On September 20th, Huawei held its annual conference with the theme "Accelerating Industry Intelligence." Meng Wanzhou, Huawei CFO and acting chairwoman, emphasized in her speech that Huawei is committed to build the foundation of China's computing power, and provide the computing power, platforms, and various research tools for large language models (LLMs) in China.
According to Chinese media reports, Meng sees computing power as the core driver of artificial intelligence (AI) development, but it is scarce and expensive, which is a major obstacle to AI development. Currently, various industries are accelerating the adoption of LLMs, with some models having parameters exceeding billions or even trillions. The development of AI is also approaching a turning point.
The Huawei chairwoman believes that the era of large models has arrived, transitioned from the era of small models. In the past, different application scenarios required the development of different models, but large models can acquire a vast amount of knowledge, making it possible for a single model to be used in multiple scenarios, significantly lowering the threshold for AI and applications.
According to Meng, what Huawei aims to do now is to allow these AI models to "blossom," supporting businesses or organizations in using their own data to train their specialized large models, focusing on specific areas of expertise.
In 2003, Huawei introduced the "All IP" strategy, driving technological standards and protocols in the communications field. With the development of cloud technology, Huawei shifted to the "All Cloud" strategy in 2013 and became one of the significant players in cloud services.
According to Canalys data, in the second quarter of 2023, the size of China's cloud infrastructure services market was US$8.7 billion. Huawei's market share in the cloud market was 19%, second only to Alibaba Cloud's 39%, with Tencent Cloud at 15%.