Ethereum Public Goods Funding

Ethereum core development is underfunded relative to market cap and the complexity of what’s being built. Multiple proposals were presented at ETHDenver 2026 to establish sustainable, protocol-aligned funding: from staking-reward redistribution to AI-assisted prediction markets for allocation. (→ [[ethdenver]])

The Protocol Guild Crisis (Peter Vecchiarelli)

Current state:

184 core developers across 25 teams, 13+ organizations maintain Ethereum
Median cash salary: ~$140,000 — significantly below competing offers
Competing offers: 200%+ cash plus 1%+ tokens (VC-backed projects can offer much more)
Zero equity/token upside outside Protocol Guild distributions

Protocol Guild cumulative distributions:

$37M cumulative distributed
$12M in 2025 alone
~$60,000 per median member (provides ~33% of total comp)

Funding gap:

Tier	Annual Need
Minimum viable	$30M/year (bridge cash gap)
Competitive	$60M/year
Ideal	$100–200M/year

Benchmark: Palantir and Netflix each spend more on R&D than the entire Ethereum core dev ecosystem receives in total compensation.

Infinite endowment proposal: Route yield from DeFi vaults (Morpho, Lido) to core dev. Creates self-sustaining funding engine. Comparable to university endowments funding research without relying on tuition fees.

Validator Reward Redistribution (Devansh Mehta / EF)

The prisoner’s dilemma: Validators collectively benefit from funding Ethereum growth (more demand = higher block space value = higher ETH price), but individually have no incentive to voluntarily contribute.

Proposed mechanism:

51% threshold flag: validators vote (like gas limit voting) on whether to redirect a portion of rewards to a public goods fund
Solves free-rider: if 51% agree, everyone participates equally — no unilateral sacrifice
Condorcet voting over validator preferences: splitter contract distributes to weighted portfolio of projects based on preference aggregation

Risks:

Staker cartelization: 51% coalition could collude to steal from the fund
Principal-agent problem: staking operators vote, but underlying ETH holders bear the cost
Formalizing overpayment: validators already overpaid relative to security budget; adding another deduction may be difficult politically

Philosophical question: Should the Ethereum protocol serve as coordination venue for collective action (public goods funding), or remain a pure incentive-alignment mechanism?

AI-Assisted Prediction Markets for Allocation (Devansh Mehta / EF)

Problem with traditional grant committees: monopoly on judgment; slow; opaque; subject to capture.

Proposed three-track system:

Human jury evaluation: evaluates specific projects (slow but authoritative)
Prediction markets: participants bet on what the jury will decide; market prices reveal community disagreement
AI competitions: models predict jury judgment across the entire dependency graph

Ethereum dependency graph: 10,000+ edges (packages, libraries, protocols depending on each other). Vitalik spot-checked 30 paths; AI models predict weight of every edge against jury consensus.

How it works in practice:

Solidity market: 11% market price vs. jury average → market can force jury to re-evaluate, improving transparency
AI model that best predicts jury consensus gets used for distribution weights
Humans make few decisions (jury evaluations); AI extrapolates to thousands of allocations

Technical needs: LMSR AMM (liquidity-independent pricing), agent invocation without CSV re-upload, Uniswap-style eligibility filtering

Markets for Public Goods vs. Grants

The broader category of “retroactive public goods funding” (RPGF), quadratic funding, and futarchy-based allocation:

Optimism RPGF: retroactive funding for measurable impact; works for things that already exist; doesn’t fund early-stage risky research
Gitcoin quadratic funding: community-weighted, Sybil-resistant; works for visible community projects; gaming risk
Protocol Guild: direct distribution to identifiable contributors; low overhead; hard to expand beyond core dev
DAO Security Fund approach (see TheDAO Security Fund): curated rounds with domain experts; bottom-up via operators; quarterly cadence

What’s missing: A mechanism that funds early-stage risky protocol research that isn’t yet measurable, isn’t community-visible, and may take 5+ years to pay off. This is the gap the Ethereum Foundation’s strategic grants currently fill — but EF itself is underfunded relative to the complexity of the roadmap.

The Protocol Guild Endowment Proposal

Proposed structure:

Protocol Guild holds a DeFi vault (e.g., Morpho pool) seeded by donations from DATs, protocols, and individuals
Yield from vault flows to core devs automatically without requiring recurring fundraising
Initial capitalization target: $1–2B (yielding $30–60M/year at 3–5% yield)
Bootstrapping path: large ETH holders (DATs, EF) seed the vault; yield is self-sustaining thereafter

Comparison to University endowments: MIT’s endowment funds ~30% of operating budget; a Protocol Guild endowment at comparable scale would fully fund competitive compensation.

ETHPrague 2026: Kernel framework, dependency funding, and EIP-shaped reallocation

ETHPrague 2026’s Funding Coordination team (Martin Hansen, Sophia Sukhinina, Devansh Mehta) consolidated three coordinated proposals. Full coverage at Ethereum Kernel & Public Goods Funding.

The Ethereum Kernel ($18M/yr floor, Martin Hansen)

The minimum-viable-funding framework for the protocol’s critical functions. Components: clients (CL + EL), coordination, research (short- and long-term), security/specs/testing, language stack, core wallet infrastructure, other critical tooling and dependencies. Floor of $18M/year in neutral funding to keep the kernel operational without strings attached. Above the floor → faster development; below → abrupt-shutdown risk and strategic-capture exposure. Protocol Guild’s $30M/year minimum-viable target maps directly onto this kernel framework.

Dependency-Aware Funding (Sophia Sukhinina)

The structural argument: open-source funding currently reaches ~3,000 maintainers worldwide at ~$3,200/year per maintainer (after orgs, contractors, ops, legal, audit overhead). The EU Cyber Resilience Act — partial September 2026, full December 2027 — makes Software Bill of Materials mandatory for every commercial digital product. Funders will see the full dependency graph; ignoring deep dependencies becomes legally indefensible.

Demo at ETHPrague: 10 Schleswig-Holstein-funded projects have 4,000 dependencies across 6 continents and 12 currencies. SEPA/SWIFT cannot route those funds; Ethereum + stablecoin rails are the natural infrastructure. The UNICEF Drips Network pilot is the prototype: smart-contract-auto-split distribution with dependency-aware weighting.

This is the meta funding mechanism — funds Ethereum tools indirectly because they sit in dependency graphs of EU-funded products.

Sourcing Funding for Ethereum Growth (Devansh Mehta)

Three layered EIP-shaped proposals, focused on sourcing (not allocation):

Consensus layer — Validator yield reallocation EIP:

Validators flag a percentage of yield for ecosystem reinvestment (similar to gas-limit flag mechanism).
Nash equilibrium argument: validators benefit from ecosystem value increase → commit non-zero amount.
Condorcet aggregation of validator preferences picks the best-fitting allocation contract; anyone can propose a replacement.
Solves the prisoner’s-dilemma problem (taxation is the real-world analog) — 51% vote → 100% participate.
Caveat: principal-agent problem between stakers and ETH-holders is unresolved.

Smart-contract layer — Dependency prediction markets:

Each smart contract publishes its dependency graph (CRA-style SBOM).
Live prediction market on which dependencies are most critical; traders predict expert-assigned weights.
Manipulation defense: attempts to push a weight too high get shorted by other traders chasing the profit opportunity.
Live deep.co.pm market with $20K in prizes during ETHPrague window.

Application layer:

Frontends and aggregators voluntarily route a percentage of fees to public-goods funds.
DAOs commit a default percentage of treasury yield; turning off requires explicit vote, not turning on.
Users vote with their feet: aggregators that don’t route fees lose to those that do.

The cumulative vision: $100 swap → application routes some → smart-contract layer routes more to dependencies → consensus layer validators reallocate yield.

Diversifying Funding via FRCs (Raul Romanutti)

Funding Round Contracts — standardize the funding round itself as an on-chain primitive. Each FRC specifies eligibility, voting/matching, allocation, and reporting. Multiple FRCs run in parallel for different ecosystem segments (security, language stack, research, education). QF round becomes QF protocol rather than QF event.

Funding the ETHPrague tooling cluster

Several of the Smart Contract Fuzzing and AI Smart Contract Auditing tools (Wake, Akidna/Echidna, Greg AI’s open-source components, x402 SDK) are funded primarily through ETH Security Quadratic Funding rounds. Grieco’s blunt admission: tools without revenue depend on QF rounds; a single round shrinking would silently kill production-grade audit infrastructure. The kernel framework explicitly names “critical tooling and dependencies” to address this fragility.

Connections

TheDAO Security Fund — Parallel effort funding security specifically; shares the bottom-up, community-governed design philosophy
Ethereum Staking Dynamics — Validator reward redistribution and issuance rate are coupled; changing one affects the other
DeFi Institutional Transition — DATs and large DeFi protocols are the natural donors to a Protocol Guild endowment; their interests are aligned with Ethereum’s long-term health
Digital Asset Treasuries (DATs) — Sharplink’s “for-profit but benevolent capitalist” framing includes supporting core development; Bitmine at 4% of ETH also flagged as potential contributor

Open Questions

Can the 51% validator flag mechanism resist gaming (e.g., staking pools capturing the vote for their own benefit)?
Is $1–2B endowment achievable from voluntary donations, or does it require protocol-enforced issuance redirection?
How does AI-assisted prediction market allocation handle cases where jury and market systematically diverge on underserved areas (e.g., P2P networking, which is unglamorous but critical)?
Should Protocol Guild expand beyond core client teams to infrastructure tooling (testing frameworks, auditing, documentation)?