Gasgoo Munich- The 8th AI Smart Cockpit Conference and 2026 4th Automotive Display & Perception Technology Innovation Summit wrapped up in Shanghai on June 26, 2026. Hosted by Gasgoo, the event brought the industry's focus to the future of in-car intelligence.
As the global auto industry undergoes a profound reshaping driven by electrification and intelligence, the smart cockpit has emerged as the core battleground for innovation and competition. It serves as the central hub for user experience and the deployment of intelligent technologies. Yet, the industry faces bottlenecks like feature stacking and severe product homogeneity. Rapid iterations in large models, cockpit agents, cabin-driving fusion, high-definition displays, and multimodal perception are pushing the industry beyond hardware comparisons into a contest of full-stack software, AI interaction, and ecosystem integration. With China’s auto market entering a phase of stock competition and surging exports, domestic brands face fresh opportunities and challenges in taking their intelligence global. At this critical juncture, the summit convened players across the supply chain—from OEMs to chipmakers and software developers—to explore landing paths for AI-native cockpits and vehicle-grade large models. The focus: finding breakthrough solutions amidst homogeneity and charting a course for the long-term development of intelligent, visualized, and integrated perception cockpits.
The conference’s second day dove into core issues like the evolution of emotional interaction, intelligent light control, spatial reconstruction, AI-native ecosystem building, and the transformation of in-vehicle entertainment. Discussions centered on how China’s auto industry can accelerate the construction of future-ready smart cockpit ecosystems—driven by technological innovation and anchored in user experience—as the era of "AI-defined cars" takes hold.
Special thanks go to our 26 ecosystem partners: Sunlord, SONATUS, QNX, CETC Fenghua, Yipu Optoelectronics, iFlytek, Lanrun, Hiwave, Renesas Electronics, Rockchip, BOS Semiconductor, Zhican Technology, Silux, Zhiyuan Electronics, Ruiping, Chuanjing Technology, Lidao Technology, Coloda, Anhui Xiangyu, Xinxinteng, Tongxing Intelligent, Yangjie Technology, Beijida, ETAS, Fushi Electronics, and Haoou Electronics.
From Hearing Commands to Reading Emotions: How Emotional Interaction Redefines the Smart Experience
Zhang Hui, Director of the Digital Cockpit Development Department at FAW’s R&D General Institute, pointed out significant limitations in current smart interaction evaluation systems. Mainstream quantitative metrics fail to reflect real user experience, focusing solely on system capability while ignoring emotional shifts and cognitive load. The emphasis on task completion often overlooks the impact of the process itself. In typical scenarios, a system’s calm tone during a traffic jam fails to detect driver agitation, or active inquiries during long-haul fatigue disrupt rest. Repetitive voice confirmations and mechanical responses expose a critical gap in emotional perception.
Zhang emphasized that human-machine interaction is essentially two-way communication built on emotional resonance. As the cockpit becomes a private "third space," the relationship between human and car should be one of partnership. The Kano model suggests current voice indicators only meet basic needs; achieving product loyalty and delight requires reaching "attractive quality" levels—proactive emotional care and emotional value. Data backs this up: negative emotions increase risk by 10 times, while emotional intervention can cut risk by 60%. Moreover, 58% of car owners recognize emotional care as a core safety feature, 38% believe in-car conversations can satisfy emotional venting, and emotional value ranks among the top three factors in car selection. Crucially, 67% of users are willing to pay for advanced intelligent experiences.
Interaction evolution spans three layers: command-based, state perception, and emotional resonance. The industry currently sits at the state perception stage, merely identifying external emotions. With AI and large models, systems can track user behavior and personality over time to achieve deep emotional resonance. Zhang introduced the H-K2 model to systemize this shift: H (Humanity) is the core; K (Basic Execution) focuses on accuracy and efficiency; A (Awareness) captures emotions; R (Resonance Response) provides warm feedback; and E (Evolution) achieves personalization through long-term memory. This model serves as both an evaluation tool and a design guideline, driving interaction from functional implementation to holistic emotional connection and co-creation of experience.
Zhang Hui | Director, Digital Cockpit Development Dept., FAW Group R&D General Institute
From "Blocking Light" to "Understanding Light": Third-Gen EC Solid-State Dimming Glass Reshapes Cockpit Light Control
Wang Chong, co-founder and COO of Lanrun, shared the company’s latest progress in smart cockpit light control, detailing the principles behind its third-generation electrochromic (EC) technology. The product features a five-layer structure—a 0.4-millimeter privacy film similar to a thin-film solid-state battery. When charged, ion migration triggers a redox reaction that darkens the glass. As a current-driven device, its color-changing speed continues to iterate: side windows can now change color within 20 seconds at room temperature, with the latest EC3.0 PLUS slashing that to just 10 seconds.
Regarding the smart cockpit, Wang believes the future will rely primarily on voice commands and light sensors. The Li Auto L9’s light-sensing privacy glass, which supports AI and voice control, serves as a prime example. He noted that as competition in the smart cockpit intensifies, companies in the optical chain must empower the cockpit through light control experiences, turning it into a new selling point and a critical variable. Lanrun plans to continuously innovate and iterate EC technology to meet diverse dimming demands, transforming dimming glass from a mere sunshade into an interactive entry point for smart glass displays.
Wang also outlined the "ultimate experience" of EC smart dimming across five dimensions: Privacy, where the interior can see out but the exterior cannot see in; Comfort, which prioritizes ergonomics over just speed; Energy Saving, featuring low voltage and zero power consumption in static states, along with thermal insulation that aids vehicle thermal management; Safety, maintaining transparency after power loss for observation; and Aesthetics, achieved through a neutral gray material system. The company holds a full chain of patent barriers, controlling the discourse from molecules to mass production.
Wang added that the film technology applies to architecture, consumer electronics, and vehicles. The company is collaborating with a leading domestic AI eyewear brand and will soon launch four-color changing and AR glasses products. By 2026, Lanrun aims to become the industry leader in annual shipments of EC smart dimming side windows. The goal is to become a global leading supplier of smart dimming components and solutions, creating a more comfortable, safer, and smarter third living space for users.
Wang Chong | Co-Founder & COO, Lanrun
Applications for In-Vehicle Photoelectric Interaction
Lu Rui, President of Market Strategy at Yipu Optoelectronics (Fujian), noted that while W-HUD is currently common, its display content overlaps with the instrument cluster. For safety and regulatory reasons, the two must coexist on the same screen. AR-HUD, despite being proposed for years, faces significant technical barriers; achieving a complete AR image through the windshield in scenarios like high-speed daytime driving or night alerts remains nearly impossible. By specification, AR-HUDs are divided into mini (about 10 degrees), standard (about 15 degrees), and full (spanning three lanes). The latter two are difficult to mass-produce due to volume constraints.
To tackle these challenges, Yipu Optoelectronics is attacking the problem from three angles: software, optics, and light sources. On the software side, ADAS and map data are processed and integrated into the HUD, combined with anti-shake tracking, and a free virtual environment software has been developed to simulate display effects under various lighting and speed conditions. Optically, while waveguides and holography show promise, traditional free-form mirror combinations remain the mainstream stable solution. Lu pointed out that the small HUD window struggles to balance safety, interaction, entertainment, and utility—it needs to work with car lights. Lights can achieve adaptive lighting and projection, such as projecting safety zones when reversing or trajectory lines on narrow roads. Combined with the HUD, they can present a comprehensive view on the windshield. The vision is this: PHUD handles infotainment and information, HUD handles driving data, and lights carry external-internal interactive feedback. Together, they make the vehicle provide feedback like a robot, helping the driver judge when to take over.
Yipu Optoelectronics started with free-form mirrors and now covers lenses, PGUs, optical chips, and complete solutions across four group brands. Future goals include achieving a 20–45 degree field of view (FOV) while shrinking volume, solving issues related to nonlinear films, optical engines, and AR algorithms, and expanding applications to car bodies and side windows. The company looks forward to collaborating with the industry to find even better solutions.
Lu Rui | President of Market Strategy, Yipu Optoelectronics (Fujian)
Fully Automated Assembly Line Solutions for Smart Automotive B-Pillars
Liu Jian, Deputy General Manager of the Intelligent Equipment Business Unit at CETC Fenghua Information Equipment, noted that in the field of smart cockpit display equipment, the company provides production lines for triple-screen displays. These lines can continuously vacuum-bond three display modules onto a single CG cover plate, covering processes like glass cleaning, soft-to-hard lamination, and hard-to-hard vacuum bonding. The center console production line is fully automated, enabling unmanned factories. The process flow covers cleaning, bonding, de-bubbling, backlight, film application, dispensing, and assembly, forming five major process series: cutting, bonding, bonding, de-bubbling, and backlight.
Liu emphasized that the fully automated assembly line for smart vehicle B-pillars is a flagship product. The existing 60-meter line contains 17 processes, while the latest 70-meter line includes 23. These lines have served top-tier automakers like Tesla, Xiaomi, and Huawei-backed brands, with over 1 million sets of pillar products shipped. The process covers the full workflow: glass processing, cleaning, coating, assembly, pressing, and packaging. After glass arrives, it undergoes brush washing, protective film application, power testing, and bending. Both the glass and plastic baseboard are cleaned with isopropyl alcohol and plasma, then coated with primer and PU glue. After visual alignment and high-precision pressing, the assembly rests for 12 hours. Semi-finished products undergo air tightness and dimensional precision testing, followed by waterproof foam attachment, Bluetooth board thermal riveting, harness welding, and UV glue curing, culminating in automated inspection and packaging. Core equipment includes the visual alignment high-precision pressing machine and a flexible waterproof foam curved applicator, which shapes the foam before attaching it to the pillar's curved surface. The company continues to drive automation and intelligence upgrades, providing high-precision, high-reliability equipment support for smart cockpit manufacturing.
Liu Jian | Deputy General Manager, Intelligent Equipment Business Unit, CETC Fenghua Information Equipment
AI Cockpit: The Partner-Like Evolution of Intelligent Agents
Lin Weichao, Product Director at Foryou General Electronics, noted that new energy vehicles now account for over 60% of sales—a figure that keeps climbing, marking an irreversible trend. For Gen Z buyers, intelligence is the core decision-making factor, spanning both assisted driving and cockpit experience. The development of smart cockpit multimodality is divided into five levels (L0–L5): L0–L1 is the perception stage, while L2–L4 is the cognitive stage. Multimodal large models are currently facilitating the shift from perception to cognition. Most of the industry has moved past the stages of no AI and simple machine learning fusion, and is now at the initial stage of fusing multimodal AI large models. The next challenge is edge-cloud collaboration services, with the ultimate vision being a fully intelligent and cognitive cockpit.
Lin pointed out that large model applications still face issues: application scenarios are converging with limited coverage, and the shift from single-vehicle intelligence to cross-device intelligence is just beginning. Additionally, introducing large models can lead to interaction errors, such as voice commands mistakenly turning off lights, which impacts safety. In response, Foryou has launched an edge-cloud large model and a full-domain AI collaborative architecture that supports A to A and MCP standard AI protocols. The cockpit architecture is divided into a data chain layer, a capability layer, and an application layer. The capability layer integrates data fusion processing, multimodal scheduling strategies, and AI middleware to facilitate access to various AI models. The application layer integrates customized third-party applications, focusing on user experience and intelligence levels while balancing natural interaction, fuzzy command resolution, and cross-scenario, cross-device, cross-service experiences.
In terms of product layout, Foryou Group offers cockpit services including external perception, HUD, VPD, and multimedia infotainment systems. It uses AI to collaboratively schedule applications, focusing on smart cockpits, assisted driving, and intelligent connectivity. The philosophy is to empower customers and build ecosystems with partners. Currently, Foryou General’s cockpit ranks ninth, but is expected to rise to fifth or sixth once mass production projects ramp up; its HUD and display products are already industry leaders. Earlier this year, the company began a front-loading mass production project with a southwestern customer. By late July, it expects to mass-produce an edge co-processor computing box that supplements general computing power with AI, continuously driving the evolution of cockpit intelligence.
Lin Weichao | Product Director, Foryou General Electronics
Screen and Sense as One: AI Multimodal Interaction Driving the Fusion of Display and Perception
Wang Kan, Head of Innovation Development Department 3 at Tianma Microelectronics’ R&D Center, stated that smart cockpits are evolving from L0 to L4 along the paths of intelligence and connectivity, currently sitting at L2. Full L4 cognitive cockpits are expected after 2032. As high-level assisted driving evolves from L2 to L3 and L4, freeing up driver attention and time, the cockpit is transforming from a driving domain into a comprehensive mobile living space. On the market front, the global automotive display market is projected to reach $17.3 billion in 2025 and $19.1 billion in 2026, growing at 12.2% annually. The installation rate of pre-installed smart cockpits in China has surpassed 80%, driven by the iteration of base technologies, the upgrade of interaction ecosystems, and the diversification of AI mid-ends.
Wang emphasized that AI interaction is shifting from single-modal to multimodal and from passive to active. Screens must evolve from simple output devices to perception terminals capable of seeing, sensing, and understanding. Leading manufacturers have already begun fusing display, perception, and AI to drive the evolution of interaction into proactive intelligent agents. Due to demands for larger, multiple, and freely shaped screens, displays must now carry interaction and perception capabilities. Perceiving humans and the environment is crucial for reducing accident risks, making the screen the optimal interface for integrating perception devices. By combining symbiotic algorithms with on-device large models, it is possible to achieve localized, low-latency, and more private interactions.
Tianma has proposed the "Tianxuan Five-Star" standard, defining automotive display quality across five dimensions: safety, reliability, usability, personality, and health. This covers specifications like eyes-on-the-road display, automotive-grade lifespan, high-precision touch control, irregular curved designs, and low blue light/power consumption. Cutting-edge products include light-responsive smart displays, 17-inch LCI smart light control, MicroLED Aurora HUDs, and the Tianxuan series’ panoramic window screens, flowing curved center screens, and flexible rollable screens that switch between 14 and 17 inches. Looking ahead, once screens can see, sense, and understand, they will read emotions and predict needs. Combined with AR, VR, and AI, this will provide a multi-sensory immersive experience, turning every screen into an intelligent agent. The goal is imperceptible interaction with tangible display—this is the original intent behind Tianma’s Five-Star standard.
Wang Kan | Head, Innovation Development Dept. 3, R&D Center, Tianma Microelectronics
Emotional Expression in the Era of AI Smart Interaction
Zhao Hengli, Smart Cockpit Chief Engineer at BAIC Research General Institute, stated that the smart cockpit is upgrading from command-based passive interaction to vehicle-level AI agent interaction featuring deep intent parsing, scenario prediction, and autonomous reasoning. This enables the vehicle to understand and proactively serve the user. While cockpit hardware moves toward spatial integration and architecture adopts on-device large models with edge-cloud fusion—transforming the car into a third living space—bottom lines like safety first and "human master, car assist" must be upheld. AI development is evolving from engineering models to general foundational models and then to vertical models. Automotive applications are shifting from mechanization to cognitive intelligence and ecological intelligence, ultimately achieving unconscious adaptation, emotionality, and active empathy. The market for cockpit multimodal AI large models is expected to grow more than tenfold by 2030 compared to 2026. While Qualcomm remains the mainstream for cockpit SoC chips, domestic chips are rapidly breaking through, with large models forcing the iteration of chip computing power and storage.
Addressing pain points like technological silos and the fact that most manufacturers are stuck at the L1–L2 level, BAIC has chosen to focus on real-world scenario pain points, adhering to a "product-first" philosophy and anchoring the complete transformation chain from model to scenario to product. BAIC launched a "1+N" large model platform architecture: a central large model unifies intent understanding and cross-domain scheduling, while "N" vertical models cover scenarios like driving, cockpit, and music. Third-party models can be quickly adapted via standardized interfaces. The platform consists of a bottom hardware and data security layer, a middle self-developed engine layer, and an upper AI Agent layer, featuring edge-cloud integration, multimodality, private deployment, and full-vehicle unified scheduling.
Zhao noted that the platform already serves 440,000 vehicles, has processed 2.2 billion requests cumulatively, and sees 6.84 million daily active users, with positive feedback on voice interaction and AIGC. Looking ahead, BAIC has defined a three-generation iteration roadmap: digital wisdom, embodied wisdom, and ecological wisdom. The aim is to use AI to elevate the experience dimension, providing warm emotional services through anticipation and companionship, and to reconstruct the human-car relationship with an open architecture—bringing mobility back to a human-centric focus.
Zhao Hengli | Chief Engineer, Smart Cockpit, BAIC Research General Institute
The Next-Generation Smart Cockpit Driven by Agentic AI
Li Jie, Strategic Deputy General Manager at Jidou Technology, introduced the company’s 12-year journey of business transformation. Starting with integrated software and hardware BOX products, Jidou gradually entered the pre-install market and deepened its layout in the AI smart cockpit sector. To date, it has partnered with 33 OEMs. Li noted that the company has achieved significant scale in user-side commercialization; in 2023, during the pre-agent stage, the GMV of its in-vehicle systems reached approximately 230 million yuan, covering scenarios like refueling, parking, charging, and media—placing it at the forefront of the industry.
Regarding the state and future of the smart cockpit, Li pointed out persistent industry pain points: stacked functions with low usage rates, insufficient user awareness, and safety hazards from screen operation while driving. Additionally, high-frequency scenarios like refueling, charging, and parking still rely on mobile phones, creating a fragmented experience. He divides smart cockpit development into three stages: mechanical cabin, electronic cabin, and smart cockpit. The industry is shifting from function stacking to "human-centric proactive services." In the future, voice interaction and AI agents will be core, enabling seamless access to services like coffee ordering, car maintenance, and designated driving, while reducing reliance on app stacking and screen operation. The company has completed payment channel integration and established multiple security mechanisms, verifying coffee payments in real vehicle scenarios. Future steps focus on refining functions, improving real-time navigation updates, and landing the A2A ecosystem, further driving the evolution of the head unit into a service-oriented intelligent agent.
Li Jie | Strategic Deputy General Manager, Jidou Technology
Reconstructing the Smart Space Experience with AI-Native Technology
Li Pan, Co-Founder and President of Yitu Technology, stated that the company was established in 2023 amid the rapid rise of the global AI Agent wave. Its core positioning is as an "AI-native foundational company" serving intelligent terminals like AI smart cockpits and AI vehicles. By providing model capabilities, agent systems, and full-stack technical abilities, Yitu aims to drive the evolution of automobiles and other terminals toward AI-native forms. The team currently numbers about 200, distributed across five centers nationwide, covering the complete chain from model training and AI algorithms to agent development and automotive engineering delivery. The company has completed four rounds of financing, securing strategic investment from ByteDance and forming deep technical cooperation with the Doubao large model.
At the product and industry level, Li noted that smart cockpits are currently evolving from L1 to L2, facing issues like fragmented traditional interaction chains, less-than-smooth mass production experiences, and immature agent development paradigms. He believes the industry is shifting from "manual workflow-style agents" to LLM-driven agentic architectures, emphasizing that automotive scenarios must adopt an edge-cloud integrated architecture to balance safety, privacy, and ecosystem experience.
In terms of product path, the company advances AI cockpit capabilities around the goals of being "smarter, understanding you better, warmer, and more all-around." Through end-to-end voice large models and AI-driven HMI interaction, it achieves a leap from command-based control to natural interaction and task execution. Simultaneously, Yitu is actively expanding agent-level cooperation with ecosystem partners to explore integrated service capabilities covering food, clothing, housing, and transport. This pushes the car from "software-defined" to "AI-defined" terminals and builds an AI-native ecosystem for a broader range of smart devices.
Li Pan | Co-Founder & President, Yitu Technology
From Technology to Aesthetics: The Immersive 5D Smart Cockpit
Fu Yetao, VP of the Automotive Business Unit at Danghong Technology, introduced the company as an audio-visual technology enterprise founded in 1994, deeply rooted in the industry for over 30 years. In the automotive sector, Danghong has built a comprehensive solution system centered on bringing content into vehicles, enhancing immersive cockpit experiences, and automotive-grade camera and sensor algorithms. These solutions have been applied in over 13 million mass-produced vehicles, meeting content ecosystem needs across different markets, especially for automakers exporting overseas.
Regarding technological evolution and product innovation, Fu highlighted the upgrade path toward a "5D immersive cockpit," which builds a more immersive in-car experience through the multimodal fusion of vision, hearing, touch, and even smell. Visually, super-resolution and color enhancement technologies can upgrade 720P/1080P video to near-4K quality in real-time, combined with AI recognition for ambient lighting linkage. Auditorily and tactually, sound field modeling and content understanding enable spatial audio and seat vibration feedback. Olfactorily, a digital fragrance system generates scents matching content scenarios. Additionally, leveraging AI and AIGC, the company can generate audio-visual content in 5–10 seconds and supports immersive mode construction based on scenario recognition or user customization. For vehicles with active suspension, dynamic scenario modes like roller coasters, typhoons, and earthquakes can be developed to further enhance the vestibular perception experience, driving the cockpit from functional interaction to a multi-sensory immersive entertainment space.
Fu Yetao | VP, Automotive Business Unit, Danghong Technology
Reshaping the Cockpit Audio Ecosystem with a Full-Domain Personalized Content Engine
Zhang Xinyong, Director of Business Development at CNR Cloud Listening (Internet of Vehicles), stated that the current in-vehicle audio experience faces obvious scenario matching gaps and fragmentation. On one hand, systems struggle to accurately identify vehicle status and immediate user needs, leading to content pushes that don't match the driving scenario. On the other hand, switching between multiple apps fractures the audio experience, making it difficult to form a continuous, immersive service. He noted that the in-vehicle audio industry is shifting from an "application era"—where competitiveness was defined by app quantity and copyright resources—to a new stage centered on "system-level ecosystems and intelligent scheduling capabilities." The key to future competition will no longer be content reserve scale, but rather the ability to operate a unified content operating system and distribute content based on scenarios and intent.
Against this backdrop, Cloud Listening proposes building a "full-domain personalized content engine" with a core architecture of perception, decision, and distribution layers, achieving an upgrade from "humans finding services" to "content finding humans." The system uses real-time scenario recognition and semantic understanding to adapt to various driving situations like morning rush hour commutes, highway cruising, and night drives, automatically matching appropriate content and service forms. Relying on the authoritative content resources and intelligent audit capabilities of the CNR Media Group, Cloud Listening holds unique advantages in content safety and compliance. It has constructed a three-layer business loop: "content authorization—OEM customized system—AI subscription services." In the future, Cloud Listening hopes to join forces with OEMs, Tier 1 suppliers, and AI ecosystem partners to build an open content operating system, evolving in-vehicle audio from a tool-based application into a "smart content partner" capable of emotional perception and long-term understanding, integrating naturally into the full-scenario vehicle experience.
Zhang Xinyong | Director, Business Development, CNR Cloud Listening (IoV)
AI-Native Tech Empowers the Upgrade of Smart Cockpit Interaction
Yang Wei, Director of the Cockpit-Driving Development Service Department at Lianyou Intelligent, stated that smart cockpit interaction is evolving from physical buttons and screens to voice, and finally to multimodal intelligent interaction. The physical button stage focused on mechanical control with closed functions. While screen interaction increased information richness, it introduced driver distraction risks. Traditional voice interaction freed hands to some extent but remained rigid due to three-stage architectures and fixed command patterns. With AI development, multimodal interaction is becoming the new direction, fusing voice, vision, and vehicle signals to achieve a more natural and personalized in-cabin experience.
Technologically, Lianyou proposes the MSP (Multimodal Reasoning Middleware) solution. The perception layer uniformly accesses multimodal data; the reasoning layer uses large models as the core decision engine; and the execution layer atomizes vehicle capabilities into callable skills, achieving end-to-end decoupling and efficient scheduling. The system also introduces MCP interaction logic and a dual-channel memory engine, enabling "thousand-person, thousand-face" personalization through long-term preference memory and scenario memory. On the safety front, a "safety gate mechanism" is established to boundary-check model outputs and skill calls, addressing large model hallucinations and complex driving scenario constraints.
At the application level, the architecture supports complex semantic understanding, cross-domain continuous dialogue, and memory-based services, extendable to scenarios like dangerous behavior recognition and AI sentinels. However, on-board large models still face challenges in privacy compliance and computing power limits. Therefore, the company believes the future key direction is promoting on-device model deployment, ensuring high-quality intelligent interaction without data leaving the domain.
Yang Wei | Director, Cockpit-Driving Development Service Dept., Lianyou Intelligent
The Next Battlefield for Cockpit Entertainment: YOCY’s Product Practice and Industry Insights
Huang Wei, Co-Founder of Shanghai YOCY, stated that in-car gaming is becoming a key application in smart cockpits, combining high user acceptance with significant commercial potential. Its core value lies in upgrading the car from a traditional "functional space" to an "experience and consumption space." In terms of market size, as the smart cockpit penetration rate rises to 52% in 2025, adding over 10 million new smart cockpit vehicles annually to a base of about 100 million existing users, the in-vehicle entertainment market holds billions in potential—assuming roughly 200 yuan in annual entertainment consumption per car. As autonomous driving gradually becomes widespread, freeing up in-car time, gaming is poised to become one of the core high-frequency scenarios.
Addressing current industry bottlenecks—severe hardware and system fragmentation, high adaptation costs, long automotive certification cycles, complex OTA updates, and insufficient continuous content operations—YOCY has proposed the "YOCY Automotive Entertainment Platform." Benchmarking Steam, it builds an in-vehicle gaming infrastructure system via a unified SDK, testing system, distribution channel, payment commercialization, and operation system. This connects automakers and game developers, lowering entry barriers and boosting content supply efficiency. In terms of product form, the platform covers cloud gaming, light gaming, and retro games, with plans to launch motion-sensing games based on in-cabin cameras to further enhance interaction.
Regarding progress, the platform has achieved mass production and covers multiple automakers, with plans to promote pre-installation in millions of vehicles and OTA upgrades for existing stock. In its evolution, in-car gaming will upgrade from 1.0 to 2.0, combining 5D cockpit capabilities, AI technology, and real vehicle data to enable immersive experiences like multi-car real-time track generation. This will accelerate the evolution of in-car gaming toward a "car-native content ecosystem."
Huang Wei | Co-Founder, Shanghai YOCY
The Smart Cockpit Entertainment Revolution in the AI Era
Bo Zheng, CTO of Beijing Hummingbird Interactive Technology, noted that the company's core team comes from the traditional game industry, bringing mature experience in game development and commercialization. The company subsequently expanded into VR and automotive scenarios, completing a VR cockpit replication project in 2017 and launching in-car custom game products in 2018, gradually forming capabilities in immersive entertainment and interaction solutions for automakers.
Regarding the development of smart cockpit entertainment in the AI era, Bo stated that the car is evolving from a mobility tool to a "third space," where in-car time becomes a key carrier of experience. AI is driving the cockpit from a collection of functions toward scenario-based and emotional interaction. On this foundation, the company built the "Fangka" in-vehicle entertainment platform, integrating content forms like cloud gaming, H5 games, and arcade simulators, enabled by a cloud controller for instant play. The platform is evolving from basic content distribution to an AI-capable interaction system, gradually introducing emotion perception, companion interaction, and scenario generation. It explores personalized entertainment experiences based on travel memories and LBS, moving the cockpit toward a smarter entertainment space with deeper emotional connections.
Bo Zheng | CTO, Beijing Hummingbird Interactive Technology
From "Fixed" to "Fluid": Mobile Center Consoles and the Spatial Reconstruction of the Smart Cockpit
Zhou Yibin, Mechanical and Optoelectronics Section Chief at Ningbo Inoue Huaxiang Automotive Parts, stated that the automotive center console is evolving from a traditional "function and storage center" to a "smart space and interaction carrier." In the ICE era, the console primarily integrated physical controls and storage. In the NEV era, with the proliferation of wire controls and column shifters, the fixed console is further transforming into a pivot for spatial reconstruction and intelligent interaction, gradually integrating voice interaction, head unit screens, and sensors.
Building on this, the mobile center console represents the next major direction. Compared to fixed structures, its sliding design releases redundant interior space. In L3/L4 autonomous driving scenarios, it supports various modes like office work, entertainment, and childcare, extending further to in-cabin services, health monitoring, and forgotten item reminders. Relying on electrification, connectivity, and intelligence, the console enables multi-zone sharing and full-vehicle linkage control. Relevant solutions have already achieved mass production in models like the Deepal S09. Through sliding rails and modular design, they are driving the evolution of interior space from "fixed partitions" to "reconfigurable shared spaces."
Zhou Yibin | Mechanical & Optoelectronics Section Chief, Ningbo Inoue Huaxiang Automotive Parts
With that, the 8th AI Smart Cockpit Conference and 2026 4th Automotive Display & Perception Technology Innovation Summit drew to a close. Centered on AI smart cockpits, automotive displays, and multimodal perception, the event brought together frontier insights and practical experience from OEMs, supply chain enterprises, and tech companies. It systematically presented the industry trend of smart cockpits evolving from function stacking to AI-native interaction, and from single displays to multi-perception fusion. Participants engaged in deep discussions on key issues such as interaction reconstruction, content ecosystem upgrades, spatial design innovation, and safety compliance systems, providing vital references and inspiration for the industry’s next stage of technical pathways and commercial implementation. As the convergence of multiple technologies accelerates, the smart cockpit is moving toward a new stage that is more proactive, emotional, and systematic. Industrial collaborative innovation is set to continue on a higher dimension.
