China’s AI push accelerates with DeepSeek’s breakthrough technology

DeepSeek’s recent breakthroughs have ignited a nationwide push in China to implement its large language models (LLMs) across various sectors, including hospitals and local governments, as Beijing seeks to strengthen its dominance in generative AI.

Since the debut of DeepSeek's R1 model in late January, there has been a swift uptake of the technology, with companies and state bodies rallying behind this new national AI leader, Caliber.Az reports via the Financial Times.

Key players such as major cloud providers, multiple car manufacturers, local governments, hospitals, and state-owned enterprises (SOEs) are all integrating DeepSeek’s models. This rapid adoption is particularly remarkable within traditionally conservative institutions. An SOE tech supplier explained that while AI has long had government backing, DeepSeek has accelerated the rollout of LLMs in government departments and SOEs, sparking a nationwide drive to advance Chinese AI.

The affordability of deploying DeepSeek’s models has played a crucial role in this widespread adoption. China AI expert Adina Yakefu from Hugging Face highlighted that DeepSeek is “transforming the industry landscape” by lowering barriers to AI development through its open-source approach, knowledge distillation techniques, and cost-efficient training solutions.

Hospitals such as Jilin University Hospital in Changchun are implementing DeepSeek-powered diagnostic tools that create treatment plans based on the hospital’s data and medical guidelines. Jinxin Women and Children’s Hospital has developed a tool for patients to track their ovulation and create personalised fertility plans. Additionally, public hospitals in Hubei are using DeepSeek to resolve disagreements between doctors on treatment plans by acting as an impartial third-party consultant.

In cities like Chengdu, Hangzhou, and Wuhan, less complex applications, such as digital nurses guiding patients and explaining medical reports, are also being deployed. However, some experts caution against assuming that all announcements reflect real, substantial progress, as some entities may be promoting their use of DeepSeek for investor attention rather than meaningful application.

Despite these challenges, the rapid adoption of DeepSeek represents a significant shift. In the past, institutions like government agencies and hospitals hesitated to introduce generative AI due to concerns over potential failures. However, DeepSeek’s swift integration signals a change in mindset.

Local governments such as those in Jinan and Hangzhou have launched citizen chatbots powered by DeepSeek for tasks ranging from tax inquiries to birth certificate requests. Shenzhen’s Futian district has also rolled out AI agents based on DeepSeek, including a document generator for law enforcement. Beijing’s strong backing of DeepSeek is further accelerating this adoption, with the company’s founder meeting with President Xi Jinping, and major companies like Tencent, BYD, and Great Wall Motor incorporating its models.

Despite the widespread usage, DeepSeek is not capitalizing financially, as its models are available for free download and can be run on public or private servers. This model benefits cloud providers like AliCloud and Huawei Cloud. Industry experts predict that the surge in demand for DeepSeek’s technology will drive upgrades in cloud computing and AI infrastructure.

By Vugar Khalilov