Deepc's answer to a shrinking radiology workforce
By 2050 the World Health Organization expects 22 percent of the world's population to be over 60, and Europe's radiologist workforce is not keeping pace. The EU-REST study projects imaging demand will rise 70 percent by 2030 while the workforce grows barely 1 percent a year. That gap has a catch. Under Article 6(1) and Annex I of the EU AI Act, any AI system embedded in a medical device already requiring Notified Body assessment under the MDR is automatically high-risk, and by 2023, radiology alone counted 173 CE-certified AI products from 90 vendors, each with its own oversight, logging and post-market duties. If deployed without coordination, AI does not free up clinician time so much as reallocate it into paperwork.
deepc, founded in Munich in 2019 by Dr Franz Pfister, Julia Moosbauer and Paul Mayer, built deepcOS to close that gap. The platform plugs regulator-approved AI models from multiple vendors into a hospital's existing PACS workflow through a single integration layer, validating each one against local clinical data before it goes live and folding the associated oversight and logging into one system rather than many. It now spans more than 60 clinical indications across over 30 countries, serving hospitals including Solothurn Hospital in Switzerland, the Vivantes Hospital Group and LMU University Hospital in Germany, and the NHS in England.
In 2026 deepc extended deepcOS with an agentic orchestration layer, built on Model Context Protocol and Agent-to-Agent communication, linking models, workflows and data under a single AI copilot, AIDA. It is being piloted with Mass General Brigham AI in Boston, King's College London and LMU University Hospital. The build-out follows $30 million raised to date, including a $13 million Series A extension closed in mid-2024 and co-led by Sofinnova Partners and Bertelsmann Investments.
Notified Body-assessed devices face an August 2027 deadline under the Act, and that obligation lands per device, not per hospital. Against a workforce set to shrink just as demand climbs 70 percent, infrastructure that turns 173 separately certified tools into one auditable layer is not a convenience. It is what decides whether AI ends up compounding the shortage or genuinely closing it.
Sources: deepc | EU-REST, Insights into Imaging | Crunchbase
Founders: Franz Pfister, Julia Moosbauer, Paul Mayer