Arbitrum Replaces Timeboost With Per-Transaction Priority Auctions, Routing 97% Revenue to DAO

Arbitrum has overhauled its transaction ordering architecture, replacing the Timeboost mechanism with a per-transaction priority auction system that routes nearly all generated revenue back to its decentralized autonomous organization.

The new design splits into two components. Users who want their transactions processed ahead of others can bid in real-time priority auctions on a per-transaction basis. A Fast Feed subscription option gives participants accelerated access to transaction data for a recurring fee. Together, the two streams create a dual-revenue model for the network. According to Arbitrum's official announcement, 97% of proceeds from both priority fees and paid Fast Feed subscriptions flow directly to the DAO treasury, with the remaining 3% allocated to the Arbitrum Developer Guild.

That revenue split is the headline number for governance watchers. A 97% allocation to the DAO treasury is aggressive by any standard in the Layer 2 space, and it signals that Arbitrum's contributors are positioning the protocol as a self-sustaining public good rather than a fee-extraction vehicle for insiders. The Developer Guild's 3% cut is modest, and critics have already raised the question of whether that slice is enough to meaningfully incentivize protocol-level contributions over time. If developer retention becomes a problem, the split may face governance pressure to rebalance.

The shift away from Timeboost reflects a broader rethinking of how Layer 2 networks handle MEV, short for maximal extractable value, the profit validators and sequencers can capture by reordering, inserting, or censoring transactions. Timeboost introduced a time-advantage auction where winning bidders gained a brief ordering edge over a fixed window. Per-transaction auctions are more granular: each transaction competes individually, which in theory produces more accurate price discovery for ordering priority and reduces the blunt-force advantage that bulk bidders could accumulate under the older model. The practical effect on DeFi protocols built on Arbitrum, particularly automated market makers and liquidation bots that depend on predictable ordering, will take several weeks of live data to assess fully.

Retail traders who cannot or do not want to pay priority fees may find themselves consistently outpaced by sophisticated actors willing to bid for ordering advantages. Per-transaction auctions are efficient from a market-design perspective, but efficiency and accessibility do not always point in the same direction. Arbitrum's DAO will need to monitor whether the new system creates a two-tiered network experience that pushes smaller users toward cheaper alternatives.

Optimism and its OP Stack derivatives, along with zkSync and StarkNet, are all iterating on sequencer revenue models and MEV handling simultaneously. Arbitrum still holds a substantial lead in total value locked across the Layer 2 landscape, but that lead is not guaranteed to persist if transaction cost dynamics shift unfavorably for ordinary users. The 97% DAO treasury allocation gives Arbitrum governance a meaningful war chest to fund grants, ecosystem development, and potentially subsidies, but deploying that capital effectively will require governance discipline that DAOs have historically struggled to maintain at scale.

Per-transaction auctions are more transparent than time-window advantages, and routing the bulk of sequencer revenue to a community-controlled treasury is more defensible than alternatives where proceeds accrue to a centralized entity. Whether the Fast Feed subscription model attracts sustained demand, and whether the Developer Guild can do meaningful work on 3% of an uncertain revenue base, are the two open questions that will define whether this redesign delivers on its stated goals.