Two separate developments are drawing attention across crypto governance and network infrastructure this week. Arbitrum’s Watchdog Committee has given three DeFi projects a tentative Sept. 10 deadline to answer findings of grant misuse or face votes on permanent exclusion from future Arbitrum DAO programs. The projects named are Good Entry, Limitless and APX Finance, formerly ApolloX.
The watchdog cited combined figures totaling 457,553 ARB, though it stressed the sum combines different findings rather than one recoverable or stolen amount. For Good Entry, on-chain analysis found 142,839 ARB distributed to 1,032 ineligible users during and after the Short-Term Incentives Program, along with alleged self-farming by wallets connected to team addresses. Limitless is accused of swapping 75,000 ARB into USDC and moving the funds to Base, with team members unreachable. APX Finance was linked to 239,714 ARB tied to overlapping issues, including unutilized treasury funds, late transfers to distributor contracts and alleged team-linked Sybil activity.
As of Sept. 5, none of the projects had responded in the proposal thread. The proposed bans would be decided through off-chain Snapshot votes and would not freeze wallets or disable protocols. They would instead make covered projects or individuals ineligible for future DAO programs. The watchdog said that as of Sept. 2 the broader program had received 90 reports, recovered about 532,000 ARB and distributed about 268,000 ARB in reporter bounties.
In a separate report published Sept. 1, ARK Invest and Glassnode assessed blockchain capture risk by measuring the smallest group of block-production entities needed to cross a protocol-relevant control threshold. They put that threshold at three entities for Bitcoin and Ethereum and 19 for Solana. The same framework placed Bitcoin first in its composite decentralization ranking, even though Solana had a higher coordination count for that particular metric.
A seven-day Bitcoin mining snapshot on Sept. 6 attributed 26.88% of blocks to Foundry USA, 16.91% to AntPool and 15.25% to F2Pool, meaning the three pools coordinated block templates for 59.04% of observed production. For Ethereum, Rated Network listed Lido at 21.17%, SSV at 16.56% and Binance at 7.77%, while noting Lido itself can represent many underlying entities.
Solana’s Nakamoto coefficient was reported at 19 by ARK and Glassnode, 18 by Solana Compass on Sept. 6, and 20 in the Solana Foundation’s June 2025 health report. The foundation links that threshold to the ability to censor blocks or stop consensus. The report also highlights infrastructure and software risks: Solana had about 92% of stake using Agave/Jito clients in April 2025, while Ethereum’s execution client Geth accounted for 50.17% of measured nodes. Bitcoin’s node visibility and pool coordination remain distinct from miner ownership, since miners can redirect hash rate quickly.
The findings suggest institutions should assess transaction censorship, infrastructure outages, software faults and exit speed separately rather than relying on a single decentralization score.
