Summary Bullets:
• At the Huawei Analyst Summit (HAS 2024) hosted in Shenzhen, analysts and press had the opportunity to find out more about the company’s “All Intelligence” strategy.
• The company is focusing on extending its reach across international markets outside its domestic market in China, particularly in EMEA and Latin America.
During his keynote address, Chairman Eric Xu emphasized that over the past 300 years, humans have gone through three industry revolutions: the intelligent revolution; the information revolution; and the computing revolution. In the next 20 years, we will enter an age of “all-intelligence”; with four technical breakthroughs responsible for that achievement: mathematic computing; clean energy; powerful and affordable pervasive computing; and the information explosion generated by using and creating massive amounts of data.
Huawei has a long heritage in AI, and we were not disappointed by the breadth and depth of the AI-related sessions. The company is focusing on making its Pangu suite of foundation models more automated, and delving deeper into industrial scenarios to help clients train their own foundation models on top of Pangu. Across the board, enterprises are experimenting with new ideas in AI, not just traditional fine tuning, also reverse training, mixture of experts (please see Mistral AI is not afraid of OpenAI and it has the open-source goods to prove it), retrieval augmented generation (RAG), etc. The need to have a “human in the loop” was highlighted too, since AI remains very far from replacing humans yet.
Huawei’s AI strategy is not limited to its traditional core expertise in the communications industry. Pangu has been trained for an array of scenarios, from forecasting the weather to industry-specific applications in finance, mining, manufacturing, retail, etc. During the event, new use cases driven by Pangu were showcased across a host of vertical industries outside of telecoms. The company also offers the Celia assistant to capture the broader opportunity from enterprises across many sectors adopting AI. This is par for the course and the new focus on generative AI (GenAI) will surprise nobody, given its extraordinary market potential. GlobalData forecasts the GenAI opportunity to grow 80% annually until 2027 to reach $33 billion globally.
The vendor maintains, however, an essential focus on telcos, and has also developed strategic goals around the commercial deployment of 5.5G by working with operators and extending its partner ecosystem at home and abroad. One of the important messages during the event was about delivering an AI-native cloud infrastructure to help customers unify resources across storage, networking, servers, etc. Further, the company is applying large language models (LLMs) to help make telco networks as autonomous as possible. Having a high level of autonomy in the network is paramount for CSPs. And in industrial environments, for example, remote control is of highest importance.
Huawei is also leveraging digital twin technology and GenAI for network environments. This entails classifying and cleaning the network data using natural language and finding out the status of the network by interacting with it using natural language, since Pangu makes it possible to transfer natural languages into machine-readable languages. This enables users to get timely reports on the status of the network. Offerings like this seem the right way forward, since the AI-native network is set to become an important enterprise trend in future, as carriers are enticed by the idea of having on-demand computing resources for the network with a unified orchestration of networking and computing resources.
IT Connection 