Geely will no longer develop traditional intelligent cockpits that lack AI capabilities.
Geely Auto’s “One Geely” strategy has finally extended from the brand and assisted driving levels to the smart cockpit.
At tonight’s Geely AI Smart Cockpit launch event, Geely Auto released Flyme Auto 2 and revealed that ZEEKR OS 7, based on the unified AI OS architecture, will be launched soon.
Geely stated at the conference that new models under Geely Auto will be uniformly equipped with Flyme Auto 2, and will also unify the AI Agent and user ID.
This means that users can use the same account in the Geely Galaxy M9 and Lynk & Co 10-EM-P, and their personalized functions will not need to be reconfigured when switching cars.
The first models to be equipped with the Flyme Auto 2 system are the Geely Galaxy M9, which will start pre-sales on August 23, and the Lynk & Co 10 EM-P, which has already started pre-sales. In addition to new models, this function will also be pushed to old users through OTA.
According to Geely, this architecture will bring improvements in various aspects such as the initiative of voice assistant services, interactive experience, and openness of the cockpit ecosystem.
What is its specific performance and capability? Let’s take a look together.
“Services Find People”
“From today onwards, Geely will no longer develop traditional smart cockpits without AI capabilities.” This is a bold statement made by Geely before formally introducing Flyme Auto 2.
Flyme Auto 2 released by Geely tonight is not just a simple iterative product, but a dividing line for the strategic shift of Geely Auto Group.
Looking back at Flyme Auto 1, in two years, it has completed 9 version iterations, added more than 150 new functions, optimized 498 functions, accumulated 40 OTA updates for models, and covered nearly 5 million users.
Whether from the perspective of user base or technical accumulation, there is sufficient confidence for technical improvement.
In the new era, Geely requires the AI cockpit to achieve “services find people”, that is, to “understand warmth and cold”, and to judge and actively provide services for users without the need for users to speak, rather than the previous simple question-and-answer mode. This requires a bottom-up reconstruction of technology.
It is reported that relying on the cloud capabilities such as the 23.5 EFLOPS comprehensive computing power of the Xingrui Intelligent Computing Center 2.0 and the Xingrui AI large model system, as well as the on-board AI Box with 200 TOPS computing power and 200 GB/s bandwidth, the smart cockpit system can have the capabilities of complex computing, continuous learning, and rapid response.
Coupled with the hardware-software decoupling under the native AI OS architecture, cross-domain SOA service orientation, and the “AI trains AI” shadow mode and data closed loop, the cockpit can actively provide services and “become smarter with use”.
The specific embodiment of this capability is the EVA intelligent agent.
As mentioned earlier, the biggest difference between EVA and traditional intelligent agents is that it can provide active services to users under the premise of self-perception and self-planning, presenting a personalized, memory-capable, emotional and more intelligent agent.
For example, in the demonstration video, EVA can change the voice style according to the user’s direct or indirect instructions, and can actively propose to play a certain type of music according to the user’s mood.
To improve the interactive experience, Flyme Auto 2 has added functions such as live desktop, 3D fusion navigation, and one-screen dual-application. More importantly, Geely said that it has covered more than 90% of mobile phone brands and will continue to open up the ecosystem to provide users with more applications.
New Era
Geely’s unification of the AI OS architecture, EVA intelligent agent, and user ID this time may be a key measure to promote technical collaboration among its brands and reduce redundant development under the “One Geely” strategy.
For users, being able to obtain a seamless experience in any Geely model is itself building a strong ecological moat and enhancing user stickiness.
For Geely, this can not only maximize R&D efficiency and reduce software costs, but also shape a unified brand technical image, reducing the difficulty for users to understand and perceive the technologies of different brands.
This is exactly Geely’s response to the industry trend shifting from “software-defined vehicles” to “AI-defined vehicles”. From the recent statements of new energy vehicle companies such as XPeng and Li Auto, it is not difficult to feel that the future competition in the automotive industry will be about the speed of AI iteration and intelligent experience.
The “data-training-iteration” closed loop formed by the Xingrui Intelligent Computing Center and massive user data is also helping Geely build a solid intelligent barrier, joining the competition for the commanding heights of the new round of arms race.
Today, when new energy vehicle companies are keen to shift towards the “intelligent robot” industry, Flyme Auto 2 and EVA intelligent agent may be the prototype for entering advanced intelligent assisted driving in the future.
However, it has many competitors, such as XPeng, which is preparing to make great efforts in the field of intelligence in the next two years, Li Auto, which pushed VLA Preview today, the momentum of HarmonyOS cockpit, Tesla across the ocean, etc. Internally, it is the problem of Geely’s speed and ability to deliver on its promises.
Can Geely catch up with the first echelon this time?
(End)
Discover more from ChinaEVHome
Subscribe to get the latest posts sent to your email.
