Executive Statement

This is the short version of what the on-chain data implies Livepeer should optimize for.

What we believe is true (from evidence)

New delegator inflows are collapsing overall on Arbitrum (first-time bonders drop year-over-year).
Small delegators are plentiful by count but negligible by stake, so “more small accounts” is not a meaningful security lever by itself.
Security-relevant growth is bracket growth: attracting and retaining more independent participants in the 1k–10k and 10k+ bands (and improving delegate/orchestrator decentralization).
Therefore, the core strategy is LPT utility + liquidity + UX, not pure “reward boosts” that can be sybil’d.
A meaningful share of LPT inflation rewards appears to be withdrawn and routed off-chain/cross-chain, so inflation-heavy programs can create structural sell pressure unless retention-gated or paired with real fee growth.

From the latest snapshot (/research/delegation-board):

Net change (example window 2024-10-end → latest):

10k+: +37 wallets, +~5.99M LPT
1k–10k: +15 wallets, +~120k LPT
Mid-retail bands (10–100, 100–1k) lost wallets and stake in the same window.

If the goal is bonded stake + decentralization, then:

prioritize mechanisms that increase token utility (so 1k–10k and 10k+ can enter/exit without huge friction),
pair any incentives with retention hooks (time/vesting/forfeit),
avoid “pay more per small address” designs unless there’s a credible uniqueness primitive (or a real cost to splitting).

If the goal is also “long-term investor confidence”, then:

measure and publish rewards claimed vs rewards withdrawn (upper bound + conservative proxy),
treat “delta-neutral yield extraction” as a plausible market behavior and design around it (see /research/reflexivity-and-yield-extraction).