- Article

- Market & Regulatory Insights
- Spotlight Asia Pacific
AI in China – after the DeepSeek moment
The groundbreaking Chinese AI model has boosted the confidence of local innovators, as they create AI applications with real-life use cases.
China is at the forefront of AI development. Strong policy support, a deep talent pool, and the ready availability of venture capital, are all factors that combine to give the country an edge in the defining technology of our time.
And China’s AI progress has not gone unnoticed by investors. In the year-to-August, the Hang Seng Artificial Intelligence Theme Index, which tracks 40 Hong Kong-listed companies that are engaged in the AI value chain, rose by 34.8%, compared with a 28.9% gain in the broader market1.
The HSBC 12th Annual China Conference brought together the founders of three Chinese AI startups to discuss the key developments in this fast-evolving sector, as well the AI-powered solutions they are developing to make a real-life impact.
The DeepSeek moment
At the start of 2025, China surprised the global technology world by releasing DeepSeek – a large language model (LLM) that can compete with leading models using only a fraction of the computing power. One of the most notable aspects of DeepSeek is how cheap it is to train, compared to other models, which cast some doubt onto the huge investments being made in AI infrastructure.
The so-called “DeepSeek moment” is important for China for a variety of reasons. For a start, it is a source of inspiration for the country’s young scientists and engineers: “The rise of DeepSeek, and AI technology more generally in China, really gives us lots of confidence about what we can do,” said Cicy Zhang, Founding Member of cognitive training technology company BrainCo.
The impact is felt beyond technology specialists, as the launch of a high-profile homegrown LLM assures the broader public towards AI. As such, the general population is more willing than in the past to use the technology in their daily lives.
From a technological standpoint, DeepSeek stands out because it is open source – meaning that it is available for anyone to access, modify and further develop. This is important because it will promote a culture of sharing among AI practitioners and boost the application side of the technology, said Ye Tian, Founder and CEO of AI-robotics company RoboScience.
Developers will not always have to start from scratch as they will have ‘off the shelf’ models that they can use. And also, open-source software is much safer than closed source software, and this is true for AI as well.
|
The panel was in general agreement that DeepSeek was only the beginning of a long-term trend of AI innovation, which has the potential to persist into the coming decades.
Real-life use cases
While the development of LLMs gets much of the attention, developing applications where AI can be used by businesses is also very important, as it will be how companies derive revenue from the new technology.
Industrial robotics is an area where there is strong potential for China to excel. The country is already the leading market for robots, having installed more than half of the world’s units since 20212, and the next generation of automations will likely be powered by AI.
Existing robots are already capable of pre-programmed tasks in a known environment. What AI brings to the table is generalisation – the ability to perform a task in an unseen environment or with unfamiliar objects. That said, the AI robotics still lack the precision and efficiency associated with their older counterparts.
RoboScience’s Mr. Tian explained the three-stage process that his company will follow in implementing AI robots. First, robots will be used in settings where high levels of precision are not required, such as the services sector. As the robots become more accurate, they can be used more in an industrial setting. And finally, once the technology is fully refined, it can used to work closely with human beings.
Another area where Chinese companies is making progress is in multi-modality, which is where an AI system is able to process and understand information from various types of data types: including text, pictures, audio, and video.
As an example, Dr. Liwei Wang, Founder, Weitu AI, explained how his company is developing advanced multi-modal technology to understand the content from long videos accurately and efficiently. As these models develop, he expects the AI will gradually learn spatial awareness from the video files. This is particularly relevant in robotics: a more accurate understanding of the world around them will help machines better interact with their environment.
A bright future for AI
Overall, the panel presented an optimistic outlook for the Chinese AI sector, with company founders confident that they can create solutions that will make the technology a practical tool for businesses. Against this positive backdrop, investors will continue paying close attention to this increasingly important tech theme.
Subscribe and listen to HSBC Global Viewpoint
Discover more on this topic by listening to our podcast on your preferred podcast player. Make sure you're subscribed to HSBC Global Viewpoint to stay connected with our latest episodes.
HSBC 12th Annual China Conference
Get in-depth insights and important takeaways from industry leaders, policymakers, and institutional investors.

© Copyright HSBC Bank 2025. All rights reserved. No part of this document may be reproduced, stored, distributed or transmitted in any form without the prior written permission of HSBC Bank.