An explainer for KSM and DOT holders, prepared 22 October 2026. It covers edition 0.1.0 of the SDFI decision paper “Kusama on JAM Prime: The Infrastructure Decision” (commit da60fe6), the version linked from the live Kusama referenda.
Summary
The decision paper by SDFI conflicts with both referenda, but the two conflicts are of different kinds.
Kusama WFC #573: an open, intended conflict. #573 chose an independent, lightweight JAM for Kusama with its own validators. The decision paper proposes the opposite: Kusama as a protected service on Polkadot’s JAM instance, with its validator set and staking retired. Each of the paper’s four resolutions states that, if selected, it replaces #573’s direction. Superseding an earlier Wish for Change with a later one is a normal use of OpenGov. Voters simply need to know that this is what the current round does.
Polkadot WFC #1926: a conflict the paper does not address. #1926 requires every footprint allocation to be paid for at market price in DOT, with that DOT burned automatically. The paper’s packages do three things that cut against this. They include 450,000 JAMKB-equivalent of footprint inside a price calculated from coretime alone. They fix that price in DOT for terms of up to 20 years. And they contain no burn provision. The paper’s draft Polkadot-side resolution would ask DOT holders to accept the package without mentioning #1926. The issue was raised during the forum discussion of the earlier position paper, but it was not carried into the decision paper or the referenda.
The documents at a glance
| Document | Network | Status (22 Oct 2026) | What it says |
|---|---|---|---|
| WFC #573: Option A, lightweight and independent | Kusama | Passed, 21 Aug 2025 | Independent JAM: 32 cores, 1-second blocks, KSM as native token for storage, staking and coretime |
| WFC #574: Option B, full power and Polkadot friendly | Kusama | Rejected (same round as #573) | Full-spec JAM instance designed for Polkadot to join and share security costs |
| WFC #1926: 100% of DOT revenue from JAMKB sales to be burned | Polkadot | Executed | All DOT proceeds from footprint burned; every footprint allocation paid at market price in DOT; no footprint grants or subsidised loans |
| WFC #1927: DOT settlement for JAMKB (as described in #1926) | Polkadot | Rejected | Settlement currency stays open; under #1926, non-DOT proceeds are converted to DOT and burned |
| SDFI decision paper v0.1.0 | — | Published 19 Oct 2026 | Four cohabitation packages on JAM Prime, each 40 coretime-years for US$4.32M, paid in DOT from the Kusama Vision fund |
| WFCs #671, #672, #673, #674 | Kusama | Deciding; announced closing date 3 Nov 2026 | The paper’s four resolutions: 8, 6, 4 or 2 cores |
| WFC #676: independent Kusama on a reduced JAM, keeping its validators | Kusama | Preparing (awaiting decision deposit) | Continues #573; asks that the independent option be scoped, priced and compared with JAM Prime before validators are retired |
What the decision paper proposes
The paper, by Emiel Sebastiaan and Arief Ernst of SDFI BV, is the third in a series. It presents four options that trade capacity against duration: 8 cores for 5 years, 6 cores for 6 years and 8 months, 4 cores for 10 years, or 2 cores for 20 years. Each equals 40 coretime-years (CTYs; one core for one year). Every option carries the same footprint planning ceiling of 450,000 JAMKB-equivalent: 250,000 for system chains and shared host resources, plus 200,000 as initial support for 12 other parachains. Every option also includes protected always-accumulate resources, protected KSM control over Kusama’s service, and a Fellowship mandate for migration and maintenance.
All four have the same price: 40 CTYs × US$108,000 = US$4,320,000. That price is converted into a fixed DOT amount using the 90-day average DOT price before the selected referendum’s approval date. The amount is capped at 5,000,000 DOT; the cap holds if the average is at least US$0.864. It would be paid from eligible Kusama Vision fund resources committed by Web3 Foundation under WFC #498. Kusama’s validator set, staking system and staking issuance would be retired once the transition completes.
The authors advocate cohabitation and stay neutral among the four options. Among the options that pass, the one with the largest conviction-weighted Aye tally is selected. It then goes to DOT holders through the Polkadot-side resolution in the paper’s Appendix B.
Conflict 1: Kusama WFC #573
What #573 decided
#573 asked for an independent JAM upgrade sized for Kusama’s economy. Its key terms were 32 cores against Polkadot’s 341, a 1-second block time, and KSM as the native token for storage, staking and coretime. It estimated security costs at roughly 2.4% KSM inflation in the worst case. It passed with about 144,000 KSM Aye against almost no Nay. The alternative put to holders in the same round, #574, proposed a full-spec JAM instance designed for Polkadot to join and share security costs. It was rejected with about 240,000 KSM Nay.
Where the paper departs from it
| Topic | WFC #573 | Decision paper |
|---|---|---|
| Architecture | Kusama runs its own JAM instance | A Kusama-controlled service on Polkadot’s JAM Prime |
| Security | Kusama’s own validators, secured by KSM staking | JAM Prime’s validators, secured by DOT staking; Kusama’s validators, staking and staking issuance retired |
| Guaranteed capacity | 32 cores | 2 to 8 cores, plus separately purchased coretime where supported |
| Latency | 1-second blocks | Not addressed |
| Role of KSM | Native token for storage, staking and coretime | Governance and internal allocation; the base package is paid for in DOT; monetary redesign left to later KSM decisions |
| Independence | Full | Protected control within shared infrastructure |
On latency, the earlier position paper explicitly set #573’s 1-second block time aside. Northvane posted an estimate in the decision-paper thread, which the lead author called a fair analysis. It suggests that with fewer cores, the system chains would produce blocks less often than today. Under the 2-core option, that means roughly every 6 seconds for Asset Hub and every 24 seconds for People, Encointer, Bridge Hub and Coretime. If bundling several chains into one work package is supported at launch, the slowdown shrinks.
Why this conflict is in the open
Clause 1 of each resolution says that, if that option is selected, it replaces the independent, lightweight JAM path set by #573. Wishes for Change are signals, and a later WFC can replace an earlier one. In the forum, the authors have said that a departure from #573 should be put to an explicit vote with its reasons and consequences laid out. This round does exactly that.
Community members have raised three points worth weighing:
- The paper does not mention #574. Its design differed: Kusama would have hosted and Polkadot would have joined. But it shared the core idea of one instance serving both economies, and holders rejected it in the same round in which they approved #573.
- The case that an independent JAM is unaffordable is the authors’ judgement. Their September position paper acknowledged that its economic appendix did not contain a full budget for running an independent JAM, and did not quantify how the options compare on security. Forum participant flowfarm has asked why holders are choosing among JAM Prime allocations before seeing such a comparison.
- WFC #676 offers the independent path a line on the ballot. It was submitted for holders who want #573’s direction scoped and priced before validators are retired. It states that if it passes alongside one of the four, the comparison should come first. The paper’s selection rules do not address this.
The authors’ counterpoint is that an independent JAM leaves Kusama paying on its own for validators, bonded stake, implementation and upkeep. In their view, no design they have seen shows those costs to be specified, funded and affordable.
What this means in practice
An Aye on any of #671–#674 supports replacing #573’s direction, conditional on that option being selected. If none of the four passes, no new mandate is created and #573 remains Kusama’s latest expressed direction.
Conflict 2: Polkadot WFC #1926
What #1926 decided
#1926 sets linked rules for JAMKB, the proposed fixed-supply token that backs JAM state footprint:
- Burn rule. Any DOT that DOT DAO receives from JAMKB is to be burned in full by the protocol itself, without a discretionary decision along the way. This covers a sale, a lease, metered rental or the decayed part of a deposit.
- No giveaways. DOT DAO is not to give footprint away, whether as a grant, a gift, or a loan at zero or below-market interest.
- Market price. Every footprint allocation, whoever receives it and whatever it is for, must be “paid for at the prevailing market price, in DOT, with that DOT burned.”
#1926 deliberately leaves several things open: how footprint is released, how much and when, at what price, how much JAMKB DOT DAO keeps, and the settlement currency. A companion WFC requiring DOT-only settlement (#1927) was rejected. Settlement may therefore happen in another unit, but #1926 then requires the proceeds to be converted into DOT and burned. #1926 passed with about 129.48M DOT Aye against 6,600 DOT Nay. Like any WFC it is a signal rather than a protocol change, but it states that future technical implementation is bound by it.
Where the paper collides with it
| #1926 requires | The decision paper provides |
|---|---|
| Footprint paid for at market price | 450,000 JAMKB-equivalent bundled into a price calculated from coretime alone (40 CTYs × US$108,000); the footprint carries no price of its own |
| A price set by the market | A one-off DOT amount fixed from a 90-day DOT average, before JAMKB’s release mechanism and price have been defined, for footprint reserved for up to 20 years |
| DOT proceeds burned automatically, with no discretionary step | Appendix B asks DOT DAO to accept a single DOT amount “as payment for the complete package”, with no burn clause and no identifiable footprint share |
| No free or below-market footprint | Cannot be verified, because the formula never establishes what the footprint is worth |
The structure creates a dilemma. If the footprint is part of what Kusama is paying for, #1926 requires that share to be priced at market and burned. If it is not, it is a free allocation, which #1926 rules out.
The duration test
The clearest sign that the footprint is not being priced: the footprint quantity and the price are identical in all four options, while the reservation period runs from 5 to 20 years. The paper ties the footprint entitlement to the protected term, so it behaves like a lease. Under any market-priced lease, reserving 450,000 JAMKB-equivalent for 20 years costs more than reserving it for 5.
All four options buy the same 40 CTYs at the same rate. So any part of the US$4.32M attributed to footprint is the same in each. That makes the implied yearly footprint price under the 20-year option a quarter of the price under the 5-year option. At most one of them could match a market rate.
The edition history points the same way. The September position paper priced 80 CTYs plus the same 450,000 JAMKB-equivalent at US$8.64M. Halving the coretime halved the price while the footprint stayed constant. That paper itself noted that variants with equal CTYs but longer terms keep footprint available for longer and would need to be valued separately.
What the authors have said
The September position paper said the final agreement must identify “consideration and proceeds treatment for each component”, footprint included. It also observed that any DOT burned could not also fund operations or maintenance.
On 18 September, forum participant batbayar raised #1926 directly. He asked whether an allocation like this counts as a release of footprint, whose proceeds #1926 burns, or as retained usage governed by a separate policy. The authors replied with three points: the US$8.64M figure came from the compute benchmark; the formula puts no separate cost on footprint or the other obligations; and DOT DAO would “determine how to allocate and account for the one-off payment” across its delivery obligations. batbayar later described the classification question as still open.
The decision paper of 19 October does not carry these points forward. It does not mention #1926 and has no component-level pricing or proceeds clause. Its Polkadot-side template asks DOT DAO to accept one DOT amount for the whole package.
Why “DOT DAO decides later” does not resolve it
#1926 has already settled how DOT DAO treats footprint revenue: market price, DOT, burned automatically. Leaving the footprint unpriced, and the split of the payment to DOT DAO’s later discretion, is the kind of discretionary step #1926 was written to exclude.
To comply, DOT DAO would have to identify a footprint share, price it at market and burn it, leaving only the remainder for compute and maintenance. The position paper’s own observation about burned DOT shows why that could make DOT DAO less willing to accept the same total.
The Kusama resolutions also fix the package and the DOT amount, and send material changes back to the relevant governance bodies. A #1926-compliant restructuring would therefore probably need a new KSM decision.
