The Sandbox DAO was the community-governance layer of The Sandbox, a voxel metaverse platform whose developer Pixowl was acquired by Animoca Brands in August 2018 for roughly US$4.9m. It ran from May 2024 to February 2026, funded 38 proposals with more than 3 million SAND, and then voted itself into dormancy. It is one of the cleanest public records of a corporate-sponsored DAO's full arc — launch, activity, decay, pause — because every vote is still readable on-chain.
Its value to anyone studying DAO governance is not the metaverse. It is the turnout curve: the DAO's first proposals drew over a thousand voters, and the proposal that shut it down drew 33.
From studio subsidiary to DAO
SAND launched in 2020 as the platform's ERC-20 currency, with a fixed 3 billion supply used for purchases, staking, and — eventually — governance. The DAO itself came later. The Sandbox described the plan as progressive decentralization: a staged handover rather than an immediate one, with Phase One putting SIPs to a community vote while staff kept running operations, Phase Two loosening that dependency, and Phase Three intended as full community control.
Voting opened on 28 May 2024 with a 25 million SAND treasury. The first batch was constitutional: SIP-1 introduced the DAO, SIP-2 established The Sandbox Foundation, and SIP-3 set its initial budget. It never reached Phase Three.
How the voting worked
Governance was off-chain and token-weighted, run in the sandboxdao.eth Snapshot space. The mechanics were unusually legible for a corporate-adjacent DAO:
- Eligibility: 5 SAND or a single LAND plot made you a member. Voting power summed SAND and LAND holdings across Ethereum and Polygon, with 1 LAND converted to 4,500 SAND of weight — a deliberate tilt toward landowners over token holders.
- Cadence: a new batch of SIPs went live every other Wednesday, each open for a 14-day voting period.
- Quorum: 30,000,000 voting power, with weighted voting across Yes / No / Abstain.
- Delegation: holders could delegate voting power without moving tokens, the standard Snapshot pattern (see delegated voting).
- Council review: a Council including co-founder Sebastien Borget and Animoca chairman Yat Siu reviewed every SIP and could veto proposals on grounds of legality, misalignment, or redundancy.
That last item matters more than it reads. A veto held by the sponsoring company's founders is a hard ceiling on how decentralized the DAO could become, whatever the phase roadmap promised — the same structural tension covered under progressive decentralization and security councils.
What it actually funded
SIPs fell into categories the DAO framed as Game Content, Sandbox for Good, NFT, and Platform. In practice the spending mix was grants, events, and marketing rather than protocol parameters:
- Causes: SIP-4 funded carbon capture with Net Zero Company; SIP-5 supported the Red Cross.
- Platform features: SIP-6 (Magic Palette design tool), SIP-16 (in-game inventory filter), SIP-21 (window sizing options for the game client).
- Programs: SIP-11 funded the Ambassador Program, SIP-25 established the Sandbox DAO Grants Program, and SIP-27 elected Domain Allocators to steer it.
- Cross-DAO: SIP-20 set up a collaboration framework whose first phase was participating in ApeCoin DAO governance — one metaverse DAO delegating attention to another, months before both wound down.
By its own accounting in SIP-39, the DAO distributed more than 3 million SAND across 38 SIPs, drew members from more than 20 countries, and generated roughly 8,000 forum posts across 500 threads.
The turnout collapse
Every SIP is public on Snapshot, which makes participation decay measurable rather than anecdotal. Voter counts, first proposal to last:
| Proposal | Date | Voters |
|---|---|---|
| SIP-3 — Initial Budget for the DAO | May 2024 | 1,161 |
| SIP-11 — Ambassador Program | Aug 2024 | 830 |
| SIP-19 — UGC Platform Development Team | Jan 2025 | 251 |
| SIP-27 — Domain Allocator Election | Apr 2025 | 187 |
| SIP-35 — Magic Realms | Aug 2025 | 91 |
| SIP-38 — Art Commission Grant | Dec 2025 | 31 |
| SIP-39 — Pausing The Sandbox DAO | Feb 2026 | 33 |
Participation fell roughly 97% over twenty months while the Snapshot space kept accumulating followers — over 14,000 by the end. Voting power did not fall with it: SIP-3 cleared with about 40 million voting power and SIP-39 with about 70 million. Fewer people were voting, and the ones who stayed carried more weight each.
The decision to wind the DAO down was therefore made by 33 addresses. That is not a scandal — it is the ordinary end state of token-weighted, low-stakes governance, and the reason voter apathy is a structural risk rather than a community-management problem. A quorum denominated in tokens rather than voters will keep clearing long after the electorate has left.
SIP-39: pausing the DAO
SIP-39, "Pausing The Sandbox DAO", opened on 11 February 2026 and closed on 25 February 2026. It passed with roughly 56.1 million voting power for, 6.3 million against, and 7.6 million abstaining, from 33 voters.
The proposal was candid about why. The DAO's mandate had drifted from where the platform was going with "The Sandbox 3.0", and — the more concrete problem — operating costs were invoiced in USD and EUR while the treasury was denominated in SAND, so a falling SAND price meant selling steadily more of the treasury to cover the same fixed bills. This is the treasury-denomination mismatch that has caught many DAOs holding a single volatile native asset against fiat liabilities.
What "paused" means in practice: no new funding or grant SIPs are accepted, curated, or put to a vote; no new governance initiatives, elections, or mandates; existing SIPs with signed contracts run to completion or settle under their notice terms while uncontracted ones are closed out; and responsibility for finance, assets, and contracts transfers to The Sandbox and The Sandbox Foundation. The transition ran to 31 March 2026, with dormant mode from 1 April 2026. In the forum discussion, the team put the remaining treasury at about US$4.4m USDC of the roughly US$6.8m originally received, with a maintenance budget of US$44,400 to preserve assets during the pause.
Contributors in that thread were not persuaded that a pause is reversible. One drew the comparison directly to ApeCoin DAO, which had been dissolved into a Yuga Labs company eight months earlier; another described the pause as "a way to quietly make the DAO fade away". Reactivation is possible in principle — it would take a SIP, and SIPs are exactly what the pause stops.
What DAO builders can take from it
Three things generalize beyond the metaverse sector:
- A grants DAO next to a company is a budget line, not a polity. The Sandbox DAO never governed the product — it allocated a company-funded pot toward events, causes, and feature requests. When the sponsor's strategy changed, there was no independent mandate to defend, which is why the pause passed almost unopposed.
- Token-denominated quorum hides voter flight. Because quorum was 30 million voting power rather than a headcount or a share of holders, the DAO kept clearing its bar while 97% of its voters left. Any DAO that wants participation to be a real constraint has to measure it as one — see how DAOs fail.
- A treasury in your own token is a short position on your own fortunes. Fixed obligations in fiat against a reserve in SAND meant the DAO's spending power fell exactly when its ecosystem needed it most. Compare Decentraland DAO, which faces the same structural exposure through MANA, and the diversification arguments under treasury management.
How Caper approaches this
The uncomfortable part of a pause is what it does to the people who stayed. Once no new SIPs can be raised, a member's only remaining lever is to sell into whatever market exists for the token — the governance surface simply closes, and nothing about the treasury is claimable by them.
Caper's design does not depend on the DAO being open for business. A member calls exit directly: it pays out a share of the treasury computed as their canonical vote weight — (governance tokens held × vote tokens held) / (total vote supply × circulating supply) — burns their vote tokens, and sells their governance tokens back into the bonding curve for XRD. No proposal, no quorum, no council review, and no dependency on anyone else still voting. A caper that goes quiet is one its members can still leave with something, which is the point of an exit right as distinct from a market. See what is a caper and the bonding curve.