The moats are yours, not ours.
Sovereignty vs. interdependence, in nineteen places.
· by Risto Anton Päärni
The stance, in one line
A moat that only protects the vendor is a trap for the customer. We build the other kind — nineteen places where owning the thing beats renting it — and we hand them to you.
Most AI moats protect the seller.
The vendor gets the lock-in. You get the bill and the dependency. The deeper you go, the harder it is to leave — and that difficulty is sold to you as loyalty.
And because a moat you cannot hold is not yours, here is the honest part. We build moats the customer owns. The whole design has one spine: sovereignty vs. interdependence.
Sovereignty vs. interdependence
Interdependence is sold as efficiency. It is priced as control.
Every dependency is a lever a counterparty can pull. It pulls hardest at renewal, at scale, under pressure — the moment you can least afford it.
Sovereignty is not going it alone. It is owning the levers that decide your fate, and depending on others only where the dependency is reversible at near-zero cost. Every moat below is one of those levers.
Three we start with — all local
1 · Local jurisdiction
Be legally native to your own regime. A foreign vendor can move iron; it cannot move its jurisdiction of origin to sit inside yours.
2 · Sovereignty is segmentation
Sovereignty is not one product. It is the axis that segments the market, locally. Different jurisdictions need different depths — and that is the price ladder.
3 · Near-zero exit cost
Interdependence wins the full sovereignty-exit cost. We built and maintain our own libraries, so ours is near zero — and we pass that freedom to you.
The third one is the tell. The real question is never the licence you stop paying. It is the full cost to leave: the libraries you no longer understand, the maintenance muscle you let atrophy, the integrations that only ever ran on the vendor's terms.
Interdependence accrues that cost quietly, then hands you the whole bill at once. Because we build and maintain the stack ourselves, we can walk away from any single vendor for almost nothing. The sovereign party is the one who can leave — so it never has to.
The other sixteen
The three above open the stack. Sixteen more complete it. They group into four moves: own it, prove it, compound it, position it.
Own it
- 4 Weights. You own, export, and retrain your own model.
- 5 Data residency. Your data stays on substrate you control.
- 6 Substrate integrity. Hardware isolation below the agent's OS.
Prove it
- 7 Court-defensible evidence. Article 12 audit trail, not a promise.
- 8 Testable conformance. A sovereignty claim you can audit.
- 9 Linguistic provenance. Claims resolve to a legal term of record.
Compound it
- 10 Capital efficiency. Owned substrate beats capital volume.
- 11 Domain knowledge. A model trained on your industry, not the web.
- 12 Compounding alpha. Each cycle re-enters your own substrate.
- 13 Operational maturity. Calibrated oversight time can't be bought.
Position it
- 14 Regulatory depth. Regulation as a feature, not a tax.
- 15 Connector, not competitor. We compose with your platforms.
- 16 Sovereign upstream. EU-trained provenance where a partner supplies it.
- 17 Post-quantum readiness. The crypto transition, priced in early.
- 18 Track-record epistemics. Decisions from what is right, not who is liked.
- 19 Standardization. Owning the spec others conform to.
Where each moat honestly stands
Not all nineteen are shipping at the same depth, and saying otherwise is the fastest way to lose a European buyer's trust.
Moats 1–13 run in production today. Moat 16 depends on a partner's EU upstream — ours where they supply it, not a claim we make alone. Moat 17 is readiness, not a finished migration. Moat 19 is a standard we have proposed, not one the industry has ratified. We would rather name the tier than inflate it.
Why we hand the moat to you
Because a moat around us is only worth what it costs you to stay. A moat around you is worth what it saves you when everything else moves.
The moats are not independent. Jurisdiction makes segmentation legible. Owning your libraries is what makes owning your weights and your data possible. Substrate integrity is what makes the evidence hold up. Copy any single one and you have a feature. Hold the stack and you have a position.
The short version
- Most AI moats protect the seller. We build the kind the customer owns.
- The spine is sovereignty vs. interdependence — own the levers that decide your fate.
- Start local: your jurisdiction, your segmentation, your near-zero exit cost.
- Then own it, prove it, compound it, position it — sixteen more places.
- We name the tier each moat is at. The stack is the moat, not any one line.
Read next
- The moat is not the model — NVIDIA makes it possible, EU-sovereign deployment makes it defensible
- Prompt, People, Plant, Perception — the four capital classes of the agentic era
- The ontology lives in the team — why acquisition cannot extract it
- Europe's AI factory came home — and why the edge still needs a box
Risto Anton Päärni
Founder, CEO, Lifetime Oy · Editor in Chief, Lifetime Scope Journal