Value clues in the interim report of Horizon (09660): The "Arm + Android" model opens up long-term growth space.
As the platform model enters a period of deep realization, the release of scale effects and operational leverage will inevitably accelerate the conversion of Horizon's ongoing investments into a long-term competitive advantage.
The long-term growth logic of Horizon Siasun Robot & Automation (09660) is becoming increasingly clear as the company evolves its growth approach.
According to the companys recently disclosed interim results for the fiscal year 2026, Horizon still achieved counter-cyclical growth in key financial indicators despite a challenging market environment in the first half of 2026. During this period, the revenue from customer contracts in the company's continuing operations increased by 32.9% year-on-year to 2.055 billion yuan, while gross profit from continuing operations also rose by 32.9% year-on-year to 1.356 billion yuan, yielding a gross margin of 66%. Behind these impressive aggregate figures lies the deep realization of Horizons unique "ARM + Android" open platform model. In the first half of the year, revenue from licensing and services amounted to 1.129 billion yuan, representing a year-on-year increase of 52.7%, accounting for 55.0% of total revenue; during the same period, the gross margin for this segment was 90.4%, remaining at a high level.
The highly profitable licensing business once again demonstrated its growth potential, which can be traced back to Horizons platform model possessing a unique competitive advantage. From the perspective of the company's overall commercial logic, Horizons platform model is comparable to Wintel, i.e., a standard product platform, with the core being a set of mature products entering more terminals, thereby realizing platform value through scaled shipments and software licensing; focusing on the licensing and services business, its underlying logic is similar to a composite model of "ARM + Android," which is a foundational technology licensing platform, where the core lies in a set of foundational technologies giving rise to multiple products. Customers can independently develop chips, systems, and terminal products based on Horizons capabilities, and Horizon, as a platform, expands technical boundaries through licensing and ecology, thereby supporting and promoting healthy industry development.
These two models correspond to two distinct market spaces: Wintel corresponds to "direct market share," referring to the market share held by products, while "ARM + Android" relates to "extended market share," which not only brings substantial IP licensing revenue but also helps build a long-term and sticky partnership between Horizon and its customers. Horizons management anticipates that by next year, the combined "direct + extended" market share of its chips and those self-developed chips by vehicle manufacturers supported by Horizon is expected to become the top in the high-end market. These two models complement each other, enabling Horizon to continuously evolve on its own and gradually grow into the foundational base for automotive intelligence in the era of smart vehicles.
ADAS share exceeds 50% for the first time; high-end market will leap to first place next year.
Amid a significant decline in domestic passenger car retail sales in the first half of 2026, Horizon has still shown strong growth resilience during this deep adjustment period in the industry. GMTEight noted that the dual-engine structure driving Horizons high-quality growth exhibits a more balanced development trend: During this period, the companys products and solutions business benefited from its continuously rising market share, with shipments of the Journey series hardware increasing by 12.1% year-on-year to 2.218 million sets, driving a revenue increase of 14.8% year-on-year to 926 million yuan; meanwhile, the licensing and services business further realized growth momentum through its "ARM + Android" style foundational technology licensing model.
In the past, the capital market viewed Horizon more as a smart driving technology leader, but with the deep realization of Horizon's platform model, the company's more accurate positioning is the foundational base for automotive intelligence. From being a technological leader to becoming something akin to infrastructure, this transition is backed by millions of units validated through mass production. In the first half of the year, despite pressures in the auto market, Horizon achieved a market share of 50.0% in its own brand ADAS, setting a historical high; its share of full-stage smart driving chips reached 31.9%, firmly holding the top market position; while its share of smart driving chips with urban area NOA functionality (high-end) reached 22.8%, rising to second in the market.
Currently, the cumulative shipments of Horizons Journey series chips have surpassed 10 million sets, creating a scale effect where one in every three smart cars in China is equipped with Horizon chips. In terms of new developments, the J6M, which only entered the largest new energy vehicle manufacturer in China last year, is expected to account for 70-80% of its entry-level smart driving solution shipments by the end of this year. The high-performance chip and HSD have also been awarded the high-end platform designation by this customer, with mass production delivery set to begin within the year, further solidifying the scale foundation of the platform model.
Moreover, Horizon's latest partnership with the Volkswagen Group, announced recently, can be seen as a model case of Horizons platform model. In late July, Horizon and Volkswagen Group announced they would deepen their technological cooperation in the AI foundation large model field through CARIAD and their joint venture CoolCore (CARIZON). CoolCore will leverage Horizons advanced AI foundation large model capabilities to independently develop and accelerate the construction of a unified AI driving solution for Volkswagen Group in China. The essence of this collaboration lies in the white-box licensing modelCoolCore is independently developing full-link smart driving algorithms encompassing perception, decision-making, and control based on Horizons AI foundation large model, which is the best annotation of the ARM + Android style foundational technology licensing model.
It is reported that CoolCores full-scenario advanced driver assistance solution has entered the mass production stage and will gradually be equipped on seven new electrified models from three joint ventures of the Volkswagen Group in China starting in the third quarter of this year; according to the plan, the L3 level autonomous driving capabilities of the Volkswagen Group is expected to commence delivery in the second half of 2027 at the earliest. Data shows that Volkswagen, Toyota, and Honda together account for 50% of the sales of all joint venture car companies, and through partnerships with CoolCore, Denso, Autoliv, and others, Horizon has opened up this growth space. Transitioning from product collaboration to AI foundational technology cooperation, Volkswagen casts the most substantial vote for Horizons platform model through large-scale validation at a global level.
For Horizon, which is committed to becoming the "foundational base for automotive intelligence," having more ecosystem partners and a richer application will enhance the value of its own platform. So far, based on its long-term adherence to an open Tier 2 positioning, Horizon has connected with over 200 ecosystem partners such as Bosch, Denso, ZF, and Continental Group, forming a vast industrial ecosystem alliance.
Persisting in saturated research and development investment, opening up certain growth spaces.
The mid-term report data has already verified the realization capability of the platform model, but this is just the beginning; the ecological positioning of the "foundational base for automotive intelligence" will help Horizon unlock broader and more certain growth spaces in the future.
Horizons certainty comes primarily from the "technical ecological stickiness" generated by its platform model. At present, Horizon's platform capabilities have penetrated into chips, algorithms, compilers, toolchains, testing, and mass production processes. Once car manufacturers establish their R&D systems around Horizon, switching foundational platforms means a complete redo of algorithm migration, code adaptation, and test validation, leading to extremely high switching costs. For example, in the case of granting licenses for automotive chips (BPU), smart driving algorithms and BPUs are tightly coupled; car manufacturers that adopt BPU for self-developed chips subsequently have better natural compatibility with HSD and toolchains, deepening customer stickiness with each generation of products.
It should be noted that the value that Horizon as the "foundational base for automotive intelligence" brings to car manufacturers is comprehensive. From a cost perspective, under the HSD Together model, the human resources input and computing power consumption in vehicle product development can be significantly reduced; from an efficiency standpoint, Horizons mature chips, algorithms, and toolchains can dramatically shorten the mass production cycle; more importantly, once foundational technology is standardized, car manufacturers can concentrate more resources on scene functionality, user experience, and brand definition, truly achieving standing on the horizon to innovate.
From an industry perspective, the first half of smart driving is about "making smart driving a reality," while the second half is about "scaling smart driving." As high-end smart driving moves from flagship models priced above 300,000 yuan down to family cars priced around 100,000 yuan, and as urban area NOA begins to become a standard feature on more models, the industry will inevitably transition from "every automaker reinventing the wheel" to a division of labor model where "professional platforms develop once and are reused multiple times." A crucial reason why Horizon can become an irreplaceable "foundational base for automotive intelligence" is that the company maintains a long-term high-intensity R&D investment; data indicates that the companys R&D expenditure reached 2.755 billion yuan in the first half of this year. First-mover advantages, scale advantages, and continuous saturated R&D investment ensure that Horizon firmly occupies a critical foundational infrastructure position in the era of smart vehicles.
In preparation for the layout of next-generation products targeting L3/L4, the research and development of the Journey 7 chip is progressing smoothly, with expectations to launch in the market by 2027, and proactive engagement from several leading car manufacturers and Tier 1 suppliers has been secured. Additionally, the company plans to collaborate with a leading technology platform to launch L4-level Robotaxi test projects in several cities within this year, further validating the potential application of foundational technologies in high-end autonomous driving scenarios. In broader fields, Horizon has empowered over 100 downstream categories of Siasun Robot & Automation and more than 400 clients through its joint venture Digua Siasun Robot & Automation, covering more than half of the embodiment intelligent enterprises in China. From automobiles to Siasun Robot & Automation, the foundational capabilities of the platform model are replicating into the wider physical world.
Finally, upon reviewing Horizons mid-term report, there is a value point that is easily overlooked: during the period, the company's operating expenses grew by 23.3% year-on-year, a significantly lower increase than that of revenue and gross profit, indicating that Horizons operational leverage effect is steadily being released at this stage. Looking ahead, as the platform model enters a deep realization stage, the release of scale effects and operational leverage will inevitably accelerate the transformation of Horizons long-term investments into competitive advantages. In other words, regardless of how the subsequent market environment unfolds, Horizons growth trajectory is expected to remain stable, and conversely, the certainty of the companys investments will be further strengthened as the industry enters the stage of "platform competition."
Related Articles

MNSO (09896) invested $251,600 to buy back 107,600 shares on August 31.

NIO-SW (09866) delivered 35,836 vehicles in August, a year-on-year increase of 14.5%.

XPENG-W (09868) delivered a total of 39,107 new cars in August, a year-on-year increase of 4%.
MNSO (09896) invested $251,600 to buy back 107,600 shares on August 31.

NIO-SW (09866) delivered 35,836 vehicles in August, a year-on-year increase of 14.5%.

XPENG-W (09868) delivered a total of 39,107 new cars in August, a year-on-year increase of 4%.

RECOMMEND





