BNB Chain's Pasteur Fork Activates Tuesday With Bridge Fixes

BNB Smart Chain's Pasteur hard fork activates at 02:30 UTC on Tuesday. Any node still running a client older than v1.7.7 at that moment drops out of consensus. Pasteur carries three proposals, two of them about validator security and one about how blocks get built. The security pair closes a gap in the verification that cross-chain bridges depend on, and rewrites what a replaced validator key is still allowed to do. The third lets block builders hand validators work that has already been executed, which is where the upgrade's throughput claim comes from.

The bridge gap Pasteur closes

BEP-682 has the clearest failure mode behind it. Bridges on the network verify incoming state through a CometBFT light client, a compact routine that checks validator signatures without holding the whole chain. A manipulated validator set could list the same validator more than once, so an approval could clear the signature threshold with fewer real signers than that threshold implies. Crypto Briefing described the remedy without decoration: Pasteur "closes that gap by explicitly rejecting duplicate bridge validators." Nothing in the coverage says the flaw was exploited on mainnet, and no outlet this desk checked reports a loss tied to it.

What operators have to do before Tuesday

BEP-695 is the proposal validators can get wrong. After a key rotation the retired key loses its management rights. Pending penalties now follow the rotation instead of staying with the old key, and blacklisted addresses can no longer cast governance votes through signed messages. Crypto Briefing warned that validators who do not meet the new key rules risk penalties or removal from the active set. There is a configuration trap as well. Crypto Economy reported that the EnableBAL entry has to be deleted from config.toml or the node will not start, and that builders using the new block path must run full nodes rather than the lighter fast ones. The fork has been live on the Chapel testnet since 21 July, on client v1.7.6, which gives it about five weeks of public testing.

The throughput figure is a benchmark

BEP-675 is where the performance claim sits. Builders submit pre-executed blocks, cutting validator processing time from 125 ms to 15 ms, and a benchmark run recorded throughput climbing from 1,237 to 2,324 transactions per second with finality times unchanged. Where that run happened is not settled. Crypto Economy attributes the figures to a test network it calls QANet, while crypto.news and u.today present the same numbers as testnet results without naming the environment. They are benchmark results either way. Mainnet has not produced them, and no source publishes an activation block height for Tuesday, only the timestamp.

A security release in a security week

Pasteur reads differently from the two forks that preceded it this year. u.today framed January's Fermi and April's Osaka and Mendel releases as speed work, and Pasteur as an attempt to use capacity the chain already has while closing validator-security holes. That lands in a week when the network news has been about broken assumptions more than raw speed, from Harmony's rollback after a forged mint to an attacker who emptied Term Labs vaults by winning a vote. BNB traded near $696 on Monday, up about 1.8% on the day on one live price page, unmoved by the schedule. Tuesday's test is block capacity. Whether mainnet blocks after 02:30 UTC carry more transactions than the ones before them is what the benchmark has yet to prove.