Artificial Intelligence in healthcare is often hailed as an “overnight success,” but the reality is that this revolution has been in the making for over 80 years. As technology matures and finds real-world applications, demand for AI-driven solutions in healthcare has skyrocketed, creating a fertile environment for startups and established firms alike to innovate and redefine the industry.
The Maturation of Healthcare AI
Healthcare AI is not merely a fleeting trend but a burgeoning reality, evidenced by the surge of new entrants and increasing demand from well-established incumbents. CEOs from Big Tech, Big Pharma, and healthcare providers all view AI as a massive opportunity for market expansion and innovation.
Contrary to previous waves of short-lived techno-optimism, the current momentum is backed by data. According to a recent survey, 70% of payers and providers are implementing generative AI, and three-quarters have bolstered their IT investments in the past year. Nearly two-thirds of biopharmaceutical executives are constructing use cases in both commercial and R&D settings, with 60% setting immediate ROI expectations.
Laying the Groundwork
The foundation for this AI revolution in healthcare has been laid over decades. Efforts to digitize clinical and administrative data have transformed healthcare into a data-rich environment ready to support advanced AI applications. Electronic health record (EHR) adoption has created vast datasets of text, images, lab values, and other machine-readable data, serving as the “fuel” for AI systems.
The Confluence of Conditions
Several factors have converged to make this moment possible. For one, the underlying technologies have matured significantly, enabling real-world applications. In 2024, almost 40% of healthcare tech startups that raised capital are commercializing AI technologies. Stakeholder sentiment is also shifting, with a majority of healthcare leaders and physicians expecting AI to significantly enhance workflows.
Economic challenges such as inflationary costs in hospitals, rising medical loss ratios for payers, and diminishing R&D returns for biopharma further underline the need for innovative, AI-driven solutions. Mature and adaptive regulation is now imperative to ensure the viability and success of these technological advancements.
Reimagining Healthcare Delivery
AI is positioned to act as a potent catalyst in the re-platforming of healthcare, owing to decades of pioneering work in informatics and AI research. From the early days of informatics in the 1960s and 1970s to the eras of predictive AI, deep learning, and now generative AI, the healthcare landscape has been progressively prepared for this transformation. Entirely new academic journals, like NEJM AI, have emerged to bridge the gap between research and practical implementation.
However, translating high-performing models into commercial products requires more than just advanced algorithms. Effective delivery methods and business models tailored to the specific needs of the market are crucial to achieving real-world impact.
Future Prospects
The current landscape, characterized by heightened funding and rapid technological advancements, is just the tip of the iceberg. As AI research in healthcare continues to accelerate, both startups and incumbents are likely to benefit over the next three to five years. The market’s growing understanding of AI’s potential and limitations will drive the introduction of increasingly refined, novel, and high-impact solutions.
Six years after launching our Deep Health investing thesis, we aim to update our perspective on healthcare AI, focusing on the drivers of new opportunities and the challenges AI companies may face. We will delve into healthcare AI modalities, business models, and market opportunities, emphasizing strategies for overcoming hurdles and achieving defensibility.
The Drivers of Opportunity in Healthcare AI
Data Digitalization: A Treasure Trove for AI
Healthcare generates approximately 30% of the world’s data, yet up to 97% of data produced by hospitals remains unused. The digitization of clinical data via EHRs has made vast amounts of information machine-readable, laying a solid foundation for AI applications. This digitization has also accelerated cloud adoption, pivotal for AI implementation. Companies like SmarterDx and Plenful are leveraging AI to unlock value from this data, improving hospital revenue integrity and pharmacy workflows, respectively.
From Proteomes to Populations
Healthcare AI has progressed to model health at multiple levels, from single proteins to entire populations. New data types are becoming available, and regulations are bringing additional datasets online. Effective use of this data can create multidimensional AI products addressing a wide range of complex biomedical and healthcare use cases. The ability to combine and analyze multiple datasets sets healthcare up as the “holy grail” for multimodal AI development.
Translating Research into Real-World Applications
Decades of AI research are finally translating into practical, scalable products. The number of FDA-approved AI-enabled devices has multiplied 30-fold from 2014 to 2024, signifying a critical inflection point in healthcare AI. Each new research breakthrough holds the potential to solve real-world problems, urging founders and investors to engage deeply with emerging developments.
Challenges Ahead
While the opportunities are vast, healthcare AI startups face several significant challenges:
Navigating Complex Markets
The U.S. healthcare system is notoriously difficult to navigate due to complex regulations and misaligned incentives. AI companies must anticipate resistance and counter with innovative solutions. The checks and balances within the system are also being supercharged by AI, setting the stage for a rapid evolution of both innovation and pushback.
Regulatory Hurdles
Healthcare’s dense regulatory landscape creates high barriers to entry for startups but also serves as a moat for established incumbents. The evolving regulatory environment necessitates that startups remain agile and actively engage in advocacy and lobbying to influence policy in their favor.
Market Fragmentation
Despite the industry’s massive size, the healthcare market comprises thousands of smaller segments, each with its own unique challenges. Success requires identifying and targeting specific segments with substantial TAM, navigating limited budgets, and demonstrating clear ROI.
Strategic Path Forward
To maximize value, healthcare AI companies need to consider modality-business model-market fit, embrace multimodal AI, and develop vertical-specific infrastructures such as cybersecurity, data generation, benchmarking, and model monitoring.
Investment Outlook
Our investment focus includes six core technology themes and several sub-verticals with the potential for $1 billion+ TAM. Companies must aim for platform potential, positioning themselves at critical junctures where valuable data is generated and impacting workflows across the healthcare ecosystem.
With over $1 billion dedicated to AI investments, we at Bessemer are committed to supporting innovative leaders in healthcare AI, creating enduring change and substantial impact in the healthcare sector. As the industry evolves, the future holds boundless opportunities for those ready to reimagine and reinvent.