Foundation · About · Proposal

The Proposal

Shape we propose · Structure we do not

Shape-Not-StructureCurrent Stage

Foundations that overclaim weaken the shapes they're trying to propose. The Sovereignty Foundation publishes this page so that what it proposes — and what it deliberately does not — is in the public record from the beginning. The frame is single-sentence simple: we propose shape; we do not define structure. The rest of this page elaborates what that means for what we claim and what we don't.

What the Foundation claims

That substrate-class artificial intelligence requires governance structures that do not currently exist.
That a credible governance proposal in this category must demonstrate technical feasibility, not just articulate desired properties.
That the substrate work the Foundation has built — and is publishing — demonstrates that deterministic, cryptographically-auditable governance for AI is mechanically possible.
That the governance commitments the Foundation has ratified are a serious attempt at structural ethical enforcement, and are published for public scrutiny.
That the standards drafts the Foundation has published represent a substantive contribution to the standards development that this category needs.
That the Foundation's structural choices — public-trust posture, custody-not-commodity IP stewardship, anti-capture protections, intentionally asymmetric governance — are appropriate to the role the Foundation proposes to play.

What the Foundation does not claim

That the Foundation is the only legitimate governance structure for substrate-class AI. It is not. The structure that ultimately emerges will be shaped by many actors.
That the Foundation's standards drafts are finished, ratified, or industry-endorsed. They are Working Drafts, explicitly published to invite revision.
That the Foundation currently issues certification marks or operates conformance programs. The certification mechanism is proposed; it is not yet active.
That the Foundation has authority over how substrate-class AI is governed by other organizations, regulators, or jurisdictions. The Foundation has no such authority.
That the Foundation's substrate work is production-ready in all its layers. Some layers are mature; others are at varying stages of development. The Architecture section is honest about which are which.
That the Foundation has institutional standing equivalent to established standards bodies (ISO, IEEE, NIST AISI, etc.). It does not. It is a proposal-stage organization seeking constructive engagement with such bodies.
That the Foundation has resolved the open questions that any serious governance proposal in this category surfaces. Many questions remain open. The Foundation publishes its current positions and invites refinement.

The shape-not-structure framing

The Foundation's posture: we propose shape; we do not define structure. We produce mathematical, cryptographic, and institutional shape-primitives — objects open for inspection — and submit them to whatever review, adoption, refinement, or rejection the people working in this category choose to give them. The structure that ultimately governs substrate-class AI is for those people to build. Our work is to make the shapes hold up to scrutiny when they do.

This framing does real work:

It is more honest about the current stage of the work.
It is more defensible against legitimate skepticism from external reviewers.
It is more inviting of the engagement, scrutiny, contribution, and challenge the Foundation actively seeks.
It is structurally aligned with the Foundation's anti-capture commitments — shapes can be improved by others; assertions of authority cannot.
It matches the etymological structure of the work itself: the Foundation produces infra (under-shape); structure is what others build above it. Pulse Mesh, the substrate's transport layer, is named correctly — it sits below.

The Foundation will revisit this framing as the proposal matures. If and when the Foundation transitions to operating as an established standards body (or some adjacent role), the language will update to reflect that change explicitly. The shape-not-structure framing is not permanent humility; it is accurate description of the current stage.

Engagement that strengthens the shape

Because the Foundation's authority will only ever come from the quality of the shape it proposes and the breadth of external engagement, the kinds of engagement that strengthen the proposal are:

Technical critique of the substrate work that surfaces ambiguities, errors, or improvements.
Independent implementation of the standards drafts.
Substantive engagement with the governance commitments — including disagreements articulated specifically.
Participation in the Working Groups as they form.
Dialogue with adjacent standards-development bodies and regulatory institutions.

The Engage section describes how each kind of engagement can begin.

Shapes can be improved by others. Structures asserted as authority cannot.

Published 2026