Long-term Fund Oversubscribed Over 2x: Behind Desai's IPO as First Medical Imaging Large Model Stock

Why is AI·DeShi Medical AI Platform attracting long-term funds to oversubscribe?

As artificial intelligence technology continues to evolve, the integration of AI and healthcare is gradually moving from the laboratory to clinical applications. Medical AI, as a typical high-barrier sector, is becoming a focus of capital markets due to its commercialization capabilities and technological pathways.

On March 20, Hangzhou DeShi Biotechnology Co., Ltd. (hereinafter referred to as “DeShi”) officially launched its Hong Kong IPO, planning to issue 7,999,200 H-shares at a price range of HKD 95.6 to HKD 112.5 per share.

Before submitting to the Hong Kong Stock Exchange, DeShi had completed ten rounds of financing, raising approximately RMB 397 million in total. During this IPO’s international placement, the company received subscriptions from many long-term funds and achieved oversubscription, with demand exceeding supply by more than twice. In the current market environment, the entry of long-term capital reflects the recognition of the long-term value of medical AI platform companies by the capital market. If the final offering price is set at the upper limit, the company is expected to be included in the Hong Kong Stock Connect, further enhancing liquidity in the secondary market.

Full-Platform Pathway Established, Supporting Commercialization and Market Share Growth

Recognition from capital markets is primarily based on the company’s proven commercialization ability.

Since its establishment in 2016, DeShi has developed a diversified product portfolio, including six medical imaging software and three commercial medical devices. Its core product, AI AutoVision®, focuses on intelligent chromosomal karyotype analysis and was recognized as a “Class III innovative medical device” by the National Medical Products Administration in May 2025.

According to Frost & Sullivan data, based on 2024 sales revenue, DeShi holds a 30.6% market share in China’s chromosomal karyotype analysis field, ranking first.

From an industry landscape perspective, although many participants are involved in the medical imaging AI track, most companies are limited by R&D costs and data barriers, mainly focusing on lung nodules, fundus screening, and other scenarios, resulting in relatively single-use cases.

While point breakthroughs can be quickly implemented, in actual hospital applications, they often lead to system fragmentation and fragmented user experience. In contrast, DeShi has chosen a differentiated path—evolving from single-point tools to a full-platform approach. Its multi-product portfolio and cross-scenario coverage give it a certain scarcity in the current domestic medical AI market.

DeShi’s financial performance further confirms the initial success of this model. In the first three quarters of 2025, the company achieved revenue of RMB 112 million, a year-on-year increase of about 470%. Meanwhile, benefiting from the higher proportion of high-margin technology licensing business and expanded direct sales channels, the company’s overall gross profit margin increased from 42.9% in the same period of 2024 to 75.9%.

Building a Large Model Foundation to Enable Scalable Replication

If commercialization validates the feasibility of the path, then underlying technological capabilities determine the future growth ceiling.

The healthcare industry demands high accuracy, and the direct application of general large models (like ChatGPT) in medical scenarios still faces uncertainties; traditional specialized medical AI models also face high training costs and long R&D cycles.

In this context, DeShi launched the iMedImage® medical imaging base model. As a general visual model with hundreds of billions of parameters, it can cover over 90% of clinical diagnostic scenarios.

More importantly, its advantages in R&D economics are significant. Compared to traditional specialized models, its development cycle, funding input, and training data requirements are greatly reduced (by factors of ten, one hundred, and one hundred respectively). This means the model can be developed faster and cheaper, providing a practical foundation for large-scale replication.

From industry trends, the development of medical AI is resonating with hierarchical diagnosis and treatment policies. Primary healthcare institutions have long faced shortages of high-quality doctors, and devices equipped with high-level AI models are expected to significantly improve diagnostic capabilities at the grassroots level.

Through self-developed base models, DeShi combines technological capabilities with real clinical needs, addressing some core pain points in the industry.

Continued R&D Investment and Valuation Anchored in Long-term Platform Value

In its IPO fundraising plan, DeShi explicitly states it will continue to strengthen its technological and product capabilities. Public information shows that about 49% of funds will be used for R&D and commercialization of its core product AI AutoVision®, and about 20% for enhancing the iMedImage® base model and AI technology.

Ongoing R&D investment is also key to maintaining technological barriers and platform advantages.

From a valuation perspective, DeShi’s current valuation of HKD 8.5 billion to HKD 10 billion is not low, already reflecting market expectations of its “medical AI large model platform” scarcity premium.

However, from a long-term view, the reasonableness of this valuation still depends on two points: first, whether the data flywheel effect can be sustained; second, whether the low sales cost advantage can translate into profit realization.

If the company can maintain its current high growth and continuously expand application scenarios for its models, its valuation is likely to be gradually absorbed during subsequent performance realization.

Reporter: Liu Jinyang Editing: Liu Lei On-camera: Chen Yuanhao Editing: Liu Yu Proofreading: Liu Tian