Mapping AI startup investment and innovation in healthcare using a five-tier AI systems complexity framework

Abstract Artificial Intelligence (AI) is reshaping healthcare through advances in diagnostics, treatment, and operations, yet the startup ecosystem driving this transformation remains underexplored. Analyzing 3,807 AI health startups founded between 2010 and 2024, this study applies a five-tier framework of AI systems complexity to classify ventures by medical domain, AI systems level, funding, geography, and team composition. Nearly two-thirds of AI investments focus on clinical decision support, drug discovery, and diagnostics, domains associated with higher-complexity deep-learning systems, while areas such as mental health, public health, and rehabilitation attract less AI venture capital, reflecting scalability and data limitations rather than a lack of need. Startups remain concentrated in high-income countries, and founding teams are predominantly technical and business-oriented, with limited clinical representation and gender diversity. By linking these empirical patterns to the five-tier framework, we show how AI systems complexity shapes innovation pathways, offering a foundation for more equitable, evidence-driven digital-medicine ecosystems.