often miss. The company’s ability to offer integrated hardware-software solutions—the "model room" approach—has significantly reduced the time-to-market for domestic Tier 1 suppliers like NavInfo and Continental.
Looking toward 2026, the industry is bracing for a shift from technological breakthroughs to a brutal price-performance war. The entry of Momenta into the chip space with its self-developed "BMC" chip, expected to mass-produce this year, poses a direct threat to Horizon’s cost leadership. Furthermore, as the "end-to-end" paradigm matures, the bottleneck is shifting from vehicle-side inference to cloud-side training. This favors players with deep pockets for cloud computing infrastructure. While Nvidia holds the advantage in global cloud ecosystems, Horizon’s "HSD Together" model aims to build a localized ecosystem that can withstand external shocks and lower the entry barrier for smaller automakers.
The analytical conclusion is clear: the Chinese intelligent driving market has moved past the experimental phase. The 2025 rankings suggest that the future will be defined by two distinct paths: the pursuit of absolute cognitive intelligence led by Nvidia’s high-power silicon, and the rapid popularization of reliable ADAS led by Horizon’s cost-efficient architecture. For investors and industry observers, the key metric for 2026 will not just be TOPS (Tera Operations Per Second), but the efficiency of "computing power per dollar," as the industry seeks to make urban NOA a standard feature rather than a luxury option.
Explore more exclusive insights at nextfin.ai.
