From embryo selection and cancer diagnostics to remote cardiac monitoring and AI-assisted elderly care, Taiwan’s healthcare startups are moving beyond generalized AI rhetoric and into highly specific medical applications.
Artificial intelligence is often discussed in sweeping terms — productivity, transformation, disruption — but many of Taiwan’s healthcare startups are approaching the technology from a more targeted direction. Rather than building generalized AI, they are applying machine learning, data analysis, and real-time diagnostic and tracking tools to address clinical bottlenecks and introduce patients to both routine and life-saving tools.
At events like AmCham Taiwan’s AI Connect and the Meet Taipei Startup Festival hosted by Business Next Media, founders present products and software designed to support the broader network surrounding patient care, including family members, caregivers, and hospital staff. For physicians, these innovations improve diagnosis, patient oversight, and care delivery, particularly in areas strained by aging populations, rising costs, and clinical inefficiencies.
Anivance AI, a developer of organ-on-a-chip (OoC) solutions, is seeking to tackle a fundamental challenge to advances in healthcare. “Drug development is costly and not very efficient,” says Jack Chen, the company’s head of sales and marketing. “Nearly 90% of drugs fail in clinical trials. We spend years and huge investments only to fail in the end. This cost does not disappear — it is passed on.”
Anivance AI, which presented at the second edition of AI Connect in April, combines OoC with AI-driven analysis to support drug development. Using small microfluid devices containing living human cells, the technology can mimic the structure and behavior of real human organs, allowing researchers to use AI-assisted analysis tools to quantify and monitor how the cells respond to drugs, toxins, or disease conditions in real time. Anivance has so far been able to develop more than 25 organ models, including the brain, lungs, liver, and kidneys.
The OoC methodology “enables earlier and more confident decisions before clinical trials through human-relevant data,” says Chen. The technique provides a way to avoid both the cost and ethical concerns associated with animal testing.
“Animal models still cannot replace human biology,” says Chen. In recent years, U.S. regulators, including the Food and Drug Administration (FDA), have shown increasing openness to alternative methods, including human-relevant models and in-silico approaches. This creates opportunities for cross-sector collaboration aimed at accelerating the adoption of human-relevant testing models and improving the development of targeted treatments, he says.
While Anivance AI focuses on improving drug development at the earliest stages, another Taiwan startup, CancerFree Biotech, is applying similar principles much closer to the patient, where treatment decisions often carry immediate consequences.
“Our technology is simple,” says Po Chen, CEO of CancerFree. “We collect patient samples through our platform, which uses a specialized culture regeneration to grow the cancer outside of the body.” Lab technicians are then able to test various drugs on the patient’s tumor cells, including those derived from lung and colon cancers, rare diseases, and brain tumors, to observe drug responses and offer tailored precision medical services.
While the underlying concept may appear straightforward, Chen notes that maintaining stable, clinically useful approaches is technically highly complex. Laboratory processes traditionally require extensive manual analysis, but the company’s AI software allows technicians and clinicians to extract more information from patient-derived data and improve the reproducibility and consistency of preclinical research outcomes. According to a widely cited 2015 PLOS Biology study, an estimated US$28 billion is lost annually due to irreproducible preclinical research.
CancerFree currently offers these services to patients on an out-of-pocket basis. However, reducing the trial-and-error nature of cancer treatment selection could help reduce the cascading costs associated with ineffective treatments, prolonged care pathways, and the financial and emotional burden carried by families. Notably, CancerFree provides its services free of charge for patients under the age of 12 through its “Little Star” program.
That same emphasis on reducing uncertainty and avoiding costly trial-and-error treatment pathways by modeling and testing outside a patient’s body is also shaping AI applications in reproductive medicine. AB DigiHealth, another presenter at AI Connect, has developed icONE, an AI-driven embryo selection process for in-vitro fertilization (IVF) built on genomic data.
AB DigiHealth CEO Alan Tang says that the average IVF pregnancy success rate still hovers around 40%, with repeated failed cycles creating significant psychological and financial strain for patients.
“Embryo selection still relies heavily on morphology assessments and physician experience, even after embryos undergo pre-implantation genetic testing for aneuploidy (PGT-A),” says Tang.
The icONE technology integrates genomic data from PGT-A together with maternal clinical features to generate embryo-selection predictions to support physician decision-making. Tang says that internal clinical data shows pregnancy rates reaching as high as 81% in icONE-assisted cases, compared with 53.1% in non-icONE groups.
“The platform has obtained FDA Class I listing as a software medical device and is positioned as an objective, data-driven decision-support tool rather than a replacement for physician judgment,” he says.
AI and aging
Continuous cardiovascular monitoring is becoming increasingly important in preventing sudden health events that can rapidly compromise independent living and increase long-term care needs — a crucial consideration in light of Taiwan’s super-aged population. According to National Health Insurance Administration data, Taiwan recorded more than 471,000 stroke-related hospital admissions in 2022, with people aged 61 to 80 accounting for more than half the cases.
RadiRad, another presenter at AI Connect, is attempting to address this growing burden through RadiHeart, a wearable AI-driven device for consistent blood pressure assessment. Amber Lin, RadiRad’s business development manager, notes that hypertension remains one of the strongest predictors of ischemic stroke — the most common form of stroke, accounting for approximately 87% of all cases.
Referring to traditional cuff-based, manual devices, she says “You’re supposed to bring it with you after your meal or after you work out, but we barely see that happening.” Unlike conventional cuffs that capture only isolated readings, RadiHeart is designed for ongoing, real-time recording through a wearable finger cuff connected to phones, tablets, and other devices via Bluetooth. Using PPG sensing technology, the compact portable device captures high-resolution physiological data while simultaneously measuring blood pressure, pulse rate, SpO2, and heart rate variability.
Lin says patient compliance remains a major challenge for traditional cuff-based practices. She describes the wearable device as better suited to become a natural part of one’s daily routine.
RadiHeart has undergone IRB-approved clinical validation trials in Taiwan and achieved claimed accuracy rates of up to 99%. The company is also developing broader AI-driven cardiovascular management tools, including stroke risk prediction, arterial stiffness assessment, arrhythmia detection, and heart failure risk monitoring.
While RadiRad focuses on physiological tracking, startups at Meet Taipei last November approached aging-related healthcare challenges through safety infrastructure and caregiving support designed for home environments.
For its part, Stone Co., a Taiwanese IoT and smart-care technology company focused on elderly and long-term care, has developed OTTalk PLUS under its OTVerse home setup, which is for elderly individuals living independently. Beyond functioning as a one-touch emergency alert device, it constantly tracks movement, mobility, and daily activity patterns throughout their home.
The setup combines AI-assisted activity analysis, real-time alerts, remote connectivity, and mobile integration, helping caregivers and family members identify unusual inactivity, behavioral changes, or potential emergencies before situations escalate.
“For seniors, the greatest fear isn’t aging — it’s the isolation of an accident with no way to call for help,” Stone Co. has written. “The OTVerse One-Touch Group Call console is currently the only bridge in the world that instantaneously connects ‘Home’ with ‘Protection.’”
For its part, startup AI-Supporter is addressing one of the more physically demanding aspects of long-term care. Its AI-enabled smart excretion management is designed for bedside use by patients with limited mobility.
The entire package combines AI recognition cameras, wearable pads, suction systems, automated washing functions, and cloud-connected tools to automatically detect urination and defecation, perform cleaning and drying, and record excretion frequency, timing, and volume.
Beyond hygiene management, it reduces caregiver workload while supporting more continuous patient oversight. Product features — including real-time abnormal alerts, electronic medical record integration, Bluetooth and Wi-Fi connectivity, and repositioning reminders — reflect a broader effort to digitize labor-intensive caregiving tasks that traditionally rely heavily on manual observation and physical assistance.
Although visually unconventional, the product is a response to a growing healthcare challenge within aging societies: maintaining patient dignity, hygiene, and care standards while reducing physical and logistical burdens.
Scale and visibility
Despite the technical sophistication on display across both AI Connect and Meet Taipei, many of the participating companies still face challenges that extend well beyond product development itself. These include securing the visibility, funding, partnerships, regulatory support, and healthcare integration needed to scale beyond pilot-stage adoption.
At the two events, several company representatives spoke directly about those pressures. CancerFree Biotech says it is “actively looking for fundraising,” while AB DigiHealth plans to establish a U.S. entity and pursue a US$3 million fundraising round tied to expansion into California, Texas, and New York. RadiRad similarly describes ongoing efforts to expand internationally through exhibitions, partnerships, and new overseas offices.
Anivance AI stresses the importance of broader institutional coordination and scalability. “We are building an ecosystem composed of regulatory institutions, pharma, clinical, academia, and technology partners, to accelerate validation and adoption of human-relevant models,” says Jack Chen. “Our next focus is scaling and standardization.”
The emphasis on exposure and ecosystem-building is also reflected in the structure of AI Connect itself. In opening remarks, Linda Yu, associate manager of events and marketing at AmCham Taiwan, described the gathering as “a very good opportunity for the entrepreneurs to meet each other and discuss possibilities for investment.”
That same emphasis on commercialization support and international scaling was visible at Meet Taipei through Qualcomm’s Innovation in Taiwan Challenge (QITC), an initiative launched in 2019 to support Taiwanese startups by providing technical resources, business mentorship, IP training, and financial grants.
According to Qualcomm, the program has incubated 69 startups, hosted more than 1,600 mentor sessions and over 100 IP and business workshops, while participating startups have collectively raised more than US$200 million in capital. Qualcomm also reports that more than 500 patents have been filed through participating teams, with over 300 of the patents granted globally.
“What Qualcomm is truly doing is business enablement, helping teams understand how to bring ideas or products to market,” one participating founder stated. Another described QITC as “the accelerator with the richest resources” their company had encountered, while others highlighted support related to patent filing, fundraising guidance, hardware integration, and access to Qualcomm’s chip platforms and technical teams.
Qualcomm executives similarly emphasized the role of ecosystem integration and commercialization. “The convergence of edge AI and generative AI is paving the way for a new era of AI applications,” said Sudeepto Roy, vice president of engineering at Qualcomm Inc. and lead of Qualcomm’s Global Ecosystem Development Program, describing how startups are increasingly expanding AI applications into practical deployment environments.
Still, Taiwan’s startup community, despite its engineering capability, continues to face commercial realities. Many of the technologies presented at AI Connect and Meet Taipei demonstrated strong technical promise, but transforming those innovations into a widely adopted healthcare infrastructure remains a far more difficult process.
Clinical validation does not automatically translate into hospital adoption. Regulatory approval does not guarantee reimbursement. Pilot programs do not necessarily become scalable healthcare systems. And for many startups operating in highly regulated medical environments, the period between technical proof-of-concept and sustainable commercialization can be both financially and operationally difficult to survive.
Remote, real-time care
As Taiwan’s healthcare startups push deeper into remote monitoring and decentralized care, several companies are innovating to extend clinical oversight beyond hospitals and clinics.
At the Meet Taipei Startup Festival last November, VitalSigns Corp. presented SIP, a wearable ECG platform built around a lightweight adhesive sensor called VSH101. The patch continuously records heart activity and transmits data through mobile devices and cloud-connected software, supporting AI-assisted analysis and real-time abnormality detection. The company positions SIP for use across hospitals, home care, long-term care, and rural healthcare settings.
In another example, startup Cancell is focusing on extending cancer care through digital assessment and communication tools. The company describes its software as Taiwan’s first ePRO remote care digital framework for managing cancer-treatment side effects. Patients and caregivers record symptoms through a mobile application, while AI performs severity analysis and risk prediction to help medical personnel intervene earlier when conditions worsen.
The PATHOscope — developed by another Taiwan company, mesoView, to facilitate remote care from the pathology side — performs rapid digital pathology imaging using fresh tissue samples without requiring the freezing and slicing processes typically associated with conventional frozen section analysis. Integrated with the company’s mesoSync telepathology platform, it allows pathologists to remotely view and collaborate on high-resolution tissue images in real time while remaining compatible with standard H&E staining workflows.
