What the business is
Dehaze builds an AI foundation model over heterogeneous patient data (lab values, diagnoses) that identifies chronic-disease risk at population level and sells it to health insurers and providers.
The bet
Chronic disease is healthcare's biggest cost and the data to catch it early exists — a foundation model with causal explanations can sell to insurers as cost reduction.
Starting capital
€3.2M seed (April 2026), led by YZR Capital and DN Capital with Angel Invest, Zoho and Better Ventures
How it started
Founded in Munich in 2023 by Marius Klages and a team rooted in German and European deep-tech research, Dehaze started from the observation that doctors see less than 3% of available health data before making clinical decisions, leaving roughly a third of chronic diseases undetected until treatment is pricier. It is backed by Techstars and placed 4th in Phase 2 of the Munich business plan competition.
What happened
In April 2026 Dehaze closed a €3.2M seed led by YZR Capital and DN Capital, with Angel Invest, Zoho and Better Ventures participating, and with support from public funding programs of Germany's federal education and economics ministries. The money funds medical and technical hires, a sales team for growing international payer demand, and new modules for next-best-action recommendations and full traceability of AI decisions.
How it ended up
Still live and building: the platform is in use with a growing number of international payers, and the seed round expands the team and platform; no revenue figures disclosed.
What has to be true
WHO data makes chronic disease a multi-trillion-dollar, universally painful problem, so the market need is documented rather than assumed.
Positioning against chatbots: a foundation model with causal explanations differentiates Dehaze in a crowded medical-AI field.
Payers are the wedge: insurers are measured on cost, and up to 10% savings from earlier detection gives them a concrete ROI argument.
A trust and regulatory moat: German federal ministry backing plus promised traceability of AI decisions raises the bar for competitors.
What can be applied
Sell the model's output, not the model: Dehaze won insurers by framing early detection as 10% cost reduction and promising causal, traceable explanations — trust is the product.
Aftermath
As of September 2026, Dehaze is live and expanding from Munich: its platform is in use with a growing number of international health insurers, and the April 2026 seed is funding medical and technical hires, a larger sales team, and new modules for next-best-action recommendations and complete traceability of AI decisions. The company has not disclosed revenue or customer counts.
FOLLOW THE EVIDENCE