At the 7th Future Network Development Conference, Peng Song of Huawei delivered a keynote speech entitled "Building an End-to-End AI Network and Unlocking All-Scenario AI Capabilities". He pointed out that network innovation in the AI era includes the goals of Network for AI and AI for Network.
The former aims to build a network that supports AI services, realize AI large models from training to reasoning, and from dedicated to general purpose, and open up the whole scene between terminals, edges and clouds. The latter uses AI to enhance network capabilities, making network devices more intelligent, networks more autonomous, and services more efficient.
When referring to AI for Network, Peng Song shared his vision for future autonomous driving technology. First of all, in the first stage before 2025, the goal is to achieve the L4 level of network autonomous driving, and realize network element level operation and maintenance and business intelligence by combining with AI.
Secondly, in the second stage around 2030, AI Native will be realized through AI large models, and L5 level automatic driving at the application and network levels will be realized.
Peng Song emphasized that AI for Network has brought five major innovation opportunities. In the field of telecommunications, operators can develop and train their own large models of communication networks based on telecommunications data. At the network equipment level, there are innovation opportunities for AI for data center network (DCN), AI for wide area network, and AI for edge and end networks. In addition, at the network brain level, automation will cover the entire process from planning to service experience, greatly improving the efficiency of network O&M and operation, and reducing the total cost of ownership (TCO) of customers.
Peng Song also made a prediction on the intelligence penetration rate of the AI industry. He said that the third wave of AI will sweep in, and it is expected that the penetration rate will increase from 7% in 2021 to 30% in 2026, and by 2030, the penetration rate will exceed 50%. AI will unleash great potential in scenarios such as unmanned driving, weather forecasting, and medical research and development.
It is worth mentioning that Huawei Cloud recently released an autonomous driving development platform with a built-in Pangu model. The platform is based on the Pangu large model and ModelArts AI development production line, and has been applied in Changan, FAW and other car companies, as well as commercial vehicle scenarios such as mining trucks, port ARTs, and dedicated logistics heavy trucks.
In addition, the Avita 11 Hongmeng version equipped with Huawei's ADS 2.0 intelligent driving system was officially launched last night, with 4 configurations and a price range of 300,000 to 390,000 yuan. This marks Huawei's further development in the field of autonomous driving technology, providing consumers with more intelligent travel options.