The withdrawal of OpenEvidence, one of the most well-funded AI clinical decision support platforms in the US, from EU and UK markets is a significant signal for every health AI company with European ambitions. The reason given was direct: the complexity of compliance with the EU AI Act and the UK's evolving regulatory landscape made continued operation in those markets economically unviable. For health AI innovators at any stage of development, that outcome deserves careful attention. The product development decisions that make EU AI Act compliance achievable need to be made before the build, not after. Our article on Rapid Functional Prototyping in Healthcare sets out the approach that gives teams the best chance of validating those decisions before costs compound. EU AI Act compliance in healthcare is no longer a legal consideration to be addressed late in the development cycle. It is a market access question that shapes whether European launch is possible at all.
The companies that will build durable positions in European health AI markets are those treating regulatory strategy as a core competency, not an afterthought.
A well-funded, clinically validated AI platform withdrawing from two major markets is not a product failure. It is a regulatory strategy failure, and the distinction matters for how health AI companies at earlier stages interpret the lesson.
The EU AI Act introduces a risk-based classification framework for AI systems, with healthcare applications sitting in the higher-risk categories that carry the most demanding requirements around transparency, clinical evidence, human oversight, and ongoing monitoring obligations. For a product built and optimised for the FDA pathway, those requirements are not incremental additions. They represent a fundamentally different regulatory logic that affects product architecture, data governance, clinical evidence strategy, and post-market surveillance from the ground up.
A US-first health AI product that reaches European markets without having been designed for that regulatory environment faces a retrofit challenge that is frequently more costly than the original build. The OpenEvidence case is an illustration of what that calculation looks like when it reaches the point of decision: the cost of achieving compliance retrospectively exceeded the commercial case for the market.
The regulatory frameworks governing health AI in Europe and the US are not variations on the same theme. They are built on different principles, with different risk classification logic, different transparency requirements, and different expectations around clinical evidence and explainability. An AI clinical decision support tool cleared through an FDA pathway has demonstrated what the FDA requires. It has not demonstrated what the EU AI Act or EU MDR requires, and the gap between the two is not a documentation exercise. It can require material changes to the product itself.
For health AI companies planning European market entry, that asymmetry needs to be understood and planned for before development commitments are made, not discovered after a US launch has locked in an architecture that cannot readily satisfy European requirements.
The competitive instinct in health AI is to move fast, demonstrate clinical value, and build market position before the regulatory environment fully crystallises. In the US, that approach has worked for a significant number of companies. In Europe, it carries a specific and underappreciated risk.
Health AI companies entering European markets face the combined requirements of the EU AI Act and, where their product meets the definition of a medical device or software as a medical device, EU MDR. These frameworks interact in ways that require careful navigation. A clinical decision support tool that recommends or influences clinical decisions is likely to be classified as a medical device under MDR and as a high-risk AI system under the EU AI Act, triggering obligations under both frameworks simultaneously.
Meeting those obligations requires clinical evidence at a standard that supports the specific intended purpose of the product, a quality management system aligned to MDR requirements, conformity assessment through a notified body for higher-risk device classifications, and ongoing post-market clinical follow-up. None of those elements can be bolted on after a product has been built and launched. They need to be designed into the development roadmap from the outset.
The EU AI Act's requirements around transparency and human oversight for high-risk AI systems are not documentation requirements alone. They have direct implications for how AI models are built, how their outputs are presented to clinical users, and how the system logs and audits its own decision-making. A product that cannot produce an explainable audit trail of how it arrived at a clinical recommendation does not satisfy EU AI Act requirements, regardless of how clinically accurate its outputs are.
For health AI companies designing products for European markets, transparency and explainability need to be treated as product requirements with the same weight as clinical performance, not as compliance features added after the core product has been built.
The regulatory complexity that drove OpenEvidence's withdrawal is the same complexity that creates a durable competitive advantage for health AI companies that navigate it successfully. EU AI Act compliance, MDR conformity, and demonstrated regulatory readiness in European markets are not easily replicated by later entrants. They represent a meaningful barrier that protects market position once established.
Health AI companies that build regulatory strategy into their development roadmap from day one arrive at European market entry with a compliance posture that is coherent, documented, and defensible under notified body scrutiny. Those that defer regulatory strategy until late in development arrive at market entry facing the same calculation that OpenEvidence faced: a retrofit cost that may exceed the commercial case for proceeding.
The companies that will look back at this period as their European market entry point are those already embedding EU AI Act alignment and MDR compliance into their product architecture, clinical evidence strategy, and quality management systems now.
At Santegic, we work with health AI innovators to integrate regulatory strategy, EU AI Act alignment, and MDR compliance into development roadmaps before they become blockers, building the regulatory foundations that make European market access achievable rather than aspirational.
The OpenEvidence withdrawal is a live case study in what happens when regulatory strategy is treated as a market entry consideration rather than a product development consideration. The EU and UK markets are not closed to health AI innovation. They are closed to health AI products that were not designed with their regulatory requirements in mind.
EU AI Act compliance in healthcare, combined with MDR obligations where applicable, creates a demanding but navigable pathway for companies that engage with it early. The regulatory moat it creates for those that succeed is real, durable, and increasingly valuable as the European health AI market matures.
If your organisation is developing health AI for European markets and needs to ensure regulatory strategy is embedded in your roadmap from the outset, Santegic's healthcare consulting services are available to support you. Get in touch to discuss your EU and UK market access strategy before regulatory complexity becomes a blocker.
Santegic delivers specialist regulatory strategy, EU AI Act alignment, and MDR compliance advisory to health AI companies and digital health innovators entering UK and European healthcare markets.
One conversation. The right expertise.
