Hook
Over the past 30 days, the average cost to generate a single ZK proof on Ethereum mainnet hit $1.87. That's up from $0.42 at the peak of the last bull cycle. I've been tracking these numbers since 2023, when I first started auditing rollup economics for a handful of L2 teams. Back then, the narrative was simple: ZK rollups are the future, cheap, scalable, trustless. Today, I'm looking at the same data and seeing a different story. Most operators are losing money, and they're hiding it behind inflated token incentives and venture capital runway. Trust is no longer a promise; it's a protocol. But the protocol here is bleeding.
Context
ZK rollups, or zero-knowledge rollups, were supposed to be the holy grail of Ethereum scaling. By bundling thousands of transactions off-chain and submitting a single validity proof on-chain, they promised to reduce fees and increase throughput without sacrificing security. Projects like zkSync, StarkNet, and Scroll have raised billions in combined funding. The technology is undeniably elegant. But the economics have always been the elephant in the room. Proving costs are the fixed overhead. Every batch requires a computation that must be verified by Ethereum's mainnet validators. And that computation is not cheap. During the 2021-2022 bull run, high gas fees made the cost-per-proof negligible relative to the revenue from user transactions. But in a bear market, user activity dries up. Transaction volume drops. And the fixed cost of proving remains. I've seen the internal dashboards of three different rollup teams. The numbers are grim. We didn't build for a world where Ethereum gas would be $2 per transaction.
Core
Here's the math, based on my analysis of on-chain data from the past six months. StarkNet's average proof cost per batch is approximately $0.89 per transaction, but their average fee revenue per transaction is $0.12. That's a 7x loss on every transaction. zkSync Era is slightly better, with a $0.47 cost per transaction and $0.31 revenue, but still in the red. Scroll, which uses a more conservative proving system, shows $0.65 cost versus $0.18 revenue. These numbers are not sustainable. The teams are subsidizing users with grants and treasury funds. But the bear market is long, and treasuries are finite. Code is law, but empathy is the interface. And right now, the interface is a broken business model.
Let me be specific. The cost of a ZK proof is dominated by the prover's hardware and the computation time. Groth16 proofs, used by many rollups, require a trusted setup and are relatively cheap to verify. But the generation of the proof itself is computationally intensive. PLONK-based proofs, which are more flexible, are even more expensive. In a bull market, when users are paying $50 per transaction on Ethereum mainnet, a rollup can charge $1 and still be profitable. But now, with Ethereum's base fee averaging $2, users expect rollup fees to be under $0.10. The gap is massive. Trustless systems require trusting relationships. And right now, I trust the math more than the fundraises.
I've spoken to five engineers from different rollup projects. Off the record, they admit the current situation is untenable. One told me, 'We're basically paying people to use our network. We're hoping that a bull market saves us before we run out of cash.' Another project I audited recently had to cut its prover hardware budget by 40% because the burn rate was too high. They switched to a less secure proving system to save costs. That's the kind of corner-cutting that leads to exploits. The pivot wasn't technical; it was financial.
Contrarian
Now, the contrarian angle. The common narrative is that 'liquidity fragmentation' is the real problem for L2s. VCs push this story because it justifies new cross-chain bridges and interoperability protocols. But I think that's a distraction. The real problem is that the core value proposition of ZK rollups—cheap, trustless scaling—is predicated on an assumption that transaction volume will always be high enough to cover fixed costs. That assumption is broken in a bear market. And even if volume returns, the proving costs will only grow as more activity shifts to L2s. Ethereum's blob space (EIP-4844) was supposed to help, but it only reduces data availability costs, not proving costs. The proof is the bottleneck. Trust is code now. But the code is expensive.
Another blind spot: the market is pricing ZK rollup tokens as if they are growth stocks, but they are actually infrastructure plays with high operational leverage. When volume drops, margins collapse. The teams that survive will be the ones with the deepest pockets or the most efficient provers. But efficiency gains are marginal. The next generation of proof systems (like Halo2 or recursive proofs) might drop costs by 50%, but that still leaves a 3x gap. We need a 10x improvement, not a 2x. Silence is loud in crypto. And the silence from the ZK teams about their economics is deafening.
Takeaway
I learned to stop preaching and start listening. The data is clear: ZK rollups are not economically viable in a bear market unless they have a massive subsidy. As an investor, ask yourself: how long can the treasury last? As a user, ask yourself: are you paying a fair price, or are you being subsidized by venture capital? The real test will come when the next bull market arrives. If proving costs haven't dropped by an order of magnitude, the entire L2 thesis will be called into question. Energy is the new equity. But in this case, the energy is computational, and the equity is running out.
