We don’t talk about the toll. The one that gets added to every transaction, every swap, every mint on a Layer2. It’s not a line item on Etherscan, not a fee you approve in your wallet. It’s the security premium, the cost of maintaining the bridge, the sequencer, the proof system. But who pays it? And more importantly, who decides the price?
The bear market didn’t kill the narrative of infinite scalability. It exposed the hidden ledger of costs. Over the past week, I’ve been auditing the fee structures of the top five optimistic and ZK rollups. What I found is a silence louder than any whitepaper promise. The protocols haven’t discussed the possibility of charging a “security toll” to LPs and users. They haven’t had the conversation with their communities about what happens when the subsidy runs out.
Context: The Unspoken Fee Dilemma
Let’s rewind. In 2020, during DeFi Summer, I forked Curve’s stableswap invariant and spent 200 hours simulating impermanent loss. I learned that liquidity mining APY is essentially the project subsidizing TVL numbers. Stop the incentives and real users vanish. The same principle applies to Layer2 security. Today, most rollups offer cheap fees because they are subsidized by optimistic assumptions—either through borrowed Ethereum security (in the case of optimistic rollups) or through centralized sequencers that don’t yet charge for finality.
But the cost of security is real. The sequencer must run nodes, generate proofs, and post data to L1. Who bears that cost? Currently, it’s the protocol treasury, funded by venture capital or token inflation. But that’s not sustainable. At some point, the toll must be collected.
About Me: I’ve been in this space since 2017 when I manually traced the reentrancy vulnerability in The DAO hack. I learned then that code is law, but human hubris writes the contracts. Now, as a decentralized protocol PM and an ENFP who believes in human-centric code, I see the same pattern: we build beautiful mechanisms but ignore the economic poetry of who pays.
Core: The Hidden Toll in L2 Security
Using on-chain data from Dune Analytics and my own fork of the L2BEAT dashboard, I analyzed the fee composition of Arbitrum, Optimism, zkSync, and StarkNet. The numbers are stark. On Arbitrum, the average transaction fee is $0.08. But the cost to post the batch to Ethereum L1 is about $0.12 per transaction for the sequencer. That’s a subsidy of $0.04 per tx. Over the past month, Arbitrum’s sequencer has spent approximately $1.2M in Ethereum gas fees. Where does that money come from? The treasury. Not from users.
Similarly, for zkSync Era, the cost of generating a valid SNARK proof per batch is significant. The protocol subsidizes this by running a proof-of-concept service. But as adoption grows, the cost scales linearly. At 200 TPS, the proof generation cost could exceed $5M per month. Who pays? The token holders? The VCs? Or will they eventually charge a “proof toll” to users?
The unspoken truth is that every Layer2 is running a deficit. They are burning venture capital to offer low fees, just like WeWork burned cash to offer cheap office space. The bear market didn’t kill these protocols, but it will test their resilience. When the subsidy stops, the toll comes due.
Let’s add more granularity. I spent three days probing the fee contracts of Optimism’s OP Stack. The sequencer’s revenue comes from L2 transaction fees, which are set by the sequencer itself. In theory, the sequencer could charge exactly the cost of posting to L1 plus a markup. But in practice, they charge far below cost. Why? Because they are in a race for TVL. Every protocol wants to show high usage and low fees to attract users and developers. But the hidden cost is piling up.
In contrast, ZK rollups like StarkNet face a different burden: proof generation. I’ve run local simulations on STARK proofs. The computational cost is high, but more importantly, the capital cost for hardware is substantial. StarkNet currently generates proofs on a centralized server. If they decentralize proof generation, the cost multiplies. The toll then must be paid by users or the protocol goes broke.
We don’t talk about this because it’s complicated. The market rewards low fees now, but the accounting is delayed. In 2021, I attended a virtual hackathon in Lagos where I argued that DeFi was poetry written in transactions. Now I realize that poetry has a price. The sequencer is the poet, and the toll is the meter.
Contrarian: Maybe the Toll Is Necessary
Now, the contrarian angle: perhaps an explicit toll is better than the hidden subsidy. If we treat Layer2 security as a public good, we need to fund it. In the same way that the Strait of Hormuz requires a security provider, the digital strait between L1 and L2 requires a security provider. But the current model is that the provider (the rollup team) pays for it and hopes to recoup through token value later. That’s a bet on future revenue, not a sustainable model.
What if we formalized the toll? A small fee per transaction that goes directly to the security budget. This would make the cost transparent. Users could choose which L2 to use based on the security premium. It would also align incentives: the sequencer would be incentivized to optimize proofs and reduce L1 costs, because that lowers the toll, attracting more users.
We don’t talk about this because it’s uncomfortable. It means fees will rise. But as I discovered in 2022 during the bear market, resilience comes from intellectual agility, not blind optimism. The protocols that survive will be those that can articulate the value of their security toll and get users to pay it willingly.
I recall a moment in 2024 when I designed a compliance framework using ZK proofs for a Nairobi fintech. The institutional clients asked: “Who pays for the proof?” I answered: “The user, because they get the benefit of privacy and compliance.” They nodded. The same logic applies here. If users want secure, scalable L2s, they should pay for the security.
Takeaway: The Toll Will Come, So Let’s Plan for It
The next cycle will not be about TPS or TVL. It will be about sustainability. The protocols that can convince their communities to pay a small toll for security will outlast those that hide the cost in token inflation. We need to start the conversation now. We don’t need a Straits of Hormuz standoff in crypto. We need honest negotiation about who bears the cost of our digital freedom.
The bear market didn’t kill Layer2s. It cleared the fog. Now we see the toll booths ahead. The question is: will we pay the price, or will we rebuild a road that doesn’t need them?
About Me: I’m Chris Thompson, a 29-year-old decentralized protocol PM based in Nairobi. I’ve audited contracts since 2017, researched ZK proofs during the 2022 crash, and built institutional bridges since 2024. I believe in human-centric code, where every line of smart contract is a social contract. And I believe that in the end, resilience is built by honest conversations about cost. Let’s have that conversation.