Data Dividend: monetize your data without exposing it
Your data earns. Nobody sees it. You get paid for your data's contribution to an AI trained by the collective, without any data ever leaving home. The model travels, end-to-end encrypted; your data stays home.
Estimate the value: ROI calculator · How you start: how a project works.
Who is it for?
The Data Dividend is for holders of massive or sensitive datasets who own a valuable asset but can't (or won't) hand it over:
- Organizations with large volumes of proprietary data (industrial, operational, scientific).
- Holders of rare or regulated data that would be legally risky to transfer.
- Actors who want to capitalize on a dormant data asset without losing control of it, without duplicating it, without exposing it.
The problem: your data has value, but selling it is impossible
Your data is worth a lot for training an AI. But selling it means handing it over: once copied, you lose control of it forever, and you expose yourself legally. So it sits dormant: an asset you can neither protect nor monetize.
The classic dilemma: either you keep your data inert, or you sell it and lose it. The Data Dividend opens a third way.
What the Data Dividend is NOT
What you monetize is the contribution of your data to training a model, not the data itself. The difference is fundamental: in a sale, the buyer leaves with your data; with Mesh, nobody leaves with your data, only the model improves.
How does it work?
How do you get paid for your data without exposing it?
Training runs at your place, on your own data. Only the model updates (the weights) travel, end-to-end encrypted, to a coordinator that aggregates them without ever seeing your data. Your contribution to the model is measured, and it's this contribution that is compensated.
The three steps
- Your data stays at your place. Training runs locally, in your infrastructure. No central database, no line of data emitted.
- Only the model travels, encrypted. The coordinator aggregates the weights without accessing the raw data. None of your data transits.
- You get paid for your contribution. Your input to the collective model is valued, without ever having exposed a single row.
What travels, exactly?
Only the model weights, encrypted. In our demonstration: on the order of ~29 KB of encrypted updates per cycle, zero line of data emitted. demo · order of magnitude (varies by model.)
The benefit
- Revenue on an otherwise dormant asset, without ceding control of it.
- Zero loss of control: your data is neither duplicated nor transferred; you keep all your rights.
- No data in circulation: your records never leave your premises, so there is no copy to intercept and no central database to breach. Exposure is minimized by design.
- Easier compliance: since raw data is not transferred, regulatory exposure is reduced (final compliance depends on the project's governance).
The proof
An interactive demonstration exists: access on request. It runs on real, public datasets (AI4I 2020 in industry, Tox21 in pharma, EDP Wind SCADA in energy) and shows, on each cycle, that only the encrypted weights travel while the data stays in place. demo · real data
To understand the mechanism, see the pillar Federated learning and How to train an AI without sharing your data.
Frequently asked questions
Am I selling my data?
No. The Data Dividend is not a sale. Your data is neither copied, transferred, nor resold. You get paid for your data's contribution to a model, not for the data itself: you keep full control of it.
Where is my data during training?
Always at your place. Training runs in your infrastructure, on your own data. Only the model weights, end-to-end encrypted, leave your walls.
Who can see my data?
Nobody. Not the other participants, not the coordinator, not Mesh. It's the federated architecture that guarantees it: the raw data is never emitted.
How is my compensation calculated?
It reflects the measured contribution of your data to the quality of the collective model. The scale is defined within the project framework; we establish it with you, with your real volumes.
Can I also use the final model?
Depending on the collective's agreement. The Data Dividend posture is to contribute and be paid; if you also want to leave with the engine, see The Collective or Engine, Ready.
Estimate the value of your contribution
We show you the mechanism in a demo, then we calculate together, with your volumes, what your contribution can be worth, without ever exposing your data.