Should Kusama spend its treasury onboarding members instead of selling KSM?

Kusama is sitting on roughly 838,000 KSM (~4.5% of all KSM) and its default habit is to spend it down into dollars on grants that never come back; there is now a working alternative — use that KSM to pay the tiny on-chain cost of turning real people and real collectives into members of the network — which converts a shrinking spending account into a growing reason to hold and use KSM.

And there is no line between taking part and joining. This document argues for turning real people into members of the network — and taking part does exactly that. Your passkey is the join. The first time you comment on, suggest a change to, or endorse this proposal — or simply show up to a Chaos Session and sign in — it enrols you as a member of Birdbrain (community N° 1786), a real on-chain Kreivo membership. The rule is the same on every surface: participation is membership. The process, the participation and the product are one thing — to take part is to become one of the citizens this document is about.

Imagine a town with a large savings account. Right now the town council spends from that account by writing cheques to contractors, and to pay the cheques it sells the town's own shares on the open market. Every cheque makes the savings account smaller, pushes a few more shares onto the market, and — once the work is delivered — leaves nothing behind that keeps earning. The town gets poorer and its shares get cheaper, one cheque at a time.

That is, roughly, how the Kusama treasury works today. Kusama is a blockchain network. Its treasury holds a large pile of its own token, KSM. To fund things, it hands out KSM as grants; recipients usually sell that KSM for dollars to pay salaries. The money leaves. It does not return.

Here is the alternative this document puts up for discussion.

Kusama's sister network, Kreivo, has quietly built the machinery to do something a blockchain has never been able to do easily before: sign a real person up as a member in one tap on their phone — no seed phrase, no tokens in their pocket, no crypto knowledge at all. And it can sign up whole collectives as well as individuals — a co-op, a music label, a football club, a village.

Each of those sign-ups has a small, known cost paid in KSM (on the order of a third of a KSM per person). So the proposal is simple to state:

**Instead of spending the treasury down into dollars, spend it into memberships.**

Every KSM used this way does three things a grant cheque does not:

1. It stays inside the ecosystem instead of being sold for fiat.
2. It leaves behind a member — an account, with a reason to keep using and holding KSM.
3. It is measurable. "How many members did we acquire, and how many stayed?" is a harder number to fake than "how many grants did we disburse?"

The rest of this document asks the obvious follow-up questions. Is that actually affordable? What is a new member even worth? Does it create real demand for KSM, or just shuffle KSM around? And what has to be true for it to work?

This idea would have been a fantasy two years ago. Five things shipped (or nearly shipped) that make it concrete now — and a sixth, Bloque, is already running real money through them. Each is explained assuming you have never heard of it.

2.1 Kreivo — the membership machine

Kreivo is a parachain — a specialised blockchain that plugs into Kusama for security. (Kusama is the "relay"; parachains are the spokes.) Kreivo's job is communities and membership. Over the last two years its team shipped the on-chain parts that matter here:

• pallet-communities — lets a group stand up a DAO (a member-run on-chain organisation) with its own treasury, voting and membership roll.
• pallet-pass — the important one. It represents a membership as an NFT that carries a built-in "gas tank", so a member can transact without holding the network token themselves and without managing keys directly. Its security audit (by Guvenkaya, with the Polkadot Assurance Legion) is complete. The tank is seeded at enrollment with a small KSM top-up (~0.02 KSM in the reference flow) and has a configurable capacity/refill rate — so it is prefunded, not an open-ended liability. Keeping it topped up over time is a real recurring cost, but that cost is precisely the "genuine gas demand" the subsidy is meant to manufacture (§4.5): once the subsidy ends, a productive collective funds its own tank — which is the point, not a hidden bill.
• pallet-payments — reversible, card-like payments for real-world goods, with a configurable fee that can route a slice back to the community.

In plain terms: the code that turns "a person" into "a member with an account and a spending allowance" already exists and is audited. The proposal is not asking anyone to build it. It is asking what the treasury should do with it.

Why "buy" a membership at all? On Kreivo, memberships live in a pool. Standing up a community costs a small refundable registration deposit (0.5 KSM), and adding a member means acquiring a membership item from the pool (currently listed at ~0.3 KSM). Those KSM amounts are the "tiny on-chain cost" from Layer 1. They are small, known, and — crucially — they move KSM within the ecosystem rather than out of it. And the ~0.3 KSM is not revenue to a vendor: the open pool is owned by the Kreivo Treasury (F3opx…pg29), so the payment moves KSM from one on-chain public account to another — Kusama's treasury envelope to Kreivo's — never to Decent or any private party — and the pool's owner is publicly checkable on-chain.

2.2 Passkeys / Virto Connect — the reason ordinary people can now join

Historically the single biggest blocker to onboarding anyone non-technical onto a chain was this: to do anything, a user had to create and safeguard a seed phrase and hold the native token to pay for gas. That is a wall most normal humans never climb.

Virto Connect is a drop-in sign-in component (<virto-connect>) that removes the wall. It uses passkeys — the same face/fingerprint mechanism you already use to log into your phone — to create and control a Kreivo account. Combined with pallet-pass's gas tank, the user needs no seed phrase and no tokens. The team's phrase for it is "one gesture → a citizen." A live demo runs at demo.virto.dev.

This is what makes "buying citizens" a real onboarding funnel and not a thought experiment: the friction that used to make mass onboarding impossible is gone.

This page is that funnel, live. The comments below are gated by exactly this gesture. The first time you comment, suggest an edit, or endorse, your passkey doesn't only sign — it enrols you as a member of the Birdbrain collective (community 1786), with a real on-chain Kreivo membership. To argue back with this document is to become one of the citizens it argues for. The mechanism isn't hypothetical; you're standing in it.

2.3 VOS — the reason real organisations will actually use it

A public blockchain is, by default, radically public: every balance and every vote is visible to the world. Most real organisations — a business, a co-op, a fund — cannot operate that way. They need private books.

VOS (the Virto / Virtual Operating System, the successor to an earlier runtime called "Valor") is the backend layer that fixes this. It combines Noir (zero-knowledge proofs — a way to prove something is true without revealing the underlying data) with Matrix (an encrypted messaging/state layer) so that a community can vote and manage funds privately while still living on a public chain. Under the hood it is an actor runtime (javm + libp2p + a local store) that lets the same community "app" pick its own consistency/trust model.

Why it matters to this proposal: onboarding subsidy only pays off if the members you onboard are real economic actors — shops, labels, clubs — rather than crypto hobbyists. VOS is the piece that makes a public chain acceptable to those actors.

2.4 Federate — why the unit of onboarding is the collective rather than the person

Notice that the proposal keeps saying "people and collectives." That is deliberate, and it comes from how this network is designed to grow.

Federate is the growth model: each community owns its own data and its own intelligence, and communities **link up by treaty — federating upward — rather than being controlled from the centre downward.** (The design slogan is "federate up, do not delegate down.") A hundred local cells can pool what they choose to pool, while each stays sovereign over its own members and books.

The consequence for the treasury: the natural thing to subsidise is not a lone individual but a collective that brings its own members with it. One onboarding event can create a co-op with twenty members. That changes the arithmetic — the registration deposit is shared across the group, so cost-per-member falls as collectives get bigger (this is the 0.3 + 0.5/N in Layer 4).

2.5 Birdbrain — the reason a membership is productive rather than a headcount

A member who signs up and does nothing is worth almost nothing. The final piece is what makes membership do something.

Birdbrain is the network's intelligence layer — a local-first application that sits with a community and turns its activity (calls, notes, decisions) into a shared, queryable world model the community owns. It already runs in production: in the network's recurring "Chaos Sessions", Birdbrain now joins as a live, on-device attendee that transcribes and cross-references prior sessions — and it uses passkey attestation to record who was actually there. That sign-in is itself an on-ramp: showing up to a Chaos Session and signing in with your passkey enrols you in the live on-chain "birdbrain" community (N° 1786) — the very same membership that commenting on this page mints, reached from the room instead of the page. Each member gets a deterministic on-chain "seed" — a relational identity anchor — and a returning member re-anchors that same seed rather than minting a second one, so a person accumulates presence, not duplicate memberships.

Why it matters here: it closes the loop. A subsidised membership isn't a dead entry on a list — it plugs the member into an intelligence layer that produces something (attested attendance, contribution provenance, a community model) which is itself valuable and, over time, monetisable. Federate pools it; Birdbrain produces it; Kreivo anchors who owns it.

2.6 Bloque — proof the stack already carries real commercial traffic

Everything above is capability. Bloque is what turns capability into evidence: a fintech running its payment SDK/API on top of Kreivo, settling in dUSD — the dollar stablecoin native to this stack, issued on Kusama Asset Hub by Decent Partners on Brale's regulated rails. In user terms, dUSD has the stable-dollar convenience people expect from USDC and lets a user pay fees without holding KSM. But the political economy is different: instead of reserve yield disappearing into opaque, privatised structures, dUSD is designed so yield can recirculate as public funding into an emerging ecosystem of on-chain collectives. On top of it, Bloque issues VISA debit cards and runs an Asset-Hub stablecoin exchange that off-ramps dUSD to USDC and bank accounts.

What matters here is not the product sheet but the traction. Bloque publishes a live metrics dashboard (metrics.bloque.sh); as of the week of 6 July 2026 it shows:

That is a near-standing-start to ~47,000 real users in about a year, on the same Kreivo/Kusama infrastructure this proposal is about. It matters here for one blunt reason: the demand side is not a thought experiment. A fast-growing commercial business is already moving real money across the US and Latin America on these exact rails — which is precisely the kind of productive member (§4.4) the subsidy is meant to manufacture at scale.

Putting the five together

The one-line synthesis: the network can now onboard a real collective in one gesture, give it private books, let it keep and federate its own intelligence, and anchor all of it on Kusama's security — and the only scarce input the treasury has to supply is a little KSM.

And this is no longer theoretical. Bloque (§2.6) is already moving real money across the US and Latin America on exactly these rails — tens of thousands of users, growing weekly. The machine is built and it has its first commercial engine running on it.

To judge whether "buy members" beats "sell KSM," you need a shared picture of how a blockchain captures value at all. Here it is, briefly, for a general reader.

3.1 The seven ways a token network can capture value

1. Security / staking — people lock the token to secure the chain and earn rewards. Captures value by taking supply off the market.
2. Blockspace / coretime — people pay to use the chain's capacity. On Kusama/Polkadot this is coretime (buying blockspace).
3. Fees / burns — transactions pay fees; some designs burn (destroy) a portion, making the token scarcer as usage rises.
4. Collateral / liquidity — the token is used as backing or trading liquidity elsewhere.
5. Governance / access — holding or spending the token grants rights, membership, or access.
6. Retained protocol revenue — the network keeps a cut of the economic activity that happens on it.
7. Narrative / legitimacy — the token becomes the symbol and coordination point for a believable future: a story that attracts builders, users, capital, partners and political support. Narrative does not replace the six mechanisms above; it makes people care enough to use, fund and defend them.

A healthy network turns usage into token demand through several of these at once, while narrative gives the usage a shared direction. The failure mode is a network whose treasury spends but whose usage does not feed back into demand — or whose story floats free of any durable economic loop.

3.2 Where Kusama leaks

Kusama's treasury today is close to that failure mode, for reasons the ecosystem's own analyses have named repeatedly:

• Outflows are non-recoverable. A grant is spent, converted to fiat, and gone. There is no return leg.
• Coretime and transaction fees generate negligible revenue. Channel (2) and part of (3) are, in practice, close to zero.
• Loans return principal but don't diversify. They help, but they don't build a non-KSM reserve or create new users.
• Success metrics are gameable. "Grants disbursed", "milestones hit", "marketing impressions" are easy to inflate and hard to tie to network value.

The net effect: the treasury draws down, each drawdown adds a little sell-pressure to KSM, and little of it comes back as durable demand for the token. This is the "town selling its own shares to write cheques" picture from Layer 1, stated precisely.

3.3 Re-reading the burn change: mistake, or unfinished policy?

Kusama used to run a treasury burn — a mechanism that destroyed a slice of unspent treasury funds each spend period, acting as a deflationary counterweight to the KSM the network mints every block. That policy changed: after Kusama referendum #437, the Polkadot Fellowship shipped runtimes PR #511, replacing the hard-coded burn with governance-controlled BurnParameters and setting the default burn rate to 0%.

In isolation, that looks like a mistake. If the treasury is merely a spending pot, turning off the burn means the network keeps minting KSM into a fund that can later be sold into fiat-bound grants. More inflation, more inventory, same weak return loop.

But the same policy looks different if the treasury is redefined as a citizen-subsidy reserve. In that frame, newly minted KSM is not surplus to destroy and not cash to spend. It is onboarding capacity: a stock of native capital that can be converted into memberships, collective accounts, gas allowances, governance rights and productive economic actors.

So the burn change is best read as unfinished policy. It removed an automatic sink, but did not yet answer the positive question: what should the no-longer-burned KSM become? This document proposes one answer: turn it into citizens — not metaphorically, but as user-owned Kreivo memberships attached to real collectives.

3.4 What "buy members" does to that picture

The onboarding-subsidy model attacks the leak through channels (1), (5), (6) and (7) at once:

• It converts treasury KSM into governance/access (channel 5) — every subsidised membership is a granted right to participate.
• It creates the conditions for retained revenue (channel 6) — members transact through pallet-payments, whose fee can route a slice back to community (and, potentially, network) accounts.
• Over time it feeds security/staking (channel 1) — productive members and collectives are exactly the actors who end up holding and locking KSM.
• It gives Kusama a stronger narrative (channel 7) — not merely “the canary network with a treasury problem”, but a live experiment in using protocol wealth to create a membership economy of on-chain collectives.

Most importantly, it changes the direction of the money. A grant sends KSM out (to fiat). A subsidised membership moves KSM sideways and keeps it in-ecosystem — but two distinct movements happen and it is worth not conflating them: the ~0.3 KSM item price is paid to the Kreivo Treasury (F3opx…pg29), the public on-chain account that owns the open membership pool, while the membership item itself (the NFT) lands in the onboarded collective's own sovereign account. Cash to a network treasury, asset to the community — not KSM working-capital handed to the collective. And it plants something that pulls KSM in later (co-payments, gas top-ups, renewals, commerce fees). We make that "in" rigorous in Layer 4.

3.5 What Polkadot chose — and why Kusama should choose differently

While this document was being written, Polkadot answered §3.3's open question for itself. Anyone who follows Polkadot governance will otherwise wonder why it is missing, and the two answers are genuinely different rather than rival versions of the same one.

Polkadot made three moves in March 2026:

1. It capped supply. Referendum 1710 fixed a maximum of 2.1 billion DOT. On 14 March annual issuance fell from 120M to ~55.8M DOT — a 53.6% cut — on a schedule that steps down again every two years.
2. It replaced the burn with a pool. Referendum 1827 established the Dynamic Allocation Pool (DAP): a permanent on-chain account collecting newly issued DOT together with transaction fees, coretime revenue and slashes, which governance then splits across separate budgets — validators, stakers, treasury, reserve. Treasury burns stopped. In the first split, ~12.6M DOT/yr funds validator self-stake incentives, ~25.2M covers staker payouts, and ~18M is retained in the buffer as reserve, flagged as possible backing for a future stablecoin (Polkadot Forum, 2 April 2026).
3. It started buying users. Referendum 1783 — the Polkadot People Initiative, passed at 87.7% aye and executed — spends $3,009,600 of stablecoin reserves on incentives for people verified by Project Individuality, Polkadot's native zero-knowledge proof-of-personhood system: $200/$100/$50 for the earliest joiners, up to ~$10 a week for useful governance participation, and the majority into a weekly draw of roughly 140 prizes of $250 with $2,500 monthly stars. No wallet, no tokens, no identity documents. The stated end state is a "Polkadot New Deal" in which verified individuals supply cheap security so issuance can keep falling.

Two things follow.

First, Polkadot and this document agree on the diagnosis. Both conclude that a treasury should stop buying projects and start buying people. The argument here is no longer heterodox.

Second, they disagree on the instrument, and that disagreement is what there is to debate. Polkadot pays individuals in dollars for proving they are human. This RFC pays in KSM for a collective to hold a membership.

By this document's own standard (§3.2), paying cash to individuals sits closer to the failure mode it names than to the cure: the money is converted and gone, and the return leg depends entirely on whether a lottery-acquired cohort stays. That is not a cheap shot — it is the same retention question we hold ourselves to in §4.4 and §5.3, and neither network has the cohort numbers yet. But it is why "spend it into memberships" is not the same proposal in a different currency.

Where Polkadot is plainly ahead is uniqueness at the gate. Individuality's mechanisms — Proof-of-Ink and Proof-of-Video-Interaction, built on the Web3 Foundation's ring-VRF cryptography — are a stronger answer to §5.2 than what this proposal's join path uses today, which is passkey attestation and an attendance record. That is a statement about the current implementation, not a claim that the question is closed. pallet-people deliberately does not verify personhood itself: it stores verified humans, sorts them into rings, mints per-context aliases and validates proofs, and it outsources verification to pluggable verifiers — Decentralized Individuality Mechanisms, in Polkadot's own framing. Proof-of-Ink is one such mechanism, not the system, and that interface is open by design.

Two things sit above it that a registration check cannot supply. Uniqueness is not attribution: ring-VRF aliases are unlinkable across contexts by construction, which is excellent privacy and structurally unable to carry the value of a contribution back to the person who made it — the thing §2.4 and the seed depend on. And a one-shot proof is a one-off cost: pay for the ink and the video once, and the identity is permanent and free to operate forever, whereas a proof derived from continuing, witnessed participation costs more the more it is asked to claim. Those are the gaps a participation-derived mechanism fills, and they sit on top of a personhood registry rather than against it — so this RFC does not propose to replace Individuality, it assumes something like it underneath. Which is why Kusama's own attempt to deploy those pallets natively, referendum #612, rejected at 42.1% aye, reads to us as the wrong call and would have our support if it returns: proof-of-personhood on Kusama is a prerequisite for the stronger mechanisms, not a competitor to them, and Kusama already funds a Proof of Personhood Vision bounty aimed at exactly that problem.

Where Kusama is behind is not technology but decision. Polkadot resolved its monetary question; Kusama has rejected every attempt to resolve its own — the burn-based tokenomics wish for change (#627) rejected at 28.2% aye, the halving-and-maximum-supply wish (#625) timed out at 8.2% — while inflation runs near 7% and the burn stays at 0%. That is the vacuum §3.3 describes, and it has now lasted a year.

This RFC does not try to fill it. It is deliberately not a monetary proposal: it changes no issuance curve, sets no cap, adds no burn. Kusama's supply-side reformers work the supply; its validator-set reformers work the cost; this is the demand side — what the KSM the network already holds should buy. Those are complements, not competitors, and the network could adopt all three.

This layer is for readers who want to argue with the arithmetic. Every figure is sourced or reproducible (see Appendix). Prices move; treat KSM/USD as illustrative.

4.1 What the treasury actually holds

The Kusama treasury is not one number in one place. As of this writing it is roughly:

That ~838k KSM is roughly 4.5% of the ~18.49M circulating supply. This is the working capital the proposal is talking about.

4.2 What one membership costs

From the live Kreivo machinery (the figures the enrollment code actually encodes):

• Community registration deposit: CommunityDepositAmount = UNIT/2 = 0.5 KSM (refundable; paid once per collective).
• Per-membership item: currently listed at ~0.3 KSM — the open-pool price, paid to the Kreivo Treasury (F3opx…pg29) which owns the pool. This is a public on-chain account, not a vendor, and not Decent.

So the cost to onboard one member inside a collective of size N is:

cost_per_member(N) = 0.3 + 0.5 / N      (KSM)

This is exactly why Federate (Layer 2.4) matters to the maths: bigger collectives amortise the deposit toward the ~0.3 KSM floor.

4.3 The headline subsidy: is 1m memberships even affordable?

The stated ambition is to bootstrap 1,000,000 memberships — a round, memorable target. At the collective-of-20 rate:

1,000,000 members × 0.325 KSM = 325,000 KSM

That is ~39% of the treasury (~$1.07M), leaving ~513,000 KSM untouched. In other words, the entire headline onboarding target fits inside roughly a third of the fund, at current prices.

And that headline is nowhere near the ceiling. Spending the whole ~838,464 KSM treasury the same way would subsidise:

838,464 KSM ÷ 0.325 KSM  ≈  2.58 million members   (collectives of ~20)
838,464 KSM ÷ 0.30  KSM  ≈  2.79 million members   (large collectives, deposit ≈ negligible)

So the treasury could, in principle, buy the entire 1m target roughly two and a half times over. Affordability is emphatically not the binding constraint. The binding constraints are (a) whether those members stay, and (b) whether they generate return demand — which is the rest of this layer.

4.4 What is a new account worth? A four-stage funnel

A raw sign-up is not worth 0.325 KSM of value; it costs that. Its worth depends on how far it travels down a funnel:

1. Created — the account exists. (Cost incurred: ~0.325 KSM.)
2. Activated — the member did something real once (a payment, a vote, a contribution).
3. Retained — the member is still active after some months.
4. Productive — the member (or their collective) generates recurring economic activity: payments with fees, gas top-ups, staking.

Only stages 3–4 create the "return demand" that justifies the spend. So the honest acquisition metric is cost per retained member, not cost per sign-up:

cost_per_retained = cost_per_member / retention_rate

at 40% retention:  0.325 / 0.40 = 0.8125 KSM per retained member
at 20% retention:  0.325 / 0.20 = 1.625  KSM per retained member

The whole model lives or dies on retention and productivity — which is precisely why Birdbrain (something for members to do and own) is load-bearing, not decorative. It is also why Bloque (§2.6) matters as evidence: stage-4 productive behaviour — real payments, card spend, dUSD off-ramps — is already happening on this stack at the scale of tens of thousands of accounts, and not on a slide.

4.5 The distinction that makes or breaks the argument: recycled vs. genuine demand

This is the single most important idea in the document, and it is where a sloppy version of the argument would cheat.

• **Treasury-recycled KSM is not new demand. When the Treasury buys a membership item, KSM moves from a Treasury account to a community account. That is an intra-ecosystem transfer. It reduces sell-pressure (the KSM isn't dumped for fiat) but it does not** by itself create net new buying of KSM. Counting it as "demand" would be double-counting.
• Genuine external demand is KSM that flows in from outside the subsidy, i.e. someone choosing to acquire KSM they did not previously hold. The subsidy's job is to manufacture the conditions for this. It comes from four places:

1. Co-payment — members/collectives paying part of their own onboarding or upgrades. 2. Recurring gas / operating budgets — communities topping up their gas tanks and treasuries. 3. Renewals / expansions — memberships that recur, collectives that grow. 4. Commerce-linked fees & buybacks — pallet-payments volume whose fees can be used to buy and retain KSM.

So the correct claim is not "subsidising 1m members creates demand for 325k KSM." It is: subsidising 1m members spends 325k KSM of recycled, in-ecosystem capital in order to stand up a member base whose ongoing behaviour produces a stream of genuine external KSM demand. The subsidy is the fixed cost; the demand is the annuity.

4.6 Modelling the annuity: demand past the subsidy point

Suppose the subsidy succeeds and leaves a retained, productive base. What recurring external KSM demand does that base pull per year? Three scenarios, deliberately spanning pessimistic to optimistic:

These are flows, not one-off. The point is qualitative and robust to the exact inputs: a grant creates a one-time outflow; a member base creates a recurring inflow. Even the conservative flow, compounding annually against a fixed one-time subsidy cost, inverts the treasury's trajectory from "draws down forever" to "spends once, earns thereafter."

4.7 What this does to KSM demand over time (the whole argument in one shape)

• Today (grants): treasury ↓, small persistent sell-pressure, no return leg. A depreciating spend.
• Onboarding subsidy, during the spend: treasury ↓ but KSM stays in-ecosystem (recycled, not sold); a member base accumulates.
• After the subsidy: the member base emits recurring external demand (§4.6). The one-time cost is behind you; the annuity is ahead of you.

The proposal, reduced to its economic essence, is: convert a depreciating spending account into a one-time investment that installs a growing KSM sink. Whether the sink is big enough to justify the spend is exactly what §4.4–4.6 put on the table for the community to pressure-test.

A discussion document earns its name by making the case against itself well. The strongest objections:

1. "Recycled KSM to community accounts can still be sold." True. A community that receives KSM can dump it. The subsidy reduces forced sell-pressure (grantees selling to make payroll) but does not eliminate discretionary selling. Mitigation lives in gas-tank design, vesting, and giving KSM a use (gas, staking) rather than only a price.
2. "Sybil / farming." If onboarding is nearly free and subsidised, you invite fake members and airdrop-farmers. This is why the model insists on retained, productive members (§4.4) and why passkey attestation + Birdbrain's real-attendance record matter — they raise the cost of faking a productive member well above 0.325 KSM.
3. "Retention is unproven." The entire annuity depends on retention/productivity rates the network has not yet measured for subsidised members at scale. The right response is a staged pilot with published cohort-retention numbers, not a 1m commitment on day one. The stack is not starting from zero, though: Bloque (§2.6) already shows tens of thousands of real users transacting on Kreivo/dUSD, so the open question is the retention of subsidised members specifically — not whether anyone will use the rails at all. And it is being tested with real money now, not just argued: we have a pipeline of collectives lined up to onboard and have already begun buying KSM ourselves to subsidise their memberships — so the live question is less would it work than at what point, and at what size, should the treasury co-fund what we are currently bootstrapping by hand.
4. "It's still spending the treasury." Yes. The claim is not "free"; it is "better-shaped spending with a return leg." Reasonable people can hold that the current grant model, improved, is preferable — that belongs in the debate.
5. "Aren't you just building this for yourselves?" Decent Partners runs part of this stack (dUSD/Kreivo/Bloque/Birdbrain) and is a proponent — so it is fair to ask. But we custody nothing: the ~0.3 KSM moves between public on-chain accounts (never through us — see §2.1), and members now hold their own keys via self-custodied passkeys, so we hold no balances and no signing rights on anyone's behalf. The design is deliberately stablecoin- and team-agnostic — anything one team can do here, another should be able to copy. If it only works when we run it, it is the wrong design, and we would want to know.
6. "Governance load." Standing up hundreds of thousands of memberships has on-chain and human-governance costs not modelled above.
7. "Polkadot is already doing this — copy Individuality and run an airdrop." Fair, and §3.5 sets out what Polkadot actually did: $3m of stablecoin paid to proof-of-personhood-verified individuals, behind a stronger check at the gate than this proposal's join path runs today. Three reasons to do this on Kusama rather than mirror it. Currency: their spend leaves the ecosystem as dollars; ours moves KSM sideways and returns as gas, co-payment and fees — decisive for a network whose treasury is denominated in its own token and which has no stablecoin reserve to hand out. Unit: an airdrop recruits atomised individuals, and a cohort attracted by a weekly draw is the hardest of all cohorts to retain; a collective arrives with its economic activity already attached. Ownership: an airdrop leaves a balance, a membership leaves an account in a community that owns its own books and data (§2.4). The honest version of this objection is not "copy it" but "take the good half" — Kusama should want native proof-of-personhood and the membership route. These are different layers, not rival products: pallet-people outsources verification to pluggable mechanisms (§3.5), so a stronger check at the gate and a participation-derived identity above it compose rather than compete. We would rather argue for both than defend a weaker gate out of pride.

Open questions for the community — the first is the one that matters most now: What should the initial bounty request be for? Given a real, growing pipeline and members already being onboarded, what size and scope should a first rule-bound envelope (§6.2) be — enough to prove retention at scale without over-committing the treasury before the numbers are in? Then: What retention rate would make this clearly worth it? Should the fee on pallet-payments route a slice to the network, and if so how much? Who counts as a "collective" worth subsidising?

Everything above argues what the treasury should buy. This layer answers the mechanical question a sceptic asks next: once the community agrees, how does KSM actually leave the treasury and become a membership — without a fresh governance vote every time, and without handing anyone a blank cheque?

The answer turns on separating two things that are usually run together.

6.1 Authorisation is governance; execution can be permissionless

One thing here cannot be made permissionless, and pretending otherwise would be the dishonest version of this argument: spending treasury KSM at all requires the electorate's consent. KSM holders own the treasury; only they can authorise it to spend. That is a feature, not a bug.

But "authorise" and "execute" are different acts:

• Authorisation — may the treasury spend up to X KSM, on these rules, for this purpose? — is inherently a governance decision, taken once, by vote.
• Execution — given the rules already approved, fund this specific membership now — is bookkeeping. It does not need a vote each time, provided it stays inside the approved envelope and every payout is checkable after the fact.

So "permissionless draw-down" does not mean "anyone can spend the treasury." It means: the community approves a rule-bound envelope once, and thereafter memberships are funded from it automatically, by formula, without returning to governance per payout. The permission is granted at the envelope; the draws inside it are mechanical.

6.2 The instrument: a standing Onboarding Bounty

Kusama already has the exact primitive for "approve a pot once, pay it out many times without re-voting": pallet-bounties with pallet-child-bounties. A bounty is a treasury-funded envelope with a named curator who disburses against it; child-bounties are the individual payouts, and no fresh referendum is needed per payout.

This is not hypothetical for Kusama. Its own Vision bounties — Proof of Personhood, Art & Social Experiments, Zero-Knowledge — run on precisely this shape: a large standing envelope, approved once, drawn down by curators over time, with no vote per disbursement. The proposal is to add one more, purpose-built:

The Onboarding Bounty. A KSM envelope, authorised by the Wish For Change this RFC becomes, whose sole permitted use is paying the on-chain cost of memberships at the fixed formula of Layer 4. Birdbrain is its curator.

And “curator” understates it: the Onboarding Bounty is itself Birdbrain. The collective that authors this RFC, that you join by commenting on it, and that each subsidised membership is minted into is the very thing that draws the bounty down. There is no division between the process (this document and its debate), the participation (your passkey enrolling you as a member), and the product (the bounty minting the next member) — they are one loop, run by one collective. Birdbrain proposes the facility, Birdbrain is the facility, and Birdbrain is what the facility produces.

A precondition we said we would not paper over — now met. “Birdbrain is the collective, not us” is only honest if members hold their own keys — otherwise whoever holds the keys holds the votes, and the “keyless” sovereign account (community 1786) is really the operator in a mask. So self-custody is not a nice-to-have; it is the thing that makes a trustless curator true rather than rhetorical. Status, stated plainly: as of this week the live join path is self-custody by default. One passkey tap creates a Kreivo Pass account whose authenticator is the member's own device, with the first registration sponsored (paid-for, not held) so a new member needs no KSM — and Decent holds, derives, or can reconstruct no member's signing key. (A custodial derived-key path remains only as a transitional fallback for the moment, and is being retired.) The trustless-curator property above is therefore one we are now asserting, not merely a target — and you can test it yourself: comment on this page and your membership is minted to an account only you control.

6.3 A curator with no discretion over amounts

The usual objection to bounties is curator trust — a curator with discretion is a person who can be captured. This design removes the dangerous discretion entirely: the curator has zero say over how much each membership costs.

The price is not a judgement call; it is the published formula from §4.2:

cost_per_member(N) = 0.3 + 0.5 / N      (KSM)

The curator cannot pay 1 KSM for a 0.325 KSM membership; the rule is both the ceiling and the floor. What the curator does decide is the one thing a formula cannot: is this a real, eligible member? That single judgement — eligibility, not amount — is what §6.4 makes cheap and checkable. Everything else is arithmetic the chain enforces.

Birdbrain is a thin rule-executor, not a treasurer. It answers one yes/no question per applicant and, on "yes", triggers a payout whose size it cannot alter.

6.4 Eligibility: attested up front, challengeable after

Who counts as an eligible member? The stack already produces the signal:

• Passkey attestation — a membership is minted by a real device gesture (Face / Touch ID), not a script. This is the same primitive Birdbrain already uses to record who actually attended a Chaos Session (§2.5).
• Proof of Real — the network's test for whether an account is a genuine, specific, 1:1 human rather than a farm or a bot.

Together these are the eligibility oracle: the curator's yes/no is mostly read off these signals, which is why it can be automated rather than adjudicated by hand.

Crucially, eligibility is optimistic, not gate-kept. Reviewing every membership up front would reintroduce exactly the bottleneck we are trying to remove. Instead — borrowing the co-witness pattern the network already uses for consented capture — a claim is funded optimistically, and the curator (or any watcher) holds a challenge window in which a fraudulent claim can be vetoed and clawed back. You do not ask permission to onboard; you onboard, and a fraud can be undone. This keeps the common case (a real person joining) instant and the abuse case (a farm) reversible.

6.5 The rail: from Asset Hub envelope to a Kreivo membership

The KSM lives on Kusama Asset Hub; memberships are minted on Kreivo. The bridge between them already works. A single XCM call — polkadotXcm.transferAssetsUsingTypeAndThen, forcing LocalReserve because Asset Hub is KSM's reserve for Kreivo — moves KSM from the Asset Hub envelope to the operator account on Kreivo. Then the three membership calls fire:

(Asset Hub)  polkadotXcm.transferAssetsUsingTypeAndThen   →  KSM lands on Kreivo
(Kreivo)     CommunityMemberships.buy_item → transfer      →  Communities.add_member

Both halves are proven: the XCM leg dry-runs clean and forwards exactly one message to Kreivo (parachain 2281); the three-call enrollment flow has been rehearsed on a chopsticks fork and against live metadata, and needs only KSM (Appendix). Nothing here needs to be built — it needs to be authorised.

6.6 The flywheel: a facility that partly refills itself

A one-off envelope drains. The point of a standing facility is that the members it buys generate the return demand of §4.6 — and a slice of that return can be routed back to keep the system alive. Seniority matters here, and it is a deliberate design choice: returning pallet-payments fee revenue refills a collective's own member gas tanks _before_ it refills the facility. Members first, facility second — because if the facility refills but the members' tanks run dry, the members are what churns, and retention is the variable the whole annuity rests on (§4.4). A subsidy that keeps the account registered but lets it run dry the day it graduates is not an annuity; it just defers the churn to graduation day.

• pallet-payments fees from subsidised members' commerce (§3.4, channel 6) first top up that collective's own member gas tanks, then the envelope.
• Buybacks funded by that fee revenue (§4.5) refill it in KSM.
• Co-payments and renewals (§4.5) mean later members increasingly fund themselves.

Gas cost per active member per year is a named pilot metric. Kreivo fees are sub-milli-KSM per transaction, which puts it at order ~0.01–0.1 KSM per member per year at realistic light volumes (tens to low-hundreds of transactions) — small in absolute terms, but real for a collective that holds no KSM and has no inflow. It is unmeasured for subsidised members at scale, so it ships as a pilot number to be published, not asserted.

The honest claim is partial, not perpetual-motion: the facility is not guaranteed to fully self-refill, and early on it will not. But every unit of genuine return demand routed back is a unit the treasury does not have to re-authorise — so the ask to governance shrinks over time instead of recurring at full size. That is the flywheel of the title, made mechanical: spend once to install a base; let the base refill the facility that installs the next.

6.7 Why a bounty, and not the alternatives

Three mechanisms could deliver a rule-bound draw-down. The bounty wins on precedent:

The bounty is the only one of the three the Kusama electorate has already seen work at scale. That is why the on-chain ask (see the front-matter Wish-For-Change note, and §6.1) is "approve the standing facility," not "approve a one-off spend."

6.8 Where this sits: RFC → Wish for Change → bounty

This document is deliberately the first of three governance steps, in order, so that no one mistakes a discussion doc for an ask:

1. RFC — Request for Comment (now). This page and its debate. Pressure-test the logic, the numbers, the design. No ask, no vote.
2. WFC — Wish for Change. If the idea survives contact, an on-chain referendum on the Wish for Change track establishes a mandate — does Kusama actually want to direct treasury this way? — without yet moving any funds. Submitting the wish is permissionless; approving the facility needs KSM votes, and the wish carries this document's hash so the on-chain ask is verifiably the community's text.
3. Bounty proposal. Only with that mandate do we bring the concrete, rule-bound Onboarding Bounty of §6.2 for a specific figure — the "how big" question of Layer 5.

We are at step 1. Nothing here commits anyone to anything.

Treasury figures. Bucketed from the treasury's live sources (relay free balance; Asset Hub treasury; deployed Hydration loan). Cross-checkable at kusama.dotreasury.com. Circulating supply / price from coinmarketcap.com/currencies/kusama. All USD figures are illustrative at ~$3.28/KSM.

Cost constants. CommunityDepositAmount = UNIT/2 = 0.5 KSM and the ~0.3 KSM open-pool membership price are the exact values encoded by the enrollment machinery (file decent-kreivo/dryrun-full.ts, file decent-kreivo/src/memberships.ts). The full three-call flow (CommunitiesManager.register → CommunityMemberships.buy_item/transfer → Communities.add_member) has been rehearsed against live metadata and on a chopsticks fork; the flow needs only KSM.

Reproduce the cost model. cost_per_member(N) = 0.3 + 0.5/N; headline = 1,000,000 × 0.325 = 325,000 KSM; whole-treasury capacity = 838,464 / 0.325 ≈ 2.58M members (or 838,464 / 0.30 ≈ 2.79M at the bare item price); cost_per_retained = 0.325 / retention. The demand scenarios (§4.6) are flow models over a retained base — substitute your own per-member assumptions.

Disbursement mechanism (Layer 6). The permissionless draw-down uses pallet-bounties + pallet-child-bounties — the same standing-envelope-plus-curator pattern Kusama already runs for its Vision bounties (Proof of Personhood, Art & Social Experiments, Zero-Knowledge), where a single referendum approves a pot that curators disburse over time with no vote per payout. The Asset Hub → Kreivo funding leg is a single polkadotXcm.transferAssetsUsingTypeAndThen call forcing LocalReserve (Asset Hub is KSM's reserve for Kreivo, so auto-detecting transfers fail with InvalidAssetUnknownReserve); it dry-runs clean and forwards exactly one XCM to Kreivo (parachain 2281) — see file decent-kreivo/src/asset-hub-funding.ts. Execution on Kreivo is the same rehearsed CommunityMemberships.buy_item → transfer → Communities.add_member flow above. Eligibility is read from passkey attestation + Proof of Real, funded optimistically with a challenge window (the co-witness veto pattern) rather than gated by up-front review.

Treasury burn policy. The treasury-burn default was set to 0% via governance-controlled BurnParameters: community mandate in Kusama referendum #437 (~99.7% aye), implemented in polkadot-fellows/runtimes PR #511 (default burn rate 0%). The interpretation in §3.3 is that this is weak policy if the treasury remains a spending pot, but coherent if the treasury becomes a citizen-subsidy reserve.

Bloque / dUSD. Bloque runs a payment SDK/API, VISA debit-card issuance and an Asset-Hub stablecoin exchange on Kreivo, settling in dUSD — a Kusama Asset Hub stablecoin issued by Decent Partners on Brale's regulated rails. It gives users USDC-like stable-dollar convenience and fee payment without KSM, but with a different value-capture logic: yield is intended to recirculate as public funding for on-chain collectives rather than disappear into opaque, privatised reserve structures (Kusama's treasury acquired dUSD and set it a Sufficient Asset in referendum #526). Live traction from Bloque's own dashboard (metrics.bloque.sh, week of 2026-07-06): ~46,700 end-users (from 23 in July 2025), ~54,300 accounts across virtual wallets, US bank accounts, cards, Colombian rails and Polygon.

Developments referenced. Kreivo pallets and <virto-connect> (Local Incentives Protocol v2, file WHITEPAPER-V2.md; audit by Guvenkaya / Polkadot Assurance Legion). VOS = Virto OS, Noir + Matrix private backend (matrix-sessions 2026-05-17 / 2026-06-08; whitepaper §2.4). Federate = "federate up, do not delegate down" (whitepaper §6). Birdbrain live-attendee + passkey attestation (Chaos Session CS18, 2026-06-01). Wiring status: file core/vault/briefs/kreivo-stack-wiring-comparison.md.

Polkadot's 2026 economic reset (§3.5). Supply cap and stepped issuance: referendum 1710 (executed; 120M → ~55.8M DOT/yr from 14 March 2026, −53.6%, stepping down every two years). Dynamic Allocation Pool: referendum 1827 (executed), budget split and timetable in Polkadot Staking Changes: Progress & Timeline (Polkadot Forum, 2 April 2026) — ~12.6M / ~25.2M / ~18M DOT annualised across validator self-stake incentives, staker payouts and the retained reserve. Polkadot People Initiative: referendum 1783 (executed, 87.7% aye; $3,009,600 in USDT/USDC, disbursement logic and Project Individuality described in the proposal text). Kusama's own attempts, for comparison: #612 native proof-of-personhood pallets (rejected, 42.1% aye), #627 burn-based tokenomics (rejected, 28.2%), #625 halving and maximum supply (timed out, 8.2%). Tallies read from the Subsquare API on 2026-08-22 and reproducible there.

Prior treasury work this builds on. The "repayable KSM loans to stablecoin issuers" Wish (file Treasury-Strategy-v-5.md) and file Focus-Spending-On-Outcomes-Over-Activity.md — this document is the demand-side complement to those supply-side proposals.

This is a discussion document, not a proposal for a vote. Its purpose is to make the argument legible enough that the community can improve it, cost it, or reject it on the merits. Corrections to any figure are welcome — every number here is meant to be checkable.