IDC Evaluation Shows Zebra Smart Cabin Leads, Breaking the Illusion of 'Pseudo-Intelligent' Cockpits}

As the trend of intelligentization spreads, smart cockpits have become a key factor in vehicle purchasing decisions. According to KPMG, 88% of Chinese consumers consider cockpit features when buying a car, with over 47% willing to pay for value-added functions within the cockpit.

Meanwhile, although the penetration rate of intelligent cockpits in new cars has reached 73% domestically and 58% internationally—accounting for over half of the market share—the user experience still needs improvement. Most players are still in the rule-based voice interaction stage of "pseudo-intelligence."

Therefore, large models capable of fuzzy semantic understanding and autonomous task execution have become the core technology for upgrading cockpit experience. As large models pass the commercialization inflection point, their integration into vehicles is entering a practical phase.

For consumers, cars have evolved from simple transportation tools to intelligent mobile terminals capable of real-time environment perception and personalized services. The "brain" of the smart cockpit—the large model—has become a critical factor influencing interaction capabilities.

In this context, IDC conducted real-world tests on the large model capabilities of nine mainstream automakers and technology providers, releasing the "Smart Cockpit Large Model Capability Evaluation, 2025" report. The assessment covered seven high-frequency scenarios: vehicle control, reading, entertainment, lifestyle, travel, business, and social interaction. Among them, Zebra Smart Cabin achieved perfect scores of 5 in vehicle control, reading, and lifestyle, and scored 4 in three other scenarios and 3 in one. It was the first to realize task execution in lifestyle and business, demonstrating initial Level 3 cockpit capabilities, leading the industry into the "true AI" era, with a clear industry gap emerging.

The report notes that Zebra Smart Cabin has pioneered vertical applications in lifestyle and business scenarios, such as the AI food delivery service launched in collaboration with ZhiJi and Ele.me in March, which is now in mass production.

This breakthrough is attributed to Zebra’s key technology—the "Yuan Shen AI Large Model". Based on the Tongyi foundation model, it is specially developed for intelligent cockpits. Compared to basic large models, this vertical scenario large model, built from millions of users, over 20 billion voice interactions, and a professional knowledge base, is a "brain" that understands cockpit scenarios better, becoming a crucial technological advantage for Zebra’s leading position.

Zebra’s Yuan Shen AI cockpit large model features a multi-model fusion architecture, end-to-end interaction framework, and intelligent agent system, pushing the experience towards "No Touch, No App" new AI interactions.

"Seven-Sided Warrior"—Building a Competitive Edge

The IDC report states that cars are evolving from transportation tools to the user’s exclusive "third space," with the smart cockpit being the core infrastructure of this transformation.

The development of smart cockpits is divided into five stages: from L1 traditional cockpits with no intelligence, to L2 autonomous thinking and rule-based decision-making, then to L3 task automation, L4 semi-active services, and finally to L5, capable of understanding emotions and providing proactive services.

In this process, the ability of large models to understand fuzzy intentions, generate responses, and autonomously execute tasks is key to advancing from L2 to L3, and Zebra’s leading performance in this area makes it a pioneer in the industry.

Among the seven evaluation dimensions, Zebra scored full marks in three, and four in others, leading the industry. Especially in vehicle control, reading, and lifestyle scenarios, Zebra has pioneered vertical applications, forming a closed-loop assistant for users.

Beyond evaluation scores, Zebra’s practical capabilities were showcased at the Shanghai Auto Show in June. While most smart cockpits remained in basic scenarios, Zebra demonstrated AI food delivery directly on the vehicle, providing authentic AI services on-site.

In addition to "ordering coffee," Zebra’s Yuan Shen AI intelligent agents, in the form of "1 System Agent + N platform-level Agents," have successfully integrated 10 major types of intelligent agents, transforming the car into a highly imaginative third space of life.

"Three Horses"—Building a Technological Moat

Founded in 2015, Zebra Smart Cabin has collaborated with over 40 automotive brands, with more than 10 million mass-produced vehicles. In 2024, Zebra’s new car cooperation rate in China exceeded 15%.

As one of Shanghai’s "Beidou Seven Stars" AI enterprises, Zebra’s strategy for maintaining industry leadership involves AI in All.

Looking ahead, Zebra plans to launch the "New Three Horses" strategy, transforming from an OS system-level software provider to a full-stack AI technology provider, and finally to an AI software ecosystem and joint operator.

Zebra’s Chief Product Officer Cai Ming emphasizes: "Under these three strategic directions, Zebra is the largest autonomous OS producer, the earliest to cover AI scenarios comprehensively, and the most deeply integrated AI ecosystem service provider."

Behind these strategies are deep software capabilities and technological accumulation.

Software: Zebra is one of the only two fully self-developed companies in China, leading the formulation of national standards for in-vehicle OS, and jointly developing integrated intelligent cockpit and driving fusion solutions.

Technology: Zebra has created the first multi-model architecture, pioneered the System Agent star-shaped collaboration architecture, and achieved AI-powered vehicle services.

Ecology: Zebra collaborates with over 400 partners, operates the largest in-vehicle mini-program platform, and has over 4 million monthly active users.

Of course, competition in the smart automotive industry is far from over. In the short term, technological depth remains the key battlefield; in the long term, ecological breadth will be even more critical.

Zebra’s ability to maintain its leadership in the next decade depends on transforming technological barriers into user habit barriers.

Conclusion: The Deep Water of Smart Cockpit AI Competition

Smart cockpits have shifted from a technological gimmick to a real user experience. With future iterations of large models and expanding vehicle ecosystems, the evolution towards "Seamless Interaction—Proactive Service—Continuous Experience" will accelerate, creating smarter, more convenient, and safer interactions for users.

Looking forward, IDC highlights that while large model iteration is crucial, ecosystem development must also speed up. In large model engineering, companies need to focus on content closed-loop optimization and refine interaction experiences.

The AI competition in smart cockpits has entered deep waters. Whether Zebra can continue to lead and how industry players will respond remains to be seen.