The next two years of clinical AI: our thesis
Where we think capital, talent, and clinical adoption will concentrate as the field enters its next phase.
Consolidation of the base layer
Foundation model providers will further consolidate. Application-layer companies will increasingly build on a small set of clinically fine-tuned bases. The value at the base layer will accrue to a handful of platforms.
Fragmentation at the application layer
The application layer will fragment by specialty, by care setting, and by workflow. This is where most durable companies will be built.
Emergence of clinical operating companies
Hybrid software-and-service companies — those that own both the AI and the clinical delivery — will define the next generation of category leaders. Pure software companies will struggle to maintain differentiation.
The talent bottleneck
The scarce talent is not ML engineers. It is people who understand clinical operations, regulatory strategy, and modern AI simultaneously. Companies that build this talent bench early will define the next decade.
Why the base layer will keep consolidating
The capital and compute requirements to train and maintain a competitive clinical foundation model keep rising, while the number of institutions willing to fund that effort without a clear near-term commercial path keeps shrinking. That combination points toward a small number of base-layer players, most of them already well capitalized, absorbing or out-competing the rest over the next two years rather than a broad field of well-funded independents persisting.
Why the application layer will fragment anyway
Consolidation at the base layer does not imply consolidation at the application layer, because clinical workflows are genuinely heterogeneous across specialties, care settings, and payer arrangements in a way that rewards narrow, deeply integrated point solutions over broad platforms. We expect the number of viable application-layer companies to grow even as the number of foundation model providers they build on shrinks.
The rise of the clinical operating company
A category of company is emerging that does not sell software at all but operates a clinical service — staffing, protocols, and outcomes accountability — with AI embedded as infrastructure rather than as the product. These clinical operating companies capture more value per patient than a pure software vendor because they own the outcome, not just the tool, and we expect several of the most valuable new companies in the space to look this way rather than like a traditional SaaS vendor.
Why talent, not capital, is the binding constraint
Capital is abundant relative to the number of people who can credibly operate at the intersection of clinical judgment, regulatory strategy, and modern AI engineering. That scarcity is becoming the real gating factor on how fast this thesis plays out, more than fundraising conditions or model capability. Founders who invest early in building a team with genuine clinical fluency, not just technical fluency, will compound an advantage that is much harder to replicate with capital alone.
What could slow this timeline down
The biggest risk to this thesis playing out on a two-year clock is not technical but reimbursement-related: payers have been slower than expected to build durable coverage pathways for AI-driven clinical services, and a prolonged stall there would push consolidation and fragmentation timelines out by a year or more. We are watching payer policy signals as closely as we are watching model releases for this reason.
