Mission 01 System 02 Research 03 Impact 04 Careers 05 Filings 06 Contact 07
Stockholm, Sweden
Published application 202641026065
Impact

Impact you can audit, like a clinical endpoint.

Our impact and our revenue are the same unit, a dataset that can be used safely where it could not be used before. That keeps the mission and the business pointed in one direction, and it makes the impact countable.

Primary goal SDG 3, good health and well-beingReporting Annual, from first commercial release
§ 01Theory of change

From records that cannot move to medicine that reaches more people.

Problem

Data is locked.

Clinical records cannot legally leave their institution or country. Research waits months for access; some populations are barely present in global datasets.

Intervention

Certified synthesis at the source.

The engine runs inside the institution and produces cohorts under formal ε-differential privacy, each with a certificate and audit trail.

Output

Data that can travel.

Certified cohorts that researchers, sponsors and model builders can use across institutions and borders.

Outcome

Faster, wider research.

Shorter feasibility cycles, more institutions taking part in studies, models trained on broader populations.

Impact

Better medicine, privacy intact.

Treatments developed on evidence that represents more of the people they are meant for.

§ 02Five dimensions

The impact case, stated plainly.

Structured on the Impact Management Project’s five dimensions, the common language of impact investors.

WhatHealth data becomes usable for research and AI development without exposing any individual patient. The outcome is faster, broader clinical evidence.
WhoPatients whose data is protected and whose populations become visible to research; researchers and sponsors who today wait months for access; institutions locked out of studies by data constraints.
How muchFor every certified release we measure the source records kept behind the institutional boundary, the institutions and jurisdictions contributing, the time from request to analysis-ready cohort, and the studies it enables.
ContributionWithout a provable privacy standard, the data stays where it is. De-identification is losing regulatory acceptance; certified synthesis offers a route that did not exist before.
RiskThe main risk is synthetic data used for a purpose it does not support, leading to a wrong conclusion. We mitigate it with measured utility and printed contraindications on every certificate.
§ 03SDG alignment

Three goals, one mechanism.

3.b

Good health. Support research and development of medicines, including for diseases that mainly affect developing countries.

9.5

Research capacity. Enhance scientific research and the technological capability of institutions in every country.

16.10

Fundamental freedoms. Protect personal data with a mathematical bound, not only a policy.

§ 04What we will report

Six numbers, every year.

Baselines start at zero with our first commercial release. We will publish these figures annually, alongside the method used to count them.

01

Certified datasets released Cohorts issued with a full certificate and audit trail.

UnitDatasets per year
02

Institutions and jurisdictions contributing Hospitals and registries whose data is used without leaving their boundary.

UnitInstitutions · countries
03

Time to analysis-ready data From a research request to a usable cohort, against the access route it replaces.

UnitMedian days, before and after
04

Research enabled downstream Studies, feasibility assessments and models built on certified cohorts.

UnitProjects per year
05

Representation Share of cohorts drawn from populations under-represented in global health datasets.

Unit% of records released
06

Privacy breaches from released cohorts Confirmed re-identification or membership-inference breaches traced to a released cohort. The target is zero, and we will report the real figure.

TargetZero
§ 05Commitments

What we will not do, at any price.

Impact that can be traded away is not impact. These commitments hold as we grow.

  1. We will never sell or move raw patient records. The engine runs at the source. Only certified synthetic cohorts leave.
  2. No dataset ships without its certificate. Privacy budget, measured utility and contraindications travel with every release.
  3. We will not support re-identification or the targeting of individuals. Our data is for population-level research and model development.
  4. No black-box quality claims. Every certificate discloses its concordance metrics, its privacy accounting method and its contraindications, so its claims can be checked.
  5. We report impact annually. The six figures above, with their method, whether they flatter us or not.

Building with impact investors who think in decades.

If your fund backs science-led companies on SDG 3, we would like to hear from you.

Talk to us about impact →