The author, a medical doctor by training, advises GenInnov Funds on biotech strategy and health innovation.

Artificial Intelligence (AI) has captured the excitement of the healthcare industry and has been touted as the panacea that will revolutionise everything from diagnostics to drug discovery. Yet, while many aspects of healthcare will continue to see GenAI progress, for an investor, not all verticals are necessarily equally monetizable.

Of the five key themes in the healthcare business, it is AI-driven medical devices and instrumentation products that are likely to see the fastest and most impactful near-term investment opportunities.

‘Medical instrumentation/devices’ is a high-potential vertical for actionable healthcare investment returns 

Medical instruments and devices across multiple specialties have been and will continue to be infused with AI. Furthermore, because mature companies innovate such medical devices, patient monitors, and robotic surgical tools in an established space such as Medtronic, Intuitive Surgical, Mindray – these companies already have a foothold in the distribution and marketing networks. This will provide investors a valuable proposition for near-term returns

In July 2023, Intuitive Surgical launched its first digital tool, harnessing AI to enable surgeons to study their procedure data and adapt approaches to achieve better results for patients. Called Case Insights, the technology works with data from Intuitive’s surgical robots (da Vinci 5) and hospitals to identify correlations between surgical techniques, patient populations, and patient outcomes, and develop objective performance indicators.

Unlike other verticals, such as diagnostics, where the adoption cycle is slow and complicated, the medical instruments’ replacement cycle is comparatively shorter, as healthcare providers look to upgrade their products amidst rising patient caseload turnover. Companies such as Shenzhen-headquartered Mindray^[1] benefit from their global reach and trusted reputations amongst their end users, making it easier for them to integrate AI into their existing products and expand market share to scale quickly. Some of their boldest innovations target emerging markets such as in Latin America, allowing for high quality medical advances to meet otherwise neglected patient needs.

With shorter replacement cycles, strong business networks and moats, medical device companies that leverage AI offer investors one of the strongest opportunities to ride the growth story in healthcare.

Barriers to achieving clinician acceptance of emerging technology in healthcare: adoption headwinds for AI-guided diagnostics and clinical workflow processes

Healthcare is undoubtedly a high-stakes field, and regulators are understandably cautious. AI-driven tools must undergo rigorous validation before they can be used in clinical settings. This will take time, additional data, clinical validation, and significant investment. Regulators argue that such rigour is necessary to address existing weaknesses in the emerging field and safeguard patients and healthcare professionals’ trust.^[2]

In verticals where diagnostic testing results determine downstream actions, clinicians are hesitant to adopt AI due to concerns about the reliability and explainability of algorithms (black boxes). Healthcare professionals who have spent years honing their diagnostic skills are not yet comfortable relying on a machine to make critical decisions.

In a March 2024 study,^[3] on physicians’ behavioural intentions towards AI-based diabetes diagnostic interventions – perceived usefulness, perceived ease of use and subjective norms directly impact a doctor’s behavioural intention to use AI technologies. Furthermore, perceived risk associated with AI technologies exerts a negative influence. In a systematic review^[4] of using machine learning and deep learning in MRI for brain tumor diagnosis, Satushe et al identified challenges including dataset bias, model interpretability issues, and computational limitations in AI-based brain tumor diagnosis.

This is also true for improving clinicians’ workflow when it comes to case note-taking, extracting patient information, and providing predictive analysis for insurance and billing using AI. Whether it be in LLM-mediated data extraction of clinical records^[5], or how individuals have begun to overtrust AI-generated medical advice despite low accuracy ^[6] - clinicians’ use case continues to be affected by change management resistance.

This is a fragmented space where the business moats of emerging companies remain difficult to pin down, given that AI models continue to evolve. Competition is fierce. Whilst there may be advantages today, competitive advantages are eroded quickly, and it is difficult to predict which winner could achieve a defensible market share.

AI-driven drug discovery does not negate the key milestones of drug approval, testing, and adoption  

Beyond diagnostics, another area where AI has captured attention is in novel drug discovery. The idea that AI can help identify new drug candidates, predict molecular interactions, or even design molecules has caught the attention of many in the biotech industry.

Many pharmaceutical companies have begun to focus on AI-driven drug discovery, and it is tempting for investors to look for research teams that have several new AI-discovered drug molecules entering clinical trials. Recent advancements^[7] in using AI include deep learning, with the identification of drug targets, prediction of their structure, estimation of drug-target interaction, estimation of drug-target binding affinity, design of de novo drug, prediction of drug toxicity, estimation of absorption, distribution, metabolism, excretion, toxicity, and estimation of drug-drug interaction.

Yet, the path to monetisation is unclear. Much of AI in drug discovery is still in the theoretical phase. While AI tools can predict which compounds may be effective, they do not circumvent established drug approval protocols.

One major issue is the availability and quality of data. AI models typically require large, high-quality datasets for effective training^[8]. In many cases, data may be scarce, inconsistent, or of low quality, which can compromise the reliability and accuracy of AI-generated outcomes^[9]. Ethical concerns, particularly around fairness and bias, present significant hurdles. If the data used to train AI models is biased or unrepresentative, it can lead to inaccurate or discriminatory predictions^[10]. Addressing these ethical considerations is crucial for ensuring that AI contributes to the development of new, fair, and effective therapeutic treatments.

Drug discovery remains a labour-intensive and lengthy process. The bulk of AI-driven breakthroughs in this space are in the realm of pre-clinical studies. Only a small fraction of potential drug candidates ever make it to clinical trials, and even fewer make it to market.

In January 2025, the Food and Drug Administration (FDA) issued a draft document “Considerations for the Use of Artificial Intelligence (AI) To Support Regulatory Decision-Making for Drug and Biological Products^[11].” Whilst it is undoubtedly good news that regulators aim to optimise the process for new therapies, investors may find themselves in an unwelcome position of waiting for long-term results, despite the high upfront costs of AI investments in the drug discovery, trial, and analysis phases. AI may hasten the development process, but not necessarily yield better odds when clinical trial failure rates stay at 90 percent^[12].

Therefore, while AI promises to revolutionize the field eventually, the truth is that monetization in this space sees bottlenecks at the clinical efficacy and effectiveness phases. We are unlikely to see a change to the risk-return payoff time horizon for investors.

Solutions need to be plugged into the broader healthcare system of care

In healthcare, principles of patient privacy and data protection are sacrosanct. Many health apps – such as those offering therapy and counseling – operate more like platforms, similar to gig marketplaces, rather than scalable clinical solutions. For users, adoption may prove a novelty. Still, it will ultimately be met with suspicion or outright rejection by doctors or regulators who may be deeply uncomfortable with how their patients’ data are used behind the scenes. There is also limited connectivity between such apps and the larger integrated healthcare systems of clinics and pharmacies.

Therefore, healthcare apps that purport to monetise mental health therapy via quick 30-minute sessions find themselves disjointed from the broader mental health continuity of care.  Investors betting on these platforms for sustained margins and defensible growth may be disappointed with where to place their bets in a space where newer products are emerging every quarter, and switching costs are being lowered due to the lack of integration with actual physicians and therapists. The ethical scrutiny also places such app companies in a situation where they may have to defend and reassure users of their data privacy and AI-driven models in the years ahead.

Conclusion: Hype versus tangible benefits of AI in medicine

For investors in the AI space, there is always a need to distinguish between hype and genuine investment potential. From a fundamental analysis point of view, medical devices are already infused with AI today. Investors will find it easy to bet on companies rolling out such products, as they are already in a durable market position.

Even within the field of medicine itself, medical devices and instruments are evolving in multiple directions. From AI-enhanced robotic surgical tools to wearable health monitors, the landscape is diversifying in ways that were once considered distant possibilities.

In a 2023 literature review^[13] by Ahmed et al, the study authors highlighted barriers in six key areas: ethical, technological, liability and regulatory, workforce, social, and patient safety barriers. Such barriers will least impact on medical devices and instrumentation.

In developed economies such as the United States and Europe, the infrastructure and resources are already in place for faster adoption of AI medical devices. Conversely, in emerging markets such as Africa and parts of Asia, companies like Mindray achieve a foothold by providing value-for-money products that can be implemented at scale.  

The growth story for healthcare AI investors is in medical devices. By understanding the macro-industry challenges of rising patient loads and cost pressures, we can further see how AI-guided medical instruments can mitigate the problems of today.

Investors who identify these companies can capture value from the patient benefits of tomorrow.

^[1] Mindray. (2025, May 24). Mindray showcasing its commitment to smart healthcare technology in Brazil. Mindray. https://www.mindray.com/en/media-center/news/mindray-showcasing-its-commitment-to-smart-healthcare-technology

^[2] Derek L G Hill, AI in imaging: the regulatory landscape, British Journal of Radiology, Volume 97, Issue 1155, March 2024, Pages 483–491, https://doi.org/10.1093/bjr/tqae002

^[3] Roy, M., Jamwal, M., Vasudeva, S. et al. Physicians behavioural intentions towards AI-based diabetes diagnostic interventions in India. J Public Health (Berl.) (2024).

^[4] Satushe, V., Vyas, V., Metkar, S., & Singh, D. P. (2025). AI in MRI brain tumor diagnosis: A systematic review of machine learning and deep learning advances (2010–2025). Chemometrics and Intelligent Laboratory Systems, 263, 105414. https://doi.org/10.1016/j.chemolab.2025.105414

^[5] Azar, W. S., Junkin, D. M., Hesswani, C., Koller, C. R., Parikh, S. H., Schuppe, K. C., Williams, N., Nethala, D., Mendhiratta, N., Kenigsberg, A. P., Turkbey, B., Merino, M. J., Zaki, G., Cortner, J., Gurram, S., & Pinto, P. A. (2025). LLM-mediated data extraction from patient records after radical prostatectomy. NEJM AI, 2(6), AIcs2400943. https://doi.org/10.1056/AIcs2400943

^[6] Shekar, S., Pataranutaporn, P., Sarabu, C., Cecchi, G. A., & Maes, P. (2025). People overtrust AI-generated medical advice despite low accuracy. NEJM AI, 2(6), AIoa2300015. https://doi.org/10.1056/AIoa2300015

^[7] Chakraborty C, Bhattacharya M, Lee SS, Wen ZH, Lo YH. The changing scenario of drug discovery using AI to deep learning: Recent advancement, success stories, collaborations, and challenges. Mol Ther Nucleic Acids. 2024 Aug 8;35(3):102295. doi: 10.1016/j.omtn.2024.102295. PMID: 39257717; PMCID: PMC11386122.

^[8] Vamathevan J, Clark D, Czodrowski P, Dunham I, Ferran E, Lee G, Li B, Madabhushi A, Shah P, Spitzer M, Zhao S. Applications of machine learning in drug discovery and development. Nat Rev Drug Discov. 2019 Jun;18(6):463-477. doi: 10.1038/s41573-019-0024-5. PMID: 30976107; PMCID: PMC6552674.

^[9] Gómez-Bombarelli R., Wei J.N., Duvenaud D., Hernández-Lobato J.M., Sánchez-Lengeling B., Sheberla D., Aguilera-Iparraguirre J., Hirzel T.D., Adams R.P., Aspuru-Guzik A. Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules. ACS Central Sci. 2018;4:268–276. doi: 10.1021/acscentsci.7b00572. 

^[10] Silvia H., Carr N. When Worlds Collide: Protecting Physical World Interests Against Virtual World Malfeasance. Michigan Technol. Law Rev. 2020;26:279. doi: 10.36645/mtlr.26.2.when.

^[11] DLA Piper. (2025, January). FDA releases draft guidance on use of AI. DLA Piper. https://www.dlapiper.com/en/insights/publications/2025/01/fda-releases-draft-guidance-on-use-of-ai

^[12] Dowden H., Munro J. Trends in clinical success rates and therapeutic focus. Nat Rev Drug `Discov. 2019;18:495–496. doi: 10.1038/d41573-019-00074-z.

^[13] Ahmed MI, Spooner B, Isherwood J, Lane M, Orrock E, Dennison A. A Systematic Review of the Barriers to the Implementation of Artificial Intelligence in Healthcare. Cureus. 2023 Oct 4;15(10):e46454. doi: 10.7759/cureus.46454. PMID: 37927664; PMCID: PMC10623210.