Market share surpasses 50% for the first time; from leading in share to monetizing the ecosystem, Horizonrobot's foundational value is being re-evaluated.

In the first half of 2026, domestic passenger car retail sales fell 20.2% year-on-year, putting pressure on total industry volume. However, intelligent driving emerged as a rare sector with structural growth, with its penetration rate rising against the trend to 76.1%, an increase of 8.5 percentage points compared to the full previous year. $HORIZONROBOT-W (09660.HK)$

Against this backdrop, leading intelligent driving companies have become the focus of capital market attention. On August 31, Horizonrobot released its interim results for 2026, reporting year-on-year growth in both revenue and gross profit for the first half of the year, while maintaining a high gross margin, demonstrating resilient growth during an industry adjustment period.

However, what is even more noteworthy is the value leap behind this growth. Today, Horizonrobot not only possesses a full range of products covering vehicle-level AI agents but has also built a foundational technology platform, completing its transition from a chip supplier to an industrial ecosystem platform. The company's growth logic and valuation framework are undergoing profound changes.

Delivering high-quality results against the headwinds, achieving value-driven growth on top of leading scale.

When assessing growth logic, performance during industry downturns is more convincing than during upcycles. Growth in favorable conditions can often be seen as 'a rising tide lifting all boats,' whereas growth in adverse conditions better reflects a company's core competitiveness.

In the first half of 2026, Horizonrobot achieved revenue of RMB 2.055 billion, a year-on-year increase of 32.9%; gross profit reached RMB 1.356 billion, also up 32.9% year-on-year, with the overall gross margin maintained at 66%. A breakdown of these results reveals three positive developments:

First, a breakthrough in market share, establishing absolute industry leadership.In the ADAS market for domestic brands, Horizonrobot's share exceeded 50% for the first time, approximately double that of the second-place competitor. In the overall intelligent driving chip market, the company secured the top position with a 31.9% share. The high-end segment is particularly significant: the market share for urban NOA (Navigate on Autopilot) computing platforms jumped from 17.9% to 22.8%, moving the ranking from third to second, surpassing Huawei. Together with NVIDIA, they have become the 'duopoly' in high-end intelligent driving, collectively accounting for over 60% of the market share, essentially forming a two-player dominant structure in the high-end market.

This signifies that Horizonrobot's technical competitiveness and customer recognition have been validated under stress testing. After crossing the 50% threshold, the head-start effect will gradually manifest, with customer resources concentrating and R&D and production costs continuing to be amortized. As the current core stronghold of intelligent driving technology application, urban NOA features high barriers and high per-vehicle value. Breaking into the top two in this sector will further highlight the company's product premium capability.

Second, optimization of revenue structure, forming dual engines for counter-cyclical growth.Currently, Horizonrobot's revenue comes from two major business segments: Revenue from products and solutions amounted to RMB 926 million, a year-on-year increase of 14.8%; revenue from licensing and services reached RMB 1.129 billion, a year-on-year increase of 52.7%, raising its proportion of total revenue to 55.0%.

The former corresponds to the shipment of products such as autonomous driving chips and solutions, serving as the foundation for scale growth; the latter covers algorithms, IP, toolchains, and technical services, acting as a source of high-margin profits.

The value of this dual-engine structure lies in risk hedging and mutual empowerment. Higher product shipments drive greater demand for accompanying algorithm licenses, toolchain services, and iterative upgrades, thereby boosting licensing business growth. Conversely, a mature algorithm foundation and toolchain reduce development complexity for clients and shorten mass production cycles, attracting more customers to procure hardware solutions. This is the core logic behind the company's ability to maintain high overall gross margins during industry downturns.

Thirdly, improved R&D efficiency and capability reuse leverage operating leverage.In the first half of the year, the company maintained intensive R&D investment, totaling RMB 2.755 billion, focusing on general underlying technologies covering chips, AI models, and toolchains to enable capability reuse.

As the number of customers and shipment scale increase, the output per unit of R&D cost continues to rise, gradually revealing the effects of operating leverage. The 23.3% year-on-year growth in operating expenses, which is lower than the growth rates of revenue and gross profit, directly validates this trend.

Overall, Horizonrobot's counter-trend growth in the first half was driven by increased market share, optimized revenue structure, and improved operational efficiency. This quality of growth is superior to mere scale expansion, laying a solid foundation for sustained future growth.

Three growth curves叠加: diversifying risks and raising the growth ceiling

Where will future growth momentum come from? In the past, Horizonrobot's growth relied mainly on domestic independent brands. The signals released by this performance indicate that its growth boundaries are comprehensively expanding.

The first growth curve is the upgrade toward advanced autonomous driving.Demand for autonomous driving features is shifting from 'availability' to 'quality.' The popularity dividend of basic ADAS is gradually peaking, while high-end NOA is taking over as the new growth engine. Data shows that in the first half of the year, the proportion of assisted driving vehicles equipped with NOA functions rose to 46.8%, with urban NOA accounting for 64.1% of assisted driving capabilities.

Horizonrobot has seized the window of opportunity presented by this round of industrial upgrading, rapidly increasing its market share in urban NOA (Navigate on Autopilot). As products such as the Journey 6P chip and the Horizon SuperDrive (HSD) full-scenario advanced driver-assistance system achieve scaled deployment, the company's share of the high-end market and the value per vehicle are expected to rise in tandem, serving as a key growth driver.

The second growth curve is breaking into the joint venture market.Joint venture brands account for approximately 35% of the domestic market share. While there is an urgent demand for intelligent upgrades, the supply chain barriers are extremely high, with core solutions previously monopolized by overseas manufacturers.

By collaborating deeply with top-tier global Tier-1 suppliers, Horizonrobot has systematically entered the supply chains of joint venture automakers. A cooperative ecosystem comprising global Tier-1 suppliers such as Hitachi Astemo, Bosch, and Denso is gradually taking shape. Meanwhile, the company has established deep ties with leading automakers through joint venture platforms: it formed the autonomous driving venture neueHCT in partnership with Amovio Group, and established CoreAuto with CARIAD, a subsidiary of Volkswagen Group.

During the reporting period, the company achieved two landmark breakthroughs in the joint venture market: First, the full-scenario advanced driver-assistance solution, developed with CoreAuto based on the Journey 6 series and HSD, has entered mass production and will be deployed in seven new electric vehicle models across Volkswagen's three joint ventures in China. Second, the Journey 6B chip achieved its global mass-production debut at GAC Toyota, passing the stringent quality system validation of Japanese automakers.

Securing a core position in the supply chains of these two giants holds significance far beyond the orders themselves. As the world's two largest automakers, Volkswagen and Toyota have extremely high supply chain certification thresholds; gaining entry signifies that technical capabilities have been recognized against the industry's highest standards, which is likely to encourage other joint venture brands to follow suit. Opening up the joint venture market effectively creates an incremental market of comparable size, providing ample room for Horizonrobot's medium-term growth.

The third growth curve is strategic positioning in overseas markets.In the first half of 2026, China's complete vehicle exports grew by 65.3% year-on-year, with the top six automotive groups accounting for over 70% of the total share, accelerating globalization. As a core differentiating feature, intelligent driving is a key competitive advantage for Chinese automakers going global, yet there are only a handful of intelligent driving suppliers capable of supporting such international expansion.

Horizonrobot has already covered models from all six leading export-oriented automakers, securing design wins for over 60 export models cumulatively, many of which feature mid-to-high-end intelligent driving hardware platforms, thereby deeply aligning with the wave of Chinese automakers' global expansion. Additionally, the company exports technology through its global Tier-1 partners; its front-view integrated unit based on the Journey 6B has secured a design win for a mainstream German automaker's global platform, with vehicles sold to markets including Australia, New Zealand, and India.

To date, the Journey 6B has secured mass-production design wins from over 25 mainstream automakers globally, with lifetime design win volume exceeding 20 million units, widely validating its global adaptation capabilities.

Unlike direct international expansion, Horizonrobot secures its position in overseas markets by partnering with global Tier-1 suppliers and leading Chinese automakers. This strategy not only circumvents direct competitive barriers but also allows the company to share in the dividends of the global upgrade in intelligent driving, representing a more prudent and scalable path for globalization.

Shifting from a single domestic market to a triple-pronged strategy encompassing high-end segmentation, joint ventures, and overseas expansion, Horizonrobot has diversified its customer base and continuously strengthened its risk resilience. Consequently, the company's mid-term growth ceiling has been significantly raised.

The technological foundation of the Physical AI era: A long-term value reassessment from automobiles to robotics.

From a value investment perspective, the growth model is more fundamental than the growth rate. If a company merely delivers product solutions following industry cycles, it remains essentially a cyclical growth stock, with valuations fluctuating alongside industry sentiment. However, if its growth is underpinned by reusable technology platforms and ecosystems, it qualifies as a platform company, offering greater certainty in growth and a higher valuation中枢 (central tendency).

Reviewing the evolution of the global technology industry, mature sectors invariably form a division of labor based on underlying platforms—a repeatedly verified pattern. The PC era saw the formation of the Wintel model (Intel + Microsoft), while the smartphone era established an architecture of chip manufacturers plus mobile operating systems. The semiconductor industry developed a specialized division of labor involving IP licensing, chip design, and wafer fabrication. The universality and reusability of underlying technologies make specialized division of labor an inevitable trend. Companies that master these underlying platforms often possess the highest industrial value and the strongest moats.

The intelligent driving industry is at a critical juncture, transitioning from differentiated configurations to foundational capabilities. Early-stage intelligent driving served as a differentiator for automakers, with each developing proprietary solutions or sourcing discretely. As intelligence levels rise, system complexity increases exponentially, driving up R&D investments in underlying chips, AI models, operating systems, and toolchains. It is economically inefficient for every automaker to redundantly build a complete technology stack. At this point, specialized underlying platforms emerge, achieving economies of scale by amortizing R&D costs across more customers and larger shipment volumes through one-time development and multiple reuse instances.

Horizonrobot's platform model aligns with this trend of industrial division of labor, centered on a dual-platform architecture:

On one hand, Horizonrobot is building a Wintel-like model.At the hardware level, the Journey and Skyline series of automotive-grade chips provide a universal computing power foundation, akin to Intel's role in the PC era. At the software level, HSD and KaKaXia serve as the operating systems for next-generation vehicles, delivering value similar to Microsoft Windows. This model follows a proven commercialization path: hardware generates revenue based on shipment volume, while software and algorithm solutions can evolve into subscription-based revenue streams through feature iterations and OTA upgrades.

Platform value is being continuously realized. Financial reports indicate that the company's integrated cockpit-and-driving whole-vehicle intelligence solution is expected to enter mass production in Q4 2026. R&D for the Journey 7 chip is progressing smoothly, achieving multiple technical breakthroughs over the previous generation. It features native optimizations for the on-device deployment of the next-generation large-parameter HSD intelligent driving model and the KaKaXia intelligent cockpit agent model, with tape-out expected to be completed in early Q2 2027.

On the other hand, Horizonrobot has built an "ARM + Android" model.It licenses core underlying technologies—such as the BPU architecture, AI foundation models, and toolchains—to customers and ecosystem partners, while also supporting their independent development at the hardware, algorithm, software, and system levels. The core value of this model lies in ecosystem stickiness: once customers establish their R&D systems based on Horizonrobot’s underlying technology, switching platforms would incur substantial costs for algorithm migration, code adaptation, and testing validation, creating extremely high switching barriers.

The greater value of the platform model lies in the cross-scenario reusability of its underlying technology, which is the key differentiator between platform companies and single-product companies. Automobiles are the first complex intelligent terminals to achieve large-scale commercialization in the era of Physical AI, and they represent the most demanding scenario for perception, decision-making, and control technologies. A BPU architecture, AI foundation models, and toolchains that have passed automotive-grade validation naturally possess the versatility to adapt to other Physical AI scenarios.

Horizonrobot’s path is clear: first, focus on smart vehicles as the primary battlefield to validate its underlying technology platform and business model; then, replicate and extend its automotive-grade foundational capabilities to scenarios such as autonomous logistics, embodied AI, and home intelligent agents.

During the reporting period, Horizonrobot has exported its technology to a diverse range of edge computing platforms—including embodied AI, autonomous logistics vehicles, home intelligent agents, and electric two-wheelers—through partners. For instance, through its strategic partner DiGua, it has delivered foundational capabilities to over 100 robot categories and more than 400 clients, covering over half of China’s embodied AI enterprises. This move breaks through the growth ceiling of the automotive industry, expanding the company’s value boundary from intelligent driving to the broader Physical AI sector, thereby unlocking significant long-term growth potential.

Summary: Resilient in the short term, certain in the medium term, and with long-term growth space opened up by Physical AI.

The adjustment in the auto market during the first half of 2026 serves as an industry stress test, evaluating companies’ risk resilience while highlighting those with genuine core competitiveness.

For Horizonrobot, in the short term, counter-cyclical market share gains and a dual-engine revenue structure have built performance resilience against economic cycles, ensuring high-quality growth. In the medium term, the simultaneous progress in independent high-end development, joint venture penetration, and overseas expansion raises the growth ceiling, offering strong certainty for growth over the next 2–3 years. In the long term, as the platform model matures, the company is upgrading from a smart driving chip supplier to a foundational pillar of the automotive intelligence industry, with the potential to extend into the broader Physical AI market.

The second half of the intelligent driving industry will be a competition of scale and platformization. Whoever can serve more customers and scenarios with a single set of underlying technologies will become the infrastructure of the smart vehicle era. Leveraging its first-mover advantage and ecosystem layout, Horizonrobot has secured a favorable position, with its long-term value gradually being realized as the industry matures.

7337 Views