Is AI healthcare now approaching its own “DeepSeek Moment” as it rides a new wave of momentum?

When Cathie Wood, often referred to as the “female Buffett,” firmly bets on AI + healthcare, calling it the most underestimated application of AI; When “Wall Street’s oracle on Capitol Hill” Nancy Pelosi places her chips on AI healthcare stocks; When Baichuan Intelligent, once one of the “Six Tigers of AI,” begins focusing on applying large models to the healthcare sector.
In January, former U.S. House Speaker Nancy Pelosi purchased 50 call options for Tempus AI stock, with a transaction value between $50,000 and $100,000. (Source: Nai 500)
AI + healthcare is standing at a critical inflection point where quantitative changes are leading to qualitative breakthroughs: fueled by capital frenzy, technological upgrades, and policy support.
Is AI healthcare now approaching its own “DeepSeek Moment” as it rides a new wave of momentum?
Capital Rush into AI Healthcare: From Scattered Bets to Strategic Focus
Since the start of 2025, the “AI + Healthcare” track has been a bright spot in financing activities.
At the beginning of 2025, China’s AI healthcare sector experienced a surge of concentrated financing deals. Unlike previous “scattergun” investments, capital is now targeting areas with clear clinical value and high technical barriers. The rationale is evident—delivering short-term results while laying the groundwork for long-term gains.
Industry capital is building moats by combining “backing track leaders + securing critical healthcare demand.”
At the end of January, Beijing-based Deepwise Healthcare announced a new round of financing totaling nearly RMB 500 million, co-invested by Legend Capital and Hangzhou’s district-level industrial fund. Additionally, Mirxes (originating in Singapore), which focuses on miRNA-based cancer early screening, secured $40 million in strategic funding from CBC Group to accelerate the commercialization of its cancer liquid biopsy products.
It is also noteworthy that regional industrial capital is becoming increasingly active. For example, Wuxi’s venture capital platform, Xichuangtou, made consecutive investments in AI healthcare startups in early January, including Tushen Zhihui (AI protein design, completed angel+ round), Insilico Medicine (AI drug discovery, received over $100 million in strategic funding), and CyberMedTech (AI-assisted orthopedic diagnostics, completed Series A financing).
Guangzhou Industrial Investment also made a series of investments in AI healthcare firms in early January, including Bo Yin Hearing (AI hearing aid R&D, completed angel+ round) and Xi Mang Medical (brain-computer interface technology). Numerous funds with state-owned backgrounds actively participated, such as Guangzhou Radio Group, the Guangdong-Hong Kong-Macao Greater Bay Area Collaborative Innovation Institute, and the Beijing-Tianjin-Hebei Technology Innovation Center, reflecting policy-driven support for cutting-edge medical technologies.
The capital market’s sharp response to the AI healthcare revolution has created a resonance between primary and secondary markets.
In the first trading week after the Spring Festival, China’s A-share AI healthcare concept sector rose by 8.7%. Winning Health, which partnered with DeepSeek, saw a single-day surge of over 10%. Companies like Chuangye Huikang and Wanda Information, which completed DeepSeek model localization and advanced medical applications, saw cumulative stock price increases exceeding 20%. The Hong Kong stock market moved in tandem, with Yidu Tech integrating the DeepSeek model into its “AI Healthcare Brain” YiduCore, pushing its share price up by 28%—a two-year high.
The “algorithm platform + vertical scenarios + industry collaboration” triad is becoming the standard formula for medical AI investment.
In the primary market, top institutions such as Sequoia Capital and Hillhouse Venture have established dedicated funds. In February alone, financing in the AI healthcare sector reached RMB 4.7 billion, a year-on-year increase of 210%.
AI Healthcare Goes Global, Attracting Capital Inflows
In overseas markets, AI healthcare is also attracting significant investment.
U.S.-based Hippocratic AI, a startup focusing on large healthcare dialogue models, recently completed a $141 million Series B round, bringing its post-investment valuation to $1.64 billion and elevating it to unicorn status. Spain’s Quibim, an AI medical imaging firm, secured $50 million in Series A funding to boost its focus on oncology and neurological disease diagnostics.
In the Asia-Pacific region, Australia’s Harrison.ai announced $112 million in Series C financing in February 2025. According to Crunchbase data, in January 2025, the healthcare and AI sectors secured $9.4 billion and $5.7 billion in funding, respectively, together accounting for 58% of global venture capital activity that month.
AI Healthcare’s Three Major Battlegrounds
As algorithms advance and computing power grows, AI is accelerating its penetration into all aspects of the healthcare industry. Currently, AI’s primary applications in healthcare focus on medical imaging analysis, AI-assisted drug discovery, and clinical decision support/intelligent diagnostics. Each sub-sector is at a different stage of development but is moving from pilot validation to large-scale deployment.
1. Medical Imaging AI: From Assisted Diagnosis to Multimodal Analysis
Medical imaging is one of the earliest and most mature applications of AI in healthcare. AI algorithms can now assist radiologists by automatically identifying lesions and screening for diseases in X-ray, CT, MRI, ultrasound, and other imaging modalities, improving diagnostic efficiency and accuracy.
Several domestic imaging AI companies have secured regulatory certifications and are widely piloted in hospitals. Imaging AI is evolving from single-modality to multimodal fusion and intelligent upgrades. For example, Deepwise Healthcare claims to have provided AI diagnostic and screening solutions to thousands of healthcare institutions, obtaining 14 NMPA Class III medical device certifications, making it a domestic leader in imaging AI.
Internationally, Quibim’s AI-driven imaging platform plays a critical role in the early diagnosis of tumors and neurodegenerative diseases, significantly improving early detection of prostate cancer and Alzheimer’s disease.
However, challenges remain: variations in imaging data quality and resolution across hospitals hinder AI model generalization, requiring repeated training and optimization. The massive scale and high labeling cost of medical imaging datasets, coupled with the limited diversity of academic research data sources, make acquiring high-quality, multi-center datasets a major bottleneck. Moreover, liability and regulatory issues surrounding AI-based diagnosis remain unresolved.
With increasing algorithmic precision and real-world validation, regulatory approval of AI diagnostic indications is expected to grow. AI is poised to assume part of the routine imaging workload, alleviating the shortage of radiologists. Major medical equipment manufacturers will likely embed AI as a standard feature in imaging devices and PACS systems, creating integrated hardware-software solutions.
Additionally, imaging AI may be extended to grassroots healthcare settings, where cloud-based recognition services can empower imaging departments at smaller hospitals, enhancing diagnostic capabilities and promoting resource equity.
2. AI-Assisted Drug Discovery: From Target Identification to Clinical Trials
AI-driven drug discovery is considered one of the most transformative AI applications in healthcare. AI can significantly accelerate processes such as target identification, lead compound design, and clinical trial planning. For instance, Recursion has built a high-throughput experimental and machine learning platform that rapidly screens potential disease-related targets from vast biological datasets, shortening the early-phase candidate screening timeline for drug development.
Major pharmaceutical companies are also investing directly in AI drug discovery: GSK partnered with Exscientia to develop new molecules screened by AI, while Eli Lilly signed an agreement with China’s XtalPi to use AI and automation to screen innovative compounds.
This "pharma + AI" joint R&D model is becoming increasingly common. In the foreseeable future, collaborations among industry, academia, and research institutions will become the norm in AI-powered drug discovery, with pharmaceutical firms providing vast historical data and pharmacological expertise, and AI companies supplying algorithms and computing platforms.
Investors and pharmaceutical firms have high expectations for AI-driven drug development. However, AI drug discovery is still in the early stages of exploration and validation, facing significant technical and industry-specific challenges.
For example, the complexity of biological mechanisms remains a major barrier, as the scientific understanding of physiological systems is still incomplete. Additionally, unlike imaging AI, which can immediately deliver actionable diagnostic outputs, AI-generated drug discovery insights (e.g., suggested compound structures) still require chemical synthesis and biological testing to validate activity and safety—processes that remain time-consuming and costly.
In other words, AI is best positioned as a tool to empower researchers by narrowing candidate pools and providing design inspiration, but the “wet lab” component of drug R&D remains indispensable. Moreover, pharmaceutical R&D data tends to be fragmented and confidential, making smaller, task-specific models a potential future solution.
3. Intelligent Healthcare and Medical Large Models: From Doctor’s Assistant to Patient-Facing Services
The emergence of Medical Large Language Models (MLLMs) in recent years has introduced the concept of the “AI doctor assistant.” These models, trained on vast datasets of medical literature, clinical guidelines, and patient records, are now capable of performing medical Q&A and supporting clinical decision-making.
For example, Hippocratic AI in the U.S. developed the Polaris model, a large-scale language model for medical applications that can provide non-diagnostic guidance, such as medication counseling, through voice interactions with patients.
On January 25, Baichuan Intelligent released its Baichuan-M1-preview model. CEO Wang Xiaochuan revealed that a "Super Doctor Model" will launch in March or April 2025, aiming to provide every Haidian resident with a personal AI healthcare assistant in the first quarter.
However, medical large models still face significant challenges. First, healthcare requires extremely high levels of accuracy—even a 1% error rate could have severe consequences in medical settings. Therefore, such AI is currently confined to auxiliary roles, supporting non-high-risk decisions rather than independent diagnostics.
Second, large models are prone to "hallucinations" (fabricating unreliable content), which is unacceptable in medical contexts. Mitigating this issue requires the integration of knowledge graphs and stronger training supervision to enhance model rigor.
With DeepSeek exceeding expectations and breakthroughs in open-source AI and technical bottlenecks, technologies like few-shot learning and explainable AI will enhance the models’ ability to manage rare and complex clinical cases. The market is moving in lockstep.
As Baichuan Intelligent CEO Wang Xiaochuan recently noted in an interview, “Healthcare AI can’t just show off technical prowess; it must deliver perceived value—saving pharmaceutical companies billions in R&D costs or helping hospitals boost annual revenue by millions. Avoid head-on collisions with tech giants; we’re not here to challenge OpenAI on pure tech or to battle internet giants in general scenarios.”
Ultimately, this points to the core value of AI healthcare: “There are no shortcuts in healthcare. AI must immerse itself in medical records, reagent bottles, and operating rooms to truly wield the scalpel.”