In Depth: DeepSeek Integration Gives China’s Health Care Stocks Shot In the Arm

[para. 1][para. 3] The integration of DeepSeek's artificial intelligence (AI) models into Chinese health care is gaining momentum, with more than 30 mainland-based health care companies adding these models to their operations. This trend comes as medical AI technology finds broader acceptance, prompting a surge in related stock prices. DeepSeek has become particularly influential through its globally recognized AI models, released as open-source tools, thereby offering free access to startups and tech giants in hopes of spurring innovative advancements in medical technology.

[para. 3][para. 4] The companies embracing DeepSeek's AI models operate in various fields, including drug research, clinical diagnosis, chronic disease management, and genetic testing. Significant adopters include Hengrui Pharmaceuticals Co. Ltd. and Yunnan Baiyao Group Co. Ltd., among others. These companies seek to enhance efficiency and services by integrating DeepSeek's models into their existing frameworks.

[para. 5][para. 6] Considering its cost-effective development strategy, DeepSeek launched its R1 large language model (LLM) in January, competitive with global counterparts like those from OpenAI and Meta Platforms Inc., but developed at a significantly lower cost. This move has further fueled the interest of health care firms seeking to capitalize on DeepSeek's innovative approaches.

[para. 8][para. 9] The surge in AI-related health stocks, notably in Hong Kong and on the mainland, reflects the positive market response, with companies like Guangzhou LBP Medicine Science & Technology seeing a near doubling of their stock price within a month. Berry Genomics and others have also witnessed substantial gains, reinforcing confidence in the potential impact of AI in health care despite the previous downturn when COVID-related health care demands dropped.

[para. 10][para. 11] Upon expanding the narrative, analysts acknowledge skepticism from hospitals and regulators regarding AI's accuracy and usefulness. However, reports like those from Huafu Securities and Caitong Securities suggest that DeepSeek could overcome challenges like the high development costs and data limitations that have historically plagued the medical AI sector.

[para. 12][para. 13] DeepSeek's open-source strategy represents a paradigm shift, making advanced AI tools accessible to a wide range of companies rather than restrict them to tech giants. This could have a transformative effect on the industry, as noted by Yang Ruirong of Marathon Venture Partners. Prior to DeepSeek, concerns about proprietary models potentially overshadowing efforts in medical AI development dissuaded investment.

[para. 15][para. 16] Additionally, as of early March, over 100 top-tier public hospitals in China have experimented with integrating DeepSeek's AI into various applications, potentially assisting in clinical situations, patient record generation, and public education about diseases.

[para. 18][para. 20] Despite these advancements, experts still caution that AI cannot fully replace the adroit elements of human diagnosis. Recent studies highlight the relative accuracy of DeepSeek's models in specific metrics, such as the Deauville score for lymphoma patients, but still acknowledge a considerable gap compared to human clinicians.

[para. 21][para. 22] The AI field grapples with acceptance barriers due to concerns over accuracy, stalling large-scale commercialization and the inclusion of medical AI into China's national health insurance system. Current regulatory interpretations allow AI to augment but not replace human diagnostic judgement, indicating a continued need for careful integration into medical services.