Li Auto has formally introduced the open-sourcing of its self-developed Halo OS car working system. The announcement by founder and CEO Li Xiang on the 2025 Zhongguancun Discussion board in Beijing on March 27 positions Li Auto as the primary world automaker to make its complete car working system freely obtainable to the broader {industry}.
“We’re prepared to push car working methods to new efficiency heights by open-source,” Li said throughout his presentation. “Open-sourcing isn’t nearly offering builders worldwide with a brand new technical platform and improvement instruments—it’s about opening up our price system. Behind open-source is our confidence in our know-how and humanity’s capacity to construct an ‘innovation neighborhood.’ We perceive that closed methods can solely amplify coefficients, however open-source can amplify the bottom.”
The Halo OS mission started in 2021 as a direct response to the worldwide chip scarcity and challenges of adapting closed-source methods to new {hardware}. Li Auto invested over one billion yuan (138 million USD) and deployed a 200-person improvement staff to create the system, which entered manufacturing automobiles in 2024.
Li Auto says Halo OS considerably outperforms industry-leading closed-source AUTOSAR methods throughout key metrics, together with core efficiency, safety, value effectivity, adaptability, and adaptability. By revolutionary cross-system architectural design, Halo OS optimizes your entire “perception-decision-execution” chain, doubling response velocity and bettering response stability by an element of 5 in comparison with AUTOSAR methods.
These efficiency benefits translate to tangible security advantages, corresponding to decreasing automated emergency braking (AEB) stopping distances by seven meters at 120 km/h freeway speeds. The system additionally enhances Li Auto’s air suspension system by bettering physique posture management precision by 73% by clever coordination between cameras, central processing items, and the suspension system.
Concerning {hardware} compatibility, Halo OS dramatically reduces the time required to adapt to new chipsets. Whereas closed-source methods sometimes require 3-6 months to regulate to new chips, Halo OS can full chip adaptation and verification in 4 weeks. The system helps all automotive chip architectures, together with Infineon and Nvidia, and rising Chinese language chips from Horizon Robotics and SemiDrive.
“At present, there isn’t a vehicle-oriented working system within the open-source ecosystem that spans the horizontal domains of cockpit, clever driving, and chassis, in addition to the vertical layers of purposes, methods, and {hardware},” defined Huang Zhen, Li Auto’s car working system structure lead. “The technical obstacles are extraordinarily excessive.”
Li Auto’s transfer comes as the corporate accelerates its AI improvement. It lately rolled out end-to-end imaginative and prescient language fashions for clever driving and plans to launch its next-generation autonomous driving structure, MindVLA robotic massive mannequin, later this yr. Halo OS is designed to function the technical basis for these superior AI purposes, successfully addressing the core battle between computing energy calls for and provide.
The corporate plans to steadily launch Halo OS to the open-source neighborhood in late April. Its core parts embody the car management working system, clever driving working system, communication middleware, and virtualization platform. Li Auto estimates that open-sourcing these applied sciences might assist the automotive {industry} save 10-20 billion yuan ($1.4-2.8 billion USD) yearly in redundant analysis and improvement assets.
By making Halo OS open-source, Li Auto goals to interrupt down the “black field” technical obstacles of closed methods and reconstruct the technological monopoly panorama at present dominated by worldwide suppliers. The corporate’s imaginative and prescient parallels how Android remodeled the cell phone {industry}, making a common platform that accelerated innovation and diminished prices throughout the ecosystem.
As automobiles turn into more and more software-defined, Li Auto’s daring transfer positions the Chinese language automaker on the forefront of automotive working system improvement. It probably creates an “Android of vehicles” that would turn into the muse for clever automobiles worldwide within the coming AI period.
