Buterin’s AI Agents: A New Voting Era
Ethereum cofounder Vitalik Buterin is pushing for a radical shift in how decentralized autonomous organizations (DAOs) operate.
Buterin’s plan also recommends the use of privacy-preserving tools such as multi-party computation (MPC) or trusted execution environments (TEEs). These would allow AI agents to process sensitive data without exposing it on the public blockchain, while zero-knowledge proofs (ZKPs) could enable users to prove voting eligibility without revealing wallet addresses or vote choices. The proposal even suggests launching prediction markets where agents can wager on the likelihood of proposal acceptance, potentially filtering out spam or low-quality submissions.
Wall Street Piles Into Bitmine, Despite Slump
While Ethereum developers debate governance upgrades, traditional finance giants are making bold moves in crypto equities. In Q4 2025, Morgan Stanley increased its position in Bitmine Immersion Technologies (BMNR) by about 26%, reaching over 12.1 million shares valued at $331 million. ARK Investment Management followed suit with a 27% increase, now holding 9.4 million shares worth $256 million.
Other major institutions have gone even further: BlackRock boosted its BMNR holdings by 166%, Goldman Sachs by an eye-catching 588%, Vanguard by 66%, and Bank of America by a staggering 1,668% during the same quarter. Yet, on paper these bets look bullish, but BMNR’s share price actually dropped about 48% in Q4 2025 and nearly 60% over six months, trading near $19.90 in premarket action Thursday.
Every one of the top 11 largest shareholders—including Charles Schwab, Van Eck, Royal Bank of Canada, Citigroup, and Bank of New York Mellon—upped their exposure during this period.
Why It Matters: Practical Impact for Users
For everyday Ethereum users and DAO participants, Buterin’s AI steward proposal could reduce the friction of active involvement—AI models trained on personal values may vote more consistently than busy humans. The privacy features proposed—such as ZKPs and secure computation—could also make voting less risky by hiding both user identity and vote content from public view.
However, the practical impact remains uncertain until such systems are tested at scale. For instance, the suggestion to use prediction markets as a spam filter is novel but unproven; it’s unclear whether this will genuinely improve proposal quality or simply add complexity. On the investment side, Wall Street’s aggressive accumulation of BMNR shares—even as prices fell nearly 50% last quarter—signals institutional confidence in crypto infrastructure despite short-term volatility.
See Also
Quantum Threats and Ethereum’s Defensive Moves
Security concerns are not limited to governance. At ETH Denver 2024, experts discussed Bitcoin’s vulnerability to quantum computing attacks. Bitcoin relies on elliptic curve cryptography for digital signatures—a system that could be compromised if quantum computers running Shor’s algorithm become powerful enough. In December 2024, Google revealed Willow, a quantum computer demonstrating below-threshold error correction capabilities.
Ethereum is not standing still: the Ethereum Foundation has assembled a post-quantum security team to study potential risks posed by quantum advances. Coinbase has also convened an advisory board focused on these threats to digital assets. SHA-256—the hashing algorithm used in Bitcoin—is considered difficult for quantum computers to break using Grover’s algorithm alone; however, digital signatures remain a weak point across blockchains.
Prediction Markets Eyed for DAO Quality Control
One intriguing aspect of Buterin’s proposal is the idea of using prediction markets to filter DAO proposals before they reach a vote. Under this system, AI agents could bet on whether proposals will be accepted or rejected by the broader community—a mechanism designed to weed out spammy or low-quality initiatives before they consume collective attention.
As reported by coindesk.com, these mechanisms would rely heavily on privacy-preserving computation environments like MPC and TEEs to ensure sensitive information remains confidential while still allowing agents to make informed assessments.
The Gist
- •Vitalik Buterin proposed using personal AI agents for DAO voting, with privacy ensured by MPC, TEEs, and zero-knowledge proofs.
- •The proposal suggests prediction markets for filtering DAO proposals, allowing agents to bet on acceptance likelihood to reduce spam.
- •Buterin’s plan was published on X one month after he criticized DAOs for low participation and centralization.
What may drive the next phase
If Ethereum developers proceed as scheduled, the Hegota upgrade—featuring Fork-Choice Enforced Inclusion Lists (EIP-7805)—is expected to roll out in the second half of 2026; immediate implications would include enforced transaction inclusion by validators, though the precise impact on network operations remains unclear due to ongoing criticism from figures like Ameen Soleimani.