Runxinwei Releases "Domestic Software-Hardware Integrated AI Intelligent Base", Solving Industry Scarcity Problems and Driving Multi-Terminal Intelligent Transformation
On April 25th, during the Beijing Auto Show, Runxinwei held a strategic and product launch event themed "Core Vision · New Pattern". As a rare provider of full-stack solutions integrating "chips + operating systems + AI" in the industry, Runxinwei fully showcased for the first time its "domestic software-hardware integrated AI intelligent base" that can be applied to multi-terminal scenarios. Through the comprehensive presentation of strategy, technology, products, and ecosystem, this launch event fully demonstrated Runxinwei's progress in the field of multi-terminal intelligence. From the mass production of one million units in vehicles to the compatibility certification of openvela, and from in-vehicle cockpits to intelligent terminals of two-wheelers and robot computing power platforms, it reflects that the core competitiveness of Runxinwei's "domestic software-hardware integrated AI intelligent base" products and solutions continues to stand out.
At the event that day, Runxinwei officially launched the "1+4+N" strategic system. This system is centered around the Zhixin · Domestic Computing Platform + Zhiwei · AIOS + Zhirun · Edge-side Model to build a domestically controllable AI intelligent base, empowering four major sectors including vehicle terminals, mobile terminals, AI intelligent hardware, and embodied intelligence. Runxinwei mentioned that it will work with partners to promote the extension of intelligence to scenarios such as transportation and industrial manufacturing, and achieve the large-scale implementation of scenario intelligence through technology reuse and ecosystem collaboration, aiming to become a neutral and open industrial connector.
Liu Qing, the chairman of Runxinwei, said in his speech that the current automotive industry has entered a new stage driven by intelligence and domestically controllable. Industry pain points such as "being stuck in core technologies, price involution, and unstable supply chains" urgently need to be addressed. The breakthrough and large-scale application of domestic full-stack technologies are the keys to industrial development. Runxinwei will rely on its full-stack engineering capabilities, focus on core areas such as domestic intelligent bases, and work with the industrial chain to promote the large-scale application of domestic technologies in vehicles.
There is a common situation in the automotive industry where "it is difficult for domestic chips to be used in vehicles": not only do they need to break through the technical thresholds of R & D ecosystem support, stability, and ease of use for automotive chips, but also there is a lack of in-depth adaptation between the chip's underlying software and the vehicle terminal OS and domain control architecture. Currently, they are also stuck in the integrated optimization capabilities in aspects such as edge-side AI, multi-screen linkage, and global vehicle-smartphone interconnection. Chip manufacturers understand hardware, and vehicle manufacturers understand the whole vehicle, but the bridge of "software-hardware integration + experience translation" has long been occupied by international manufacturers, which has become one of the biggest bottlenecks for the large-scale application of domestic chips in vehicles.
A series of products of the domestically developed software-hardware integrated intelligent base independently developed by Runxinwei have achieved the rare "vertical integration ability" in the automotive field: from chips to systems, performance, and interaction, it is fully connected with the vehicle control service link, effectively solving core pain points such as long implementation cycles and high adaptation costs for domestic substitution. Its full-stack capabilities of "chips + operating systems + AI" can achieve in-depth collaboration among "chips - systems - hardware - applications", which can significantly shorten the R & D cycle of terminal customers such as automobile manufacturers.
Relying on this core base, Runxinwei achieved the mass production of one million intelligent cockpits in vehicles within five years of its establishment. It is equipped with models of well-known automobile manufacturers such as SAIC, BAIC, and Dongfeng, covering multiple platforms of passenger cars and commercial vehicles, and the monthly shipment volume is steadily increasing. Currently, it has completed the adaptation of mainstream domestic chips such as Unisoc, GAF, and Horizon. Runxinwei's C200 intelligent cockpit platform has become an industry-leading platform for the collaboration of "chip - cockpit - vehicle" in the commercial vehicle field.
With its full-stack technical strength, experience in mass production of one million units, and an open ecosystem, Runxinwei precisely meets the incremental needs of automobile manufacturers in areas such as intelligent cockpits and cockpit-driving integration, and has won model appointments from many automobile manufacturers. At the same time, based on the cockpit solutions using domestic chips, Runxinwei has achieved multi-platform and global implementation, and has perfect mass production and delivery capabilities as well as compliance capabilities.
At this launch event, Runxinwei centrally launched a core product matrix that can be applied to full-scenario intelligent transportation.
In the field of domestic AI intelligent bases, Runxinwei launched core products such as C200 and X100. Among them, C200, as the main mass production base, has a computing power of 60K - 90K, with a domesticization rate of 90%, and supports full-scenario interconnection and 3D immersive interaction. X100 is a high-level cockpit-driving integrated platform with a computing power of 100K + 80T, supporting functions such as high-speed NOA and automatic parking. The C200 platform has been appointed for more than 10 models.
In terms of AI and systems, Runxinwei launched the Intelligent Agent OS and the Zhirun edge-side multimodal model. As AI is moving from "generating content" to "executing tasks", the Intelligent Agent OS will reconstruct the intelligent organization method, achieve controllable execution and continuous evolution through "edge-cloud collaboration", and its three-layer AI capability system can meet the real-time and low-power consumption requirements of scenarios such as in-vehicle and industrial robots.
Runxinwei launched the latest generation of Zhiwei · AIOS Lite (Intelligent Agent OS). This system is built on openvela, with four core advantages of being lightweight, efficient, intelligent, and open. It is applied to Runxinwei's domestic AI intelligent base, meeting all transportation needs such as two-wheel intelligent meters and in-vehicle AIoT devices. It is equipped with a full-link toolchain, perfectly demonstrating the high cost-effectiveness of domestic substitution and ensuring the consistency of user experience.
Runxinwei launched the intelligent agents of the RideBrain M4 intelligent meter and the Zhiqu · Intelligent Code Meter. The RideBrain M4 is centered around the Zhiwei AIOS, including OpenClaw (crayfish). It deeply integrates AI active services into the product, combined with emotional interaction design, to comprehensively upgrade the user's cycling experience. The Zhiqu · Intelligent Code Meter relies on its seamless interconnection ability to continuously enrich the IoT ecosystem and expand the application scenarios of intelligent devices.
In the field of computing power and robots, Runxinwei launched the Rbox-S100 commercial-grade integrated computing power platform for the brain and cerebellum, targeting general intelligent robots and solving mass production pain points such as fragmented computing power and insufficient real-time performance.
At the launch event, Runxinwei signed strategic cooperation agreements with openvela and Shenduo Jizhi. Runxinwei said that in the future, it will continue to focus on the field of domestic AI intelligent bases, with technical strength and engineering capabilities as the core, and use high-domesticization solutions to help enterprises in industries such as automotive, AI intelligent hardware, and embodied intelligence achieve rapid mass production. It will work with industrial chain partners to build an industrial ecosystem and jointly create a new ecosystem for intelligent transportation.
As an IDH in the Xiaomi ecosystem and a core co-builder of the openvela ecosystem, based on the openvela technology base and the Gemini - S1 hardware support, Runxinwei has developed a full-stack system solution that can be commercialized and mass-produced, and implemented it in multiple scenarios such as AI Agent local services, in-vehicle AIoT smart screens, two-wheeler intelligent meters, and smart environment perception panels.
Wang Aijun, the general manager of the Xiaomi Vela R & D department, said that Runxinwei has made key contributions to the construction of the openvela ecosystem. Runxinwei's self-developed Gemini-S1 development board is the world's first development board to pass the official compatibility certification of openvela. He hopes that Runxinwei can take the launch of Zhiwei AIOS Lite as a starting point, continue to focus on technological innovation in transportation scenarios, explore more application scenarios of lightweight intelligent interaction, and let the openvela technology base shine with greater value.
The in-depth cooperation between Runxinwei and Shenduo Jizhi will focus on aspects such as the construction of the embodied intelligence industrial ecosystem, technology implementation, scenario empowerment, and industrial upgrading. The two sides will jointly promote the embodied intelligence technology from underlying innovation to large-scale industrial application.
Chen Kai, the founder of Shenduo Jizhi, said that the cooperation between Shenduo Jizhi and Runxinwei is based on the high complementarity of the two sides in the industrialization path of embodied intelligence. Shenduo Jizhi focuses on the R & D of physical intelligence, extracts physical common sense from first-person data, and builds a general base model and a software-hardware collaborative technology system. Runxinwei focuses on the domestic AI intelligent base and has the capabilities of a domestic chip computing platform, an AI operating system, an edge-side model, and an integrated computing power platform for the brain and cerebellum. Runxinwei can provide the underlying computing power and engineering support required for the edge-side deployment, industrial scenario verification, and large-scale application of Shenduo Jizhi's embodied intelligence model, helping the embodied intelligence technology to enter typical intelligent manufacturing scenarios such as material sorting, turntables, and assembly more quickly.