Underwriting the token economy in times of AI, superintelligence and stablecoins.

In the entire market, nobody vouches with their own money.

Promotion is free and judgment carries no liability. Whoever praises a token risks nothing, so praise is worthless as a signal — and everything built on that signal inherits the weakness.

That was tolerable while tokens were a niche. It is not tolerable now. Payment stablecoins are being brought inside licensing regimes across the major jurisdictions, tokenised value is becoming ordinary settlement, and the counterparties are changing: more and more decisions about what to hold, route and accept are being made by AI agents at machine speed. An agent cannot read a room. It needs a signal with something at stake behind it.

Tokenised settlement does not remove risk. It relocates it — into instruments that move instantly and clear without an intermediary who would once have absorbed it. Metastability exists to price that risk and to make someone answerable for the price.

Money talks. Everything else walks.

A stablecoin answers one question: is this token worth a dollar. It says nothing about whether anything else is worth holding. Metastability goes beyond the stablecoin — it is built so that real utilities and good ideas stand out, because someone has put capital behind them and someone else has not.

The key is a convergence that has only now become practical. Small, specialised language models are cheap enough to run on every submission. Agents can trade, make markets, answer a community and read a chain without supervision. And Solana settles fast and cheaply enough for all of it to happen on chain, per token, continuously. Put together, they form one system: self-underwriting rug protection that is stable by construction and efficient to keep compliant.

Built for a market where the counterparty is an agent.

Agents already route payments, rebalance treasuries and choose what to accept. They do it faster than any human review, and they will do it with whatever signal is machine-readable. Today that signal is mostly noise.

Metastability is designed so that an agent can ask a precise question — is this token underwritten, by how much, by whom, and how far into its observation window — and receive an answer that has capital behind it. Agents can work on the underwriting side as well: valuation runs on on-chain inputs only, which makes it work an agent can do. The liability never becomes artificial. Whoever commits the stake carries the loss, regardless of who or what made the call.

Agents check everything. The matrix decides what it means.

Before a single token is staked against, the submission passes a diligence desk staffed by agents. Each one reads a different part of the project and reports in a form the others — and any human reviewer — can verify.

The result is not a verdict. It is a rating: ten weighted criteria, resolved into one score that sets how conservatively the token is valued. The underwriters then decide, with their own capital, whether they agree.

One protocol surface for every agent.

The consortium is designing its interface on the Model Context Protocol, so that any MCP-capable agent can query underwriting status and take part in intake without a bespoke integration.

• stamp.getUnderwriting status of a token: stake behind it, window, outcome.
• valuation.getThe conservative valuation and the inputs it was built from.
• underwriters.listWho stands behind a stamp, and with how much.
• window.watchSubscribe to an observation window and its criteria.
• intake.submitSubmit a token for underwriting through the Provenance Gate.

The interface is in design. Tool names and fields shown here are illustrative and will be superseded by the published specification.

A DAO — and decentralised last, not first.

Metastability will be governed through a DAO structure. Governance steers parameters, risk weights, fees and activations, and holds influence over a real treasury.

The order matters. First the mechanism is proven with real fees, then the token is introduced, then control is handed over. Each stage has a gate that has to be passed before the next begins — a protocol that cannot run without daily intervention is not ready to be governed by its community.

MSTB presale — to be announced.

A presale of MSTB is in preparation. Date, terms, eligible jurisdictions and the allocation will be published here, in writing, before anything opens.

The launch plan covers what a serious launch needs: managed liquidity and market making from the first day of trading, lock-ups for early allocations, a smart-contract audit before genesis, and venue listings announced as they are confirmed. Early adopters who register below are told first.

Early adopters, by conversation.

Metastability is being built with a growing consortium of independent stakeholders rather than alone. We are opening a small number of early conversations.

• Founding underwriters — with real capital and a view of their own.
• First issuers — projects that want to be judged on evidence.
• Strategic partners — with DeFi, security and agent-infrastructure depth.
• Auditors and counsel — smart-contract audit and digital-asset law.