They say rollups are the big fix for Ethereum’s scaling problem. And to be fair, they do make things faster and cheaper. But once you look past the headlines, there’s a real concern that doesn’t get enough attention: these systems depend on sequencers, and sequencers themselves aren’t above trust issues.
Sequencers are the entities that order and publish transactions in a rollup. Ideally, they behave honestly, passing along data exactly as it happened. But if a sequencer cheats, say, by censoring transactions, front-running trades, or even manipulating ordering for profit, can the average user really prove it? And even if they can, what recourse do they have?
Optimistic rollups rely on fraud proofs, which let participants challenge bad behavior. But those challenges are technical, time-bound, and require users to be vigilant. ZK rollups use validity proofs, which are more mathematically secure, but the complexity means users still trust the underlying system of verifiers and provers. Either way, the idea that “anyone can prove fraud” is far more complicated in practice than it looks on paper.
That’s where decentralization of sequencers comes in. A single centralized sequencer is fast, but it’s also a single point of failure, and trust. The industry knows this, and there’s a big push toward shared or decentralized sequencers that reduce that choke point. Still, most rollups today are running with sequencers controlled by the teams that built them. In other words, the scaling solution we rely on is only as honest as the operator behind it.
This is the tension in crypto right now. We celebrate scalability milestones, but the trade-off is often hidden in governance and trust models. Users might never see fraud because the math is sound, or they might not notice because the systems designed to catch it are inaccessible to anyone without technical expertise.
For rollups to truly fulfill their promise, auditing mechanisms can’t just exist in theory, they need to be usable and transparent enough for everyday participants, not just insiders. Otherwise, we risk recreating the same dynamic we tried to escape from: trusting middlemen, but with extra cryptography layered on top.
