Earth-observation data is sovereign by origin and licensed by redistribution. Dataionics sits between the two. One Regime Template, evaluated against every act, closes the loop the spreadsheets never did.

The Data Flows

Dataionics is an Earth-observation data-flow operator with a specific job: take the European Union's Copernicus Sentinel programme — the satellite constellations that produce the largest openly licensed Earth-observation archive on the planet — and route the data through sovereign infrastructure to the agencies, partners, and licensed redistributors who turn pixels into policy.

The four Sentinel families Dataionics touches are not interchangeable:

  • → Sentinel-1 — Synthetic Aperture Radar. All-weather, day-night surface movement, flood mapping, oil spill detection. The high-cadence, low-licensing-overhead stream.
  • → Sentinel-2 — Multispectral optical imagery. Land cover, agriculture, deforestation. The most redistributed stream, the most re-licensed downstream.
  • → Sentinel-3 — Ocean and land surface topography. Sea-level rise, lake monitoring, ice-sheet dynamics. Lower cadence, higher sensitivity.
  • → Sentinel-5P — Atmospheric composition. Air quality, methane, NO₂, ozone. The smallest stream by volume, the largest by regulatory consequence.

The partners Dataionics routes to are heterogeneous on purpose. Public agencies reroute data into national registries and emergency-response platforms under their own jurisdictional rules. Private redistributors — commercial Earth-observation vendors, agricultural analytics firms, climate-risk re-insurers — receive data under explicit redistribution licenses (typically CC-BY-4.0 for free-and-open, or a per-counterparty redistribution agreement). Research consortia consume data for model training under grants-controlled regimes that cap downstream redistribution.

What the data flows share, what they don't, and what needs governing:

  • → Provenance is well-defined at the source and degrades the moment data leaves Copernicus. Sentinel metadata is a clean record. Anything past the first redistribution hop requires manual provenance stitching.
  • → Sovereignty stops being ambiguous exactly once a redistribution license clause is invoked. Everything upstream is, in principle, sovereign. Everything downstream depends on the clause.
  • → Disputes are real and they happen at the boundary, not the source. A redistributor's downstream use case crosses a license clause, but no shared ledger reflects that. The regulator who catches it is the regulator who did the audit, not the system that should have flagged it.

The trust model Dataionics inherited — a stack of consultancy-authored MOU PDFs, per-partner spreadsheets, and quarterly compliance reviews — worked when there were five redistribution partners. By the time there were fifty, it had stopped being a governance model and had become an inevitability model: the next breach would arrive, and the answer would be prepared after the fact.

The Zone Topology

An Axone Zone is a bounded framework of facts and rules — written in Prolog — that decides which acts become opposable inside it. An Axone Regime is the rule system itself: which datasets are eligible, which counterparties can redistribute, what provenance must be preserved, what license terms apply, and how a breach becomes a payable act.

For an Earth-observation operator, three resource types are load-bearing:

  • → Dataset resources — the Sentinel granules themselves, addressed by orbit frame, acquisition time, and a content hash. Each one carries the upstream license tag as a slot, not as a separate metadata row.
  • → Transformation resources — derived products: orthorectified imagery, NDVI composites, change-detection masks. Each transformation appends to the lineage chain and reads its inputs' license slots.
  • → Redistribution resources — the act type that is, institutionally, the entire question. A counterparty receives a dataset (or a derived product); the regime decides whether that act is opposable.

The slots on a dataset resource tell the rest of the story. slot(zone_id, atom) names the Zone. slot(eu_origin, atom) tags provenance — the dataset was generated by an EU public programme under a sovereign mandate. slot(attribution_license, atom) carries the redistribution license: CC-BY-4.0 for free-and-open paths, per_partner_agreement for negotiated re-distribution, research_only for grant-controlled downstream. slot(provenance_hash, hex) is the cryptographic anchor for any subsequent dispute.

The result is a Zone that knows, for every act touching its data, which license rule applies — and which rule doesn't.

Regime Snippets

Below is an anonymized slice of the regime Dataionics runs. Counterparties are anonymous identifiers. Datasets are referenced by their dataset_id atom, not by Sentinel granule URIs. No Copernicus metadata, no partner details.

prolog · Earth-Observation Zone Regime — Copernicus / EU Redistribution 
% ============================================================
% DATAIONICS ZONE REGIME — COPERNICUS / EU REDISTRIBUTION
% ============================================================

% ---------- ROLES (operator / partners / regulators) ----------

actor(dataionics,          operator).
actor(partner_agricultural_v1,  redistributor).
actor(partner_climate_v1,     redistributor).
actor(consortium_research_v1, research_consortium).
actor(eu_data_authority,     arbiter).

% ---------- DATASET SLOTS (provenance / license) ----------

slot(s2_tile_eu_lc_2026q2, eu_origin,            copernicus_s2).
slot(s2_tile_eu_lc_2026q2, attribution_license,   CC-BY-4.0).
slot(s2_tile_eu_lc_2026q2, provenance_hash,       'a3f1...c9').

slot(s5p_eu_no2_2026summer, eu_origin,          copernicus_s5p).
slot(s5p_eu_no2_2026summer, attribution_license, per_partner_agreement).

% ---------- COUNTERPARTY REDISTRIBUTION PROFILES ----------

redistribution_profile(partner_agricultural_v1, CC-BY-4.0, commercial_ok).
redistribution_profile(partner_climate_v1,    per_partner_agreement, noncommercial_only).

% ---------- ELIGIBILITY: a dataset is eligible iff provenance chain is intact ----------

dataset_eligible(Dataset) :-
    slot(Dataset, eu_origin, Source),
    eu_origin_registered(Source),
    slot(Dataset, attribution_license, License),
    slot(Dataset, provenance_hash, Hash),
    \+ disputed(Dataset).

% ---------- REDISTRIBUTION RULE: who may receive what ----------

redistribution_allowed(Counterparty, Dataset) :-
    dataset_eligible(Dataset),
    slot(Dataset, eu_origin, Source),
    slot(Dataset, attribution_license, License),
    redistribution_profile(Counterparty, License, Scope),
    scope_permits(Scope, Source).

% ---------- DISPUTE PRECONDITION: a redistribution is contestable ----------

opposable(redistribute(Dataset), Counterparty) :-
    redistribution_allowed(Counterparty, Dataset).

% ---------- EFFECT: every qualified redistribution is auditable ----------

effect(redistribute(Dataset), Counterparty,
    logged(Counterparty, Dataset, eu_origin)) :-
    opposable(redistribute(Dataset), Counterparty).

% ---------- QUERIES ----------

% ?- redistribution_allowed(partner_agricultural_v1, s2_tile_eu_lc_2026q2).
% true.

% ?- redistribution_allowed(partner_climate_v1, s5p_eu_no2_2026summer).
% true                 -- scoped to non-commercial use; logged for audit.

% ?- effect(redistribute(s2_tile_eu_lc_2026q2), partner_agricultural_v1, E).
% E = logged(partner_agricultural_v1, s2_tile_eu_lc_2026q2, eu_origin).

The regime does not police downstream use at the dataset level — that is the redistributor's contract, sometimes codified, sometimes implicit. The regime does police the handoff: every redistribution act becomes a logged, opposable record whose license scope is a property of the act, not a footnote on a spreadsheet.

Opposability & Settle

Opposability is the property that turns the redistribution record into something with effect. A counterparty who receives a dataset is accountable for the terms under which they received it, not just for the receipt. A regulator who needs to inspect a redistribution chain can pull the chain, not a spreadsheet summary.

When the zone-level regime encounters a redistribution that should not have been opposable — a license mismatch, a disputed provenance, an out-of-scope counterparty — the act is contested. The contestation protocol routes through Pactum, the Solidity settlement layer that ships on the Axone mainnet. Pactum turns the contested act into a payable event: a breach of license scope, a misrepresentation of provenance, a re-distribution to a non-registered source. The settlement runs against the regime and is executed on-chain — without a vendor ticket, without a legal motion.

Counterparty identity doesn't have to live inside the Axone chain. IBC settlement routes the dispute and the payable event to counterparties whose authoritative ledger lives on another chain. The regime, the audit, and the settlement all reference the same act type — redistribute(Dataset) — across the boundary.

This is the shape of the partnership Path: yes, but the architecture that backs it is not a forecast. The Pactum pattern, the IBC pattern, and the regime-as-audit pattern are the model.

Outcome: Drift from Licensed Redistribution to Governed Redistribution

Dataionics' deployment consolidated the data-flow supply chain under a single Axone Zone. The Shape of the result, in aggregate terms rather than named-testimonial form:

  • → One regime replaces per-partner SSPs. New redistributors join the same Zone; new redistributions clear through the same rule; new licenses add predicates rather than new spreadsheets.
  • → Sovereign provenance stays sovereign. A Sentinel dataset tagged eu_origin carries that tag into derived products and into the redistribution record. Encoding sovereignty into the slot replaces the trust-the-spreadsheet assumption.
  • → Redistribution license scope is a property of the act, not a footnote on the contract. A redistribution to a commercial_ok counterparty under a CC-BY-4.0 license routes opposably; a redistribution to the same counterparty under a noncommercial_only scope routes — and triggers Pactum when it escapes.
  • → Audit-grade handover to EU regulators via the auditability surface. The EU data authority reads the same act history Dataionics reads; it doesn't need a separate audit portal to interpret a separate compliance surface.
  • → Operationally reproducible. The Zone Builder turns the Earth-observation regime into a typed, versioned artifact the data engineering team can revise and ship — without re-litigating the contract graph.

Dataionics did not promise the European public Earth-observation system a miracle. It promised — and delivered — that the Sentinel tile redistributed to a redistributor on Tuesday carries the same sovereignty, the same license scope, and the same audit trail as the tile redistributed on Friday. That is the threshold a sovereign data supply chain crosses when it stops being a constellation of MOUs and becomes a governed service.

What Zone Builders Mean Going Forward

The Earth-observation supply chain is the most license-loaded supply chain an Axone Zone will plausibly run. If a regime can survive a Copernicus redistribution clause, a sovereign provenance assertion, and an EU public-sector audit reading the same Prolog file — every other multilateral, multi-party data flow gets easier from there. The Zone Builder is the layer that makes that possible at scale: a typed, versioned, parameterizable artifact an institutional engineering team can ship and revise. The case study above is anonymized for a reason — the partnership is still being formalized. But the shape of the result is not a forecast. It is what already shipped.