As AI transitions from "storytelling" to "delivering tangible results," companies that are genuinely rooted in the industry and effectively address pain points are poised to stand out. On January 5, 2026, the China Securities Regulatory Commission (CSRC) issued a filing notice, granting official approval for overseas issuance and listing, as well as a "full circulation" arrangement for domestic unlisted shares, to Hangzhou Deshi Biotechnology Co., Ltd. This signifies that the company, which specializes in medical imaging artificial intelligence, software development, and high-end medical equipment, is set to list on the Hong Kong Stock Exchange and could become the world's first publicly traded company with a "large-scale medical imaging model" as its core asset. Amid the clamor of general-purpose large model companies vying to be the "first IPO of its kind," Deshi Biotech has chosen a more vertical and pragmatic path, focusing on the high-value, high-barrier niche of medical imaging. It employs a unified large model to tackle the most labor-intensive and error-prone aspects of clinical diagnosis. Its narrative is not merely about technological breakthrough but represents a profound case study on how hard technology can be genuinely implemented and create commercial value.
Medical imaging serves as the "eyes" of modern healthcare. Statistics indicate that imaging data constitutes over 70% of clinical diagnostic information. However, the growth in the number of physicians falls far short of the explosive growth in data—between 2021 and 2023, medical imaging data surged by 120%, while the number of radiologists in China increased by only 10%. Under heavy workloads, misdiagnosis rates remain stubbornly high, with clinical misdiagnosis rates for certain imaging diagnoses reaching as high as 21%. Traditionally, the mainstream approach in medical AI has been "one model per task": developing a specific model for lung nodule detection in CT scans, and another for tumor identification in pathological slides. While seemingly precise, this method is costly, time-consuming, and suffers from poor compatibility. Developing a single-disease-specific model often requires investments exceeding hundreds of millions, hundreds of thousands of annotated images, and a timeline of 3 to 5 years, with limited reusability across different devices and scenarios. Deshi Biotech's breakthrough lies in overcoming this predicament with a "general-purpose foundation." Its self-developed iMedImage™ has been recognized by Frost & Sullivan as the "world's largest general-purpose medical imaging foundation model by parameter scale" and is the world's first commercially available cross-modal medical imaging large model. It supports 19 medical imaging modalities—including chromosome imaging, CT, MRI, ultrasound, pathology, and more—covering over 90% of clinical scenarios.
More importantly, this model significantly lowers the barrier to AI implementation. Based on iMedImage™, a hospital or equipment manufacturer can rapidly develop specialized AI models with just a few hundred imaging samples and several weeks of training, compressing the development cycle from the industry average of 3-5 years down to a few months, while reducing R&D investment by over 90%. This democratizes AI, making intelligent diagnostic capabilities accessible not just to industry giants but also to small and medium-sized institutions. Taking chromosome analysis as an example, traditional manual karyotyping accuracy is around 50%, whereas Deshi Biotech's AI system elevates this figure to over 99%. This dual leap in efficiency and precision stands as a powerful testament to its technological value.
Currently, the Hong Kong Stock Exchange is becoming a favored listing destination for AI companies, especially those focused on large models. While several general-purpose large model companies are racing to become the "first mover," capital markets are gradually shifting from a frenzy over pure technological concepts back to rationality. Investors are increasingly focused on key questions: Can the technology be implemented? Are there actual products? Is there stable revenue? As the criteria for evaluating AI narratives shift from "looking at the stars" to "keeping feet on the ground," Deshi Biotech demonstrates a rare level of substance. Unlike many AI companies still in the R&D and cash-burn phase, it has already established a complete commercial loop characterized by "certifications, products, revenue, and market presence." On November 21, 2025, Deshi Biotech successfully won the bid for the "Silver-Medicine Informatization AI Large Model Computing Platform" project at the First Affiliated Hospital of Zhejiang University School of Medicine, marking the official entry of its AI large model solution into the core medical information systems of a top-tier domestic tertiary hospital.
This is merely a snapshot of its commercialization progress. Deshi Biotech's core product, the AI AutoVision® Chromosome Karyotype Auxiliary Diagnostic Software, received the National Medical Products Administration (NMPA) "Class III Innovative Medical Device" designation in May 2025 and has entered the formal registration application stage. Another hardware product, the MetaSight® Automated Cell Microscopic Image Scanning System, has even secured market access certifications from three major regulatory bodies: China's NMPA, the US FDA, and the European Union's CE mark, making it the world's first comprehensively certified automated chromosome scanning device. Certifications are not just permits; they symbolize the authoritative recognition of the technology's reliability and clinical value. Backed by robust products, Deshi Biotech captured 30.6% of the Chinese chromosome karyotype analysis system market in 2024, ranking first and establishing a leadership position. The core driving force behind this achievement is the iMedImage™ large model.
The acceleration in commercialization is directly reflected in strong financial figures. In the first three quarters of 2025, Deshi Biotech's revenue reached 112 million RMB, a staggering increase of approximately 470% compared to 20 million RMB in the same period of 2024, indicating rapid market acceptance of its solutions. Concurrently, the company's overall gross profit margin improved from 42.9% in the same period of 2024 to 75.9%, reflecting enhanced product competitiveness and profitability optimization driven by economies of scale. Furthermore, as of September 30, 2025, Deshi Biotech held a combined total of 250 million RMB in cash and cash equivalents, short-term investments, and unused bank credit lines. This ample cash flow provides solid support for the company's ongoing R&D, market expansion, and business upgrades.
A larger stage is unfolding. The global intelligent medical imaging detection market, starting from $400 million in 2019, is projected to climb to $25.1 billion by 2035. This sustained high-speed expansion is fueled by the deep integration of AI technology into healthcare and the growing demand for efficient, precise diagnostic solutions. As AI large models become deeply integrated with medical imaging, intelligent diagnosis and treatment are gradually becoming a vital component of global healthcare systems. Currently, out of over 3,200 medical testing items, AI assistance has been introduced into 65 imaging-related tests, accounting for about 2%—indicating vast market potential remains. The industry also received a policy tailwind in 2025 when the NMPA introduced policies explicitly supporting innovation in high-end medical imaging and AI medical devices. The true driving force, however, stems from real-world challenges: an accelerating aging population, the concentration of high-quality medical resources in major cities, and the lengthy training周期 for specialist physicians... These structural problems are turning AI-assisted diagnosis from an "option" into a "necessity." In many primary care hospitals, AI is not a luxury but a crucial support system.
Deshi Biotech's journey towards an IPO coincides with a period of deep reflection in capital markets regarding AI investment. While general-purpose large models engage in an internal race for parameters and computing power, investors are beginning to ask more sobering questions: Is the ultimate value of technology found in impressive conversations, or in solving tangible, critical problems related to livelihoods and lives? Deshi Biotech's choice provides an answer: In specialized fields like medical imaging, iMedImage™, though less flashy than general conversational models, effectively addresses real clinical pain points—shortening diagnosis times, reducing misdiagnosis risks, and freeing up physician productivity. This pragmatic technological approach may be more appealing to long-term investors. Capital markets are shifting from "hype-driven speculation" to "value verification." And Deshi Biotech has already demonstrated, through product certifications, market share, and hospital orders, that its technology is implementable, the demand is genuine, and its business model is sustainable. More importantly, iMedImage™, as the world's first commercially available cross-modal medical imaging large model, is not just a technological achievement but also an export of a methodology—it proves that in highly specialized fields, Chinese teams are fully capable of building globally competitive underlying infrastructure.
Deshi Biotech's planned Hong Kong IPO is not merely a capital markets event; it is a signal: In the era of deep AI-healthcare integration, those who can genuinely solve clinical pain points will win the market and the future. As AI moves from "storytelling" to "delivering tangible results," companies truly rooted in the industry and addressing core challenges are ultimately the ones that will emerge successfully. In an environment rife with uncertainty, true certainty stems from a profound understanding of the real world and the resilient capability to translate technology into value. Deshi Biotech's story begins with medical imaging, but its implications extend far beyond. It illuminates a growth path for hard technology: focus, pragmatism, achieving a closed loop, and then riding the waves of opportunity. This journey has only just begun.