Corti Brings Sovereign Healthcare AI to France With Scaleway and Dedalus
Corti is deploying its Sovereign Healthcare AI on Scaleway's European cloud, with Dedalus integrating clinical AI into French healthcare software, the companies announced on November 17.

Why it matters
- Corti announced its Sovereign Healthcare AI offering for France on Scaleway's European cloud on November 17.
- Dedalus is already integrating Corti's clinical AI into software used by French healthcare organizations.
- Corti Models offers a single OpenAI-compatible API for accessing AI models.
- Kommodity, Corti's open-source Kubernetes engine, spans Scaleway, Azure, Hetzner, and bare-metal environments.
- Corti's documentation API supports SOAP notes, patient summaries, and follow-up letters in French and other languages.
Corti is bringing its Sovereign Healthcare AI offering to France on Scaleway's European cloud, with healthcare software provider Dedalus already integrating Corti's clinical AI into software used by French healthcare organizations, the companies announced on November 17. The three-way partnership pairs a European AI platform with European infrastructure and a distribution channel into hospitals and clinics. At stake is a question every healthcare technology team in Europe now faces: how to use advanced AI while keeping control over sensitive data, model access, and the cloud systems that process each request.
The deal matters because it targets the request path, not just the data center. A clinical document can sit on a local server while an application quietly sends its contents elsewhere for an AI response. Sovereign AI architecture, in Corti's framing, concerns the route a request takes through the model, the runtime, and the infrastructure — as well as where the original record lives.
What did the companies actually announce?
Three roles, one stack:
- Corti supplies the AI platform: clinical speech recognition, documentation generation, medical coding, and tools for building AI agents.
- Scaleway, part of the iliad Group, provides the cloud and AI infrastructure for the French offering.
- Dedalus connects those capabilities to software healthcare professionals already use.
According to the companies, the arrangement lets developers building on Corti's platform keep sensitive healthcare data and AI workloads within a European environment under European jurisdiction.
The distinction between storing data and processing it sits at the center of the pitch. Sovereignty hinges on the request path through model and runtime, not just data storage location. Corti's sovereign cloud platform describes European cloud and on-premises deployment options using the same APIs, so a healthcare software company can pick an acceptable hosting arrangement without redesigning every application feature around a new interface.
"Sovereignty isn't just about where data is stored," said Lars Maaløe, Corti's co-founder and CTO. He described model choice, hosting choice, and the ability to switch environments as central to the company's strategy.
What can the clinical AI actually do?
Corti's flagship clinical AI model, Symphony, supports clinical and administrative applications across electronic health records, virtual care, practice management, and life sciences. Developers embed these capabilities into their own products rather than deploying standalone tools.
The clinical documentation API offers the most concrete example. It turns transcripts, facts, or structured information into notes and other documents shaped by predefined or custom templates. Corti lists supported formats including:
- SOAP notes
- patient summaries
- follow-up letters
The API supports French and other languages — a practical requirement for the French healthcare market the partnership targets.
The integration path is straightforward. An application captures information from an encounter, passes relevant material to the AI service, and presents a structured draft within the clinician's workflow. Clinicians then review, edit, and sign the document. The AI drafts; the human stays accountable.
Dedalus' role is to wire those capabilities into software that healthcare professionals already use. Guillaume Reynaud, the company's director of the CIS4U Business Unit, Life Science, Data and Connected Care, said the partnership is intended to reduce administrative work while improving control over how the AI is deployed and hosted.
Why does an OpenAI-compatible API matter here?
Corti Models provides a single OpenAI-compatible API for accessing AI models. That compatibility describes the software interface: developers use a familiar request format and compatible client tooling while directing requests to Corti's service. It does not mean OpenAI operates the hosting infrastructure.
The compatibility layer does real architectural work. Separating an application's integration from a particular hosting arrangement makes it easier to change the underlying service later, because the integration code doesn't have to change with it.
Caveats remain, and Corti's own material acknowledges them. Developers still need to evaluate the behavior of the model they select, including how it handles clinical terminology, structured outputs, and the specific tasks their software supports. A compatible API makes models swappable at the protocol level; it doesn't make them interchangeable at the clinical level.
For healthcare organizations, the design separates two decisions that are usually bundled together: which model is suitable for the work, and which infrastructure is acceptable for running it.
What is Kommodity, and how does it enable portability?
The other major component is Kommodity, Corti's open-source cloud infrastructure and Kubernetes engine. Kubernetes coordinates containerized software across groups of machines, helping operators deploy applications, manage resources, and scale workloads.
Kommodity's repository describes a common Kubernetes-based control plane spanning providers including:
- Scaleway
- Azure
- Hetzner
- virtualized environments
- bare-metal environments
The engine combines Cluster API, which manages the lifecycle of clusters and their machines, with Talos Linux and services for machine attestation and disk-encryption key management. Kommodity's portable Kubernetes control plane reduces provider-specific work when healthcare teams switch clouds.
Machine attestation deserves a note. It checks evidence about how a machine booted before releasing configuration or secrets. Alongside common deployment tooling, these mechanisms aim to make the infrastructure more inspectable and reduce the amount of provider-specific work needed to operate it — a meaningful property for organizations whose auditors ask pointed questions about where clinical computation happens.
Portability is not a magic wand, and the companies don't claim it is. Data transfer, networking, identity systems, and performance testing all still need attention during any migration. Kommodity's value is giving operators a more consistent foundation when they change where their applications run.
How does this fit the European sovereignty debate?
The announcement lands amid growing European demand for digital sovereignty in health data. Hospitals and healthcare software vendors face regulatory and institutional pressure to keep patient information under European jurisdiction, but they also want access to frontier-grade AI capabilities that have largely been built and hosted elsewhere.
Corti's answer is architectural rather than rhetorical: keep the model API common, keep the runtime open source, and keep the hosting within reach. European hosting addresses where computation happens; portable infrastructure and a common API give teams more options for how that deployment evolves over time.
The division of labor across the three companies makes the partnership relevant beyond a choice of data center. Corti supplies the AI platform, Scaleway provides European cloud infrastructure, and Dedalus integrates clinical AI into healthcare software that clinicians open every day.
What is the practical test going forward?
The partnership's real measure is whether healthcare developers can deliver useful features within existing clinical workflows while keeping model selection and infrastructure choices manageable. Dedalus' integration work in French healthcare organizations will be the first evidence.
The November 17 announcement also signals how the sovereign AI market is likely to develop: not as a single vendor offering a sealed stack, but as composable layers — model, runtime, cloud — that healthcare teams can recombine as their requirements and options change. If Corti, Scaleway, and Dedalus can show that clinicians get useful AI-assisted documentation while CIOs keep jurisdictional control, the French deployment becomes a template other European health systems can copy.
Source: unite.ai
More from Elena Vasquez
Show full bio
Market editor covering technology and product news in artificial intelligence and machine learning at AI In Context.
708 articles
Related articles
- League's Healthcare AI Agents Go Live With 99.72% Escalation Accuracy
- Ohio State, Microsoft Launch Five-Year Clinical AI Partnership
- Google DeepMind Unveils AI Co-Clinician Research Initiative
- Axway launches Amplify Platform to control AI costs and data risk
- Medical AI Needs More Than Accuracy to Work at the Bedside