Summary

• Headline led a $50 million Series B with existing and new investors.
• Scarlet specialises in assessing software and AI-enabled medical devices.
• Expansion ambitions include the US and Asia-Pacific, alongside more reviewers and certification capability.

London-based medical technology certification company Scarlet has raised $50 million in a Series B round led by Headline, with participation from Notion Capital and existing investors including Atomico, Creandum, Kindred Capital and Entrepreneurs First. The financing will support expansion of the organisation's assessment capabilities and its international presence as more medical products incorporate artificial intelligence.

Scarlet's commercial role differs from that of a typical health technology software developer. It assesses products made by other organisations against regulatory requirements and can issue relevant conformity decisions where it holds the necessary designation. That puts it inside the assurance process on which access to medical device markets depends, rather than simply selling a tool that helps manufacturers compile paperwork.

The company's founders, James Dewar and Jamie Cox, previously worked on medical AI and say they experienced long delays in the assessment of software products. The argument behind Scarlet is that review systems designed around more static devices are poorly matched to software that is updated frequently or retrained as new information becomes available. Such products can change their performance profile after initial development, making lifecycle management particularly important.

Scarlet is recognised in different roles across the EU, UK and US regulatory systems, according to the company and its investors. Those regimes are not interchangeable. European conformity assessment, UK medical device oversight and US third-party review operate under distinct rules, and any specific authorisation must be checked against the relevant regulator's register and the product category involved.

The company describes its process as AI-enabled, but qualified assessors retain responsibility for reviewing technical evidence and making certification decisions. An AI system might help organise submissions, compare documents and identify missing material. It cannot replace the need to assess clinical performance, risk controls and whether evidence supports a manufacturer's stated intended use.

Investors argue that assessment lead times can extend to 12 to 18 months, with updates requiring additional review. Those are company and investor characterisations, not a verified average across all notified bodies or device classes. Review duration can depend on product risk, submission quality, the availability of clinical evidence and the capacity of both manufacturer and assessor.

The central operational challenge is therefore capacity as much as technology. A certifier can improve document handling but still needs experienced people who understand healthcare, software assurance, clinical evaluation and regulatory requirements. Rapid business expansion must preserve reviewer independence and consistency while avoiding commercial pressure to approve products faster than the evidence permits.

AI-enabled medical devices also create new questions after approval. Software updates, shifts in training data and changes to the population using a tool may affect how it performs. A credible assurance process must consider the boundaries of authorised change, post-market monitoring and when a product needs reassessment. Faster review only creates enduring value if these requirements remain comprehensible and enforceable.

Scarlet intends to use the new funding to increase its team and build further capacity in the US and Asia-Pacific. Its investors see an expanding market as medical technology firms adopt AI more widely, but the company has not disclosed a valuation, audited review-time comparison or a public measure of how many certifications it has completed in each jurisdiction.

The Series B supports growth in the infrastructure surrounding medical AI rather than another diagnostic product itself. Its significance will be measured by credible assessment throughput, regulatory compliance and the continued confidence of manufacturers, health systems and regulators in the independence of certification decisions.

Independence will be particularly important as the certification market expands. The organisation reviewing a device is paid to perform a service, but its decision must reflect the evidence and the applicable standards, not a manufacturer's preferred launch timetable. Governance, assessor training and separation between commercial activity and conformity decisions are therefore material parts of the business model.

Manufacturers also need clarity about what happens after a product is approved. A system may be trained on additional data, used in a new clinical setting or changed to support another patient population. Each adjustment can affect safety and performance, even if its user interface appears nearly unchanged. Certification processes able to distinguish minor maintenance from a material product change could reduce unnecessary delays without weakening oversight.

A growing portfolio of customers would provide evidence that manufacturers value Scarlet's service, but it would not independently establish better clinical outcomes. Those outcomes depend on the devices under review and their use in healthcare. Scarlet's own performance should be measured through compliant decisions, consistency, auditability and service quality, alongside turnaround time.