Why We Bet on Health Technology at the Seed Stage: The Curevstone Investment Thesis

By Alejandro Reyes, Managing Partner · February 12, 2025 · 12 min read

Every venture capital firm claims to back the future. At Curevstone Capital, we have made a deliberate choice about which corner of that future we believe in most deeply — and that choice is health technology. Not because healthcare is large (though it is — $4.5 trillion in the United States alone, over $10 trillion globally). Not because it is growing (though it is). But because it is, in most essential ways, still broken — and because the tools required to fix it have finally arrived.

This essay is an attempt to articulate our investment thesis as clearly and honestly as we can. We want founders to understand not just what we invest in, but why — because the best founders we work with are not simply building companies that fit a category. They are building from a shared conviction about how the world is about to change.

The Scale of the Opportunity

American healthcare consumes 17.3% of GDP — nearly twice the OECD average — while producing health outcomes that rank near the bottom of wealthy nations on measures like life expectancy, maternal mortality, and chronic disease burden. The gap between expenditure and outcome is not a mystery: it is, in large part, a technology and coordination failure. The average US hospital still runs more than 16 separate electronic health record systems, frequently unable to communicate with one another. Physicians spend 49% of their working hours on documentation. Drug discovery takes 12 to 15 years from target identification to FDA approval, at an average cost of $2.6 billion per approved drug. And the vast majority of that cost goes toward failures — the 90% of candidates that never reach approval.

This is not a market waiting for incremental improvement. It is a market waiting for transformation. And transformation at this scale — in a sector this complex, this regulated, this consequential — does not happen all at once. It happens through a series of precisely targeted breakthroughs, each of which creates the conditions for the next. This is why we invest at the seed stage: we want to be present at the beginning of those breakthroughs, when conviction is required and capital is scarce.

Why Health AI Is Different

The general-purpose AI wave that began in 2020 and accelerated dramatically with the release of GPT-4 in 2023 has already produced significant value across industries. But the application of AI to healthcare is categorically different — in both its difficulty and its potential — for several reasons.

First, the data problem. Healthcare data is simultaneously the most valuable and the most inaccessible training resource in any industry. Clinical notes, imaging studies, genomic sequences, wearable telemetry, claims data, pathology reports — the information contained in these sources would, if properly assembled, represent the most complete picture of human biology ever compiled. But this data is fragmented across institutions, stored in incompatible formats, governed by HIPAA and a patchwork of state regulations, and frequently structured around billing codes rather than clinical reality. Companies that find legitimate paths to this data — through partnerships with health systems, through patient-consented research platforms, through synthetic data generation that preserves statistical properties while removing identifiable information — build moats that are nearly impossible to replicate.

Second, the validation challenge. An AI system that recommends the wrong movie costs nothing. An AI system that misclassifies a tumor or misidentifies a drug interaction can cost a life. The standards for clinical validation in healthcare are, appropriately, extraordinarily high. This means that health AI companies must invest in rigorous prospective and retrospective studies, FDA regulatory pathways, clinical champion relationships, and peer-reviewed publication programs that most software companies never encounter. This is a barrier to entry — and barriers to entry are, from an investment perspective, exactly what we want.

Third, the reimbursement landscape. Healthcare in the United States is reimbursed through a system of extraordinary complexity — Medicare, Medicaid, commercial insurers, value-based care contracts, capitation arrangements, fee-for-service schedules — and every one of these reimbursement structures represents a potential pathway to sustainable revenue for health technology companies. Unlike consumer technology, where revenue models are often discovered post-launch, health technology companies that understand their reimbursement pathway before they build have a structural advantage. We invest disproportionately in founding teams that have thought this through before their first line of code.

Our Three Investment Pillars

Within health technology, we concentrate our investments across three areas where we have developed genuine conviction and a repeatable edge in identifying exceptional companies.

1. Clinical AI That Improves Patient Outcomes

The most powerful application of machine learning in healthcare is not workflow automation — it is direct clinical decision support that improves the accuracy, speed, and accessibility of diagnosis and treatment. PrecisionPath Diagnostics, one of our earliest and most successful investments, exemplifies this thesis. Their liquid biopsy platform detects five cancer types at Stage I with 94.2% sensitivity using cell-free DNA methylation patterns — a performance benchmark that exceeds the clinical standard of care for all five indications. The clinical impact of early detection at this accuracy level is not incremental: Stage I cancer survival rates are typically three to five times higher than Stage IV rates across the cancers PrecisionPath detects.

We look for clinical AI investments that meet three criteria: a validated performance benchmark (not a theoretical one), a clear FDA regulatory pathway with designated precedent, and a reimbursement hypothesis supported by existing CPT codes or an emerging value-based care structure. CeruleanBio, our AI-guided drug discovery platform, meets these criteria at the asset level: each of their five clinical-stage candidates was generated by an AI platform that has itself become a competitive moat, attracting exclusive licensing interest from top-ten global pharmaceutical companies.

2. Health Infrastructure That Reduces Provider Friction

The second pillar of our thesis is infrastructure software that reduces the administrative and operational burden on health systems, physician groups, and care facilities. NeuroBridge, our neural interface rehabilitation platform, is emblematic of this category — though it sits at the intersection of clinical AI and infrastructure. By enabling non-invasive motor rehabilitation for stroke patients in hospital systems, NeuroBridge eliminates the surgical risk, implant complexity, and post-acute care burden of previous generation devices, making neural rehabilitation accessible to a vastly larger patient population.

Infrastructure companies in healthcare often fly below the radar of general-purpose technology investors because they lack the headline metrics of consumer software. A hospital system software company with $4M ARR and 14 health system customers is not a TechCrunch story — but it is, in our experience, a business with extraordinary durability. Switching costs in clinical software are genuine and material. Health system procurement cycles are long, but contract renewals are nearly automatic. And the integration depth that makes initial deployment complex also makes competitive displacement extremely difficult.

3. Biotech Platforms That Compress Drug Discovery Timelines

The third pillar is early-stage biotech — specifically, platform companies that use AI to compress the timeline and cost of drug discovery, rather than asset-focused companies bets on a single molecule. This is a nuanced distinction that matters enormously at the seed stage. A single-asset biotech at seed is essentially a binary bet on one clinical trial outcome. A platform biotech at seed is a bet on the engine itself — the AI architecture, the proprietary biology data, the wet lab validation infrastructure — that can generate multiple clinical-stage candidates over time.

We invested in CeruleanBio in 2021, when their platform had produced two pre-clinical candidates in rare oncology indications. By 2025, the same platform had generated five clinical-stage candidates, two of which are in Phase II trials. The platform thesis proved out: each successive candidate was generated faster and at lower cost than the last, as the AI model accumulated training signal from wet lab experiments. This is the compounding advantage that platform biotech investors are looking for, and it is exceptionally rare at the seed stage.

Why Seed Stage Is the Right Time to Invest in Health Tech

Health technology has historically been considered a later-stage investment category — the reasoning being that the complexity of clinical validation, regulatory approval, and reimbursement establishment required a longer runway and more capital than seed-stage vehicles could provide. We believe this conventional wisdom is now outdated, for three reasons.

First, the FDA regulatory environment has matured. The FDA's Digital Health Center of Excellence, established in 2020, has developed clear guidance for AI/ML-based software as a medical device (SaMD), including the Breakthrough Device Designation program that accelerates review for high-impact technologies. PrecisionPath and NeuroBridge both received Breakthrough Device Designations — a regulatory validation that, at the seed stage, dramatically de-risks the regulatory pathway and signals FDA receptivity to the technology. Founders who understand how to navigate this environment are building companies with structural regulatory advantages that are invisible to investors who do not understand the FDA process.

Second, reimbursement models are maturing. CMS's new technology add-on payment (NTAP) program, the expansion of Digital Therapeutics reimbursement under commercial payers, and the proliferation of value-based care contracts are creating new revenue pathways for health technology that did not exist five years ago. We are increasingly investing in companies that have a clear line of sight from seed-stage milestones to a specific reimbursement code or value-based contract structure.

Third, the cost of building has declined dramatically. Foundation models for biology, publicly available datasets like the UK Biobank and the All of Us Research Program, and cloud-based wet lab automation platforms mean that an AI drug discovery company can now generate its first clinical hypothesis with a team of eight and $5M — work that would have required $50M and three years of academic collaboration a decade ago.

What We Look For in a Curevstone Company

Our investment process is rigorous, and our criteria are specific. We are not generalist seed investors who have added a health tech vertical; we are health-focused investors who have built genuine clinical, regulatory, and commercial domain expertise over six years of concentrated investing.

The founding team is always our first and most important evaluation. We look for technical founders with genuine clinical domain knowledge — not founders who have hired a medical advisor, but founders who themselves understand the clinical workflow, the biological mechanism, or the health system procurement dynamic at a level of depth that cannot be acquired in six months of market research. The best health tech founders we have backed have a combination of computational or engineering excellence and firsthand clinical or biological expertise. This is rare, and when we find it, we move quickly.

Beyond the team, we look for defensible data assets. The single most valuable thing a health technology company can accumulate is proprietary, longitudinal clinical data that cannot be replicated by a well-funded competitor starting from scratch. PrecisionPath's methylation pattern database, compiled through partnerships with 22 hospital systems over three years of sample collection, is not replicable at any price. CeruleanBio's proprietary assay data from thousands of rare oncology patient samples is not available through any public dataset. These data assets are the moats that make health technology companies enduringly valuable.

Finally, we insist on a credible reimbursement hypothesis. Not a finished reimbursement strategy — that comes later — but evidence that the founding team has thought carefully about how their product gets paid for, by whom, under what conditions, and on what timeline. Founders who have not engaged with this question before their seed round are, in our experience, building toward a discovery they will make at the worst possible time.

"The best health technology founders understand that building a company in this sector is simultaneously a scientific project, a regulatory project, a commercial project, and a human project. The ones who succeed never lose sight of any of the four." — Alejandro Reyes, Managing Partner

The Next Decade Belongs to Health Technology

We are at the beginning of a transformation in human health that will be driven by the convergence of artificial intelligence, molecular biology, and digital infrastructure. The first generation of this transformation — electronic health records, telemedicine, consumer health apps — laid a digital foundation, imperfect as it was. The second generation, which is now beginning, will use that foundation to deliver something that was previously impossible: health systems that learn from their own outcomes, drug discovery platforms that find cures for diseases that have resisted decades of traditional research, and diagnostic tools that detect disease years before symptoms appear.

At Curevstone Capital, we have built our firm around the conviction that the most important companies of this generation will be built at the intersection of biology and computation. We invest $5M at the seed stage, we write one check at a time, and we go deep with each company we back. Our portfolio of 19+ companies across clinical AI, health infrastructure, and biotech platforms represents six years of conviction investing in this thesis — and we believe we are still in the early innings.

If you are a founder building at the frontier of health technology, we would like to hear from you. Not because you fit a category, but because you share a conviction — that health technology is the most important investment category of the next decade, and that the work of building it is worth doing with the rigor and patience it requires.