GAC announces groundbreaking AI technology at CES 2025, "domain-wide AI"

On January 8th, GAC showcased their industry-first "domain-wide AI" technology system at the 2025 Consumer Electronics Show. This comprehensive AI ecosystem integrates C, B, and A levels to create a cohesive smart eco-system, aiming to forge a seamless "lifestyle-wise intelligence life".

To support this ambitious goal, GAC has unveiled several innovative AI solutions, including Edge-AIOS intelligent scheduling engine and Flyme Auto, which can seamlessly integrate devices, such as cars, smartphones, wearables, and smart homes. The company expects "domain-wide AI" to unfold into three stages: 'assistant', 'body' and 'lifespan', with an overarching philosophy of "embodied+smart"

An industry-first end-to-end audio solution was developed by GAC. Replacing sequentially cascading traditional speech recognition methods, this state-of-the-art technology seamlessly understands nuances in expression, emotion, tone, and connotation. Moreover it brings emotional intelligence to AI systems and develops perception understanding through modeling the intermediate latent variables. Furthermore, the model has been expanded with advanced enhancements which enable audio "voice" and further enables better real-world interaction possibilities.

GAC has already introduced several applications across various models featuring these innovative technologies; starting with Flyme's latest features on new cars like silver rhino, silver e5, as well as the E8. Moreover AI smart passenger seats have integrated voice recognition that is able to predict what you need next and even automatically order food in restaurants

In the realm of computing power, GAC has established a cloud supercomputer - called GAC's Zhe Cheng SuperComputer, and successfully passed the certification of overall management processes for intelligent computational centers. It supports the needs for 500 million connected vehicles with an unmatched processing capability of around 102E21 operations per second.

In terms of data accumulation, GAC had accumulated total 10T token data; as well as a massive range of automotive domain-specific data, establishing itself to be capable in driving and enhancing large models, known as "cognitive capabilities". And most recently they published the GAC ZHE Cheng AI large model