Snapshot is the backbone of modern DAO governance, enabling gas-free voting for projects like Uniswap, Aave, and Lido. But behind the scenes, a small group of entities control the oracle that counts your vote. This is the story of how we got here—and why it matters.
The Snapshot Phenomenon
In 2021, Snapshot launched as a simple solution: off-chain voting with gas-free participation for token holders. Today, over 14,000 DAOs use the platform, processing millions of votes. The magic is in the architecture—your vote is signed with your wallet, submitted to IPFS, and counted by an 'oracle' that reads the subgraph.
How the Oracle System Works
When you vote on Snapshot, you're not just submitting a signature. The oracle—a backend service—fetches your vote from IPFS, verifies it against the subgraph data, and aggregates results. This happens in milliseconds, but the oracle's role is anything but trivial.
Who Controls the Oracle?
The answer is... Snapshot itself. The snapshot.org domain is controlled by a multisig of core contributors. The oracle endpoints (like hub.snapshot.org) run on servers managed by the same group. While technically 'decentralized' in that anyone could fork the system, the network effect keeps most users on the official infrastructure.
The Decentralization Dilemma
DAOs love Snapshot because it removes the barrier to participation. No gas fees, no wallet setup beyond signing a message, and results appear instantly. But this convenience comes with a trade-off: centralized vote counting.
The Censorship Concern
In a governance system, the oracle is the arbiter of truth. If Snapshot's oracle censors a vote, that vote simply doesn't exist in the official tally. The 2023 Lido proposal controversy highlighted this risk when certain votes appeared to be missing from preliminary tallies.
The Single Point of Failure
A government subpoena, a server outage, or a compromise of the multisig could theoretically disrupt governance for thousands of DAOs. This isn't theoretical—Tally's acquisition of Snapshot in 2023 raised questions about commercial control over public infrastructure.
Attempts at Decentralization
The DAO community has tried to solve this. Projects like Colony and Aragon have built their own vote counting mechanisms. More interestingly, some DAOs have implemented 'vote delegation verification'—using multiple oracles and requiring consensus.
The Chainlink Model
Chainlink's decentralized oracle network could, in theory, count votes. But the gas costs and latency make it impractical for mass participation. The tradeoff between decentralization and accessibility remains unsolved.
What DAOs Can Do Today
If you're a DAO participant, here are practical steps to mitigate oracle risk:
Demand transparency: Require DAOs to publish their vote tallying methodology and data sources
Build redundancy: Run your own subgraph and verify results independently
Consider hybrid models: Use on-chain voting for high-stakes proposals, Snapshot for routine matters
Vote with conviction: If your DAO's governance model is truly decentralized, support infrastructure that reflects those values
The best DAOs don't just optimize for participation—they optimize for trust minimization. Snapshot solved participation. Now we need to solve trust.
The Road Ahead
The tension between accessibility and decentralization isn't going away. As DAOs mature, we'll likely see more sophisticated approaches: verifiable vote counting, multi-oracle consensus, and perhaps even on-chain vote aggregation for critical decisions. The goal isn't perfection—it's progress toward governance that truly represents its stakeholders.
Bottom Line
Snapshot's oracle system is the price we pay for mass democratic participation in DAOs. It's not inherently bad—it's a pragmatic solution to a hard problem. But it's not decentralized either, and pretending otherwise helps no one. The healthiest DAOs acknowledge this tradeoff and work to mitigate it, rather than ignore it.
