DuneDigest036

This week, Uniswap Labs and the Uniswap Foundation introduced the UNIfication proposal, which seeks to finally activate the long-dormant protocol fee switch and channel those fees into a programmatic UNI burn via a new onchain “TokenJar → Firepit” mechanism. Rather than distributing revenue to token holders, the proposal routes fees from Uniswap v2, v3, Unichain sequencer revenue, future v4 aggregator hooks, and the upcoming MEV-internalizing Protocol Fee Discount Auctions (PFDA) into UNI buy-and-burn cycles as the core value-accrual mechanism. It also includes a retroactive burn of 100M UNI from the treasury, an estimate of what would have been burned had fees been active since launch.If approved, v2 fees would move from 0.3% for LPs to a 0.25% / 0.05% split, and v3 pools would send a small cut of LP fees to the protocol—25% on low-fee tiers, 17% on higher-fee tiers. The proposal also restructures the ecosystem by merging most Uniswap Foundation teams into Labs, committing Labs to eliminate interface, wallet, and API fees, and creating a 20M UNI annual growth budget to align Labs’ operations with protocol-wide development and usage. This shift transforms UNI from a “governance-only” token into a deflationary asset directly tied to protocol activity. The timing is notable: October saw one of the largest LP fee spikes in Uniswap’s history, surpassing $100M in weekly fees, underscoring how much potential value the protocol could capture once fees are activated. This proposal has rekindled debate over LP incentives, potential liquidity erosion, competitive pressure from fee-generating rivals like Aerodrome, and the broader question of how value capture should work in a maturing DEX landscape.

Ripples from the Stream Finance debacle continue to hit DeFi. On Nov 12, MEV Capital issued an update on its Ethereum Morpho USDC vault, removing the Elixir sdeUSD/USDC market after deUSD’s collapse and crystallizing a 3.6% bad-debt loss. The vault now has no ties to Elixir assets, with all other parameters unchanged. Dune data shows MEV Capital’s total TVL on Morpho dropping from $540M on Nov 4 to $125M on Nov 14, with the Morpho USDC cluster collapsing from $314M to $34M, and the USDC Arbitrum cluster from $48M to $3M over the same period. The episode highlights the potential vulnerabilities of curator-driven vaults, where managers deploy user funds into higher-yield, higher-risk strategies without proportional downside, leaving lenders to absorb losses when over-leveraged positions unwind or oracle failures hit. Supporters of Morpho’s design counter that the system worked as intended: losses were contained to a single vault, preventing protocol-wide contagion. They contrast this with monolithic pools like Aave, which effectively operate as a “single big vault”, centrally curated assets and parameters mean a bad oracle, collateral failure, or governance error can expose the entire pool to systemic risk. Backers also argue that modular, vault-based architectures are necessary for scaling lending to institutional and eventually multi-trillion-dollar levels, enabling permissioned or KYC-gated vaults and bespoke risk profiles. What everyone agrees on is the urgent need for clearer user understanding of vault risks and the limits of so-called “safe” yield.

Pump.fun's Mayhem Mode, introduced on November 12, 2025, is an experimental feature that doubles a new token's supply, allocates half to an AI trading agent for 24 hours of randomized buys/sells to generate volume and liquidity, then burns the bot's remaining holdings. The goal, according to Pump.fun, is to help small or low-attention projects break out, attract traders earlier, and increase the odds of reaching higher market caps, potentially accelerating paths to DEX listings like Jupiter. Any profits or losses the agent generates flow into an insurance fund, with protocol fees from Mayhem coins redirected there as well. Pump.fun stresses the agent isn’t designed to profit and is running as a transparent, permissionless beta. Still, backlash has been strong. Dune data shows over 5,800 Mayhem-mode coins launched in the first week (2,000+ in the last 24 hours alone), the bot has already executed over 130,000 trades with a net loss of ~$76K, and critics warn it can become a net seller, introduce extra supply, drain bonding-curve liquidity, and leave some holders unable to exit before a coin bins to PumpSwap, effectively socializing losses onto retail traders. More broadly, Mayhem Mode accelerates the shift toward AI-driven launchpads, introducing new volatility, new failure modes, and fresh ethical debates around artificial activity, platform incentives, and trader protection in a gambling-heavy market. Whether it becomes a genuine boost for struggling early-stage projects or a reputational liability for Pump.fun will depend on how these dynamics play out as more data comes in.

Talking about AI Agents, Virtuals Protocol’s Agent Commerce Protocol (ACP) v2, launched on October 21, 2025, upgrades its onchain framework so AI agents can discover jobs, negotiate terms, execute work, and settle payments using ERC-4337 wallets, ERC-6551 identities, and x402 for faster, cheaper transactions. In the ACP design, memos are signed onchain messages that act as the core communication primitive, powering standardized interactions across the request, negotiation, transaction, and evaluation workflow. AACP v1 was already gaining traction by mid-October, surpassing 22k daily memos on Oct 18, and ACP v2 has rapidly taken over, hitting an ATH of ~50k memos on Nov 13 and pushing the total to 934,440+ onchain memos as agents coordinate everything from research tasks to trading automation. This activity underpins a fast-growing aGDP (agentic GDP), suggesting real economic weight behind agent networks rather than toy demos. The implications are big: ACP v2 links DeFi rails with AI logic, enabling agent clusters (trading DAOs, media shops, research collectives) to build verifiable reputations and programmable workflows, while also exposing new challenges around cold-start liquidity, noisy/public transaction trails, and a potential shift in launchpads toward utility-focused AI ecosystems over pure memecoins.

LI.FI, the cross-chain bridge and DEX aggregation layer, continues to accelerate. In October, the protocol surpassed $50B in lifetime volume, adding $8.2B month-over-month, and processed 8M additional transactions, bringing total activity to 80M+ transfers across 60+ chains. Growth was fueled by integrations with a wave of new partners, including Momentum, Kamino, Flow, and Hemi. LI.FI also expanded stablecoin routing with new DEX support on Plasma (which saw a +148.7% month-on-month TX increase) and added mint/burn flows for USD* and Eco. This momentum reinforces LI.FI’s position as a key multichain coordination layer in an increasingly fragmented DeFi landscape, driven by 650+ partner integrations and routing engines like HyperCore powering hyperliquid environments. At the same time, competition from competitors such as Across and Socket continues to spark debate over whether aggregator models can sustain long-term growth without deeper liquidity optimization and more defensible revenue streams as the market matures.

Nothing in this newsletter constitutes financial advice.
Always do your own research.

Dune Digest is all about cutting through the noise and surfacing the most relevant onchain trends. If you have insights, dashboards, or data-driven stories that belong in the Digest, let us know!