There are projects that get patched and projects that get torn down to be rebuilt. Oneiro chose the second path: barely a week ago it announced Revolution Network V2, a complete rebuild of its Ethereum layer 2 on ZKsync OS, with validity proofs in every batch from the public testnet onward, according to the official statement distributed by FinanceWire from Wilmington. The company promises a standard, boring base, in its own words, to support programmable staking and a trust layer designed for commerce run by autonomous agents.
Context helps explain the move. The current network, known as Libertas, runs on EraVM, the stack ZKsync has announced it will retire, so keeping it would have meant dragging along custom pieces doomed to lose support. Oneiro now publishes what it learned from that first version, freezes its mistakes at the root and bets on the standards already used by wallets, transaction bundlers and auditors. Everything below its value proposition becomes common, proven infrastructure.
The move comes at a time when layer 2s are competing to show they inherit Ethereum's security without demanding blind trust. V2 will publish each transaction's data in Ethereum blobs and allow the entire chain to be rebuilt from what is stored on layer 1 alone, while every batch will carry its Airbender proof from the first day of the public testnet, named Virtus. If the schedule holds, Libertas holders will not have to move anything until a migration with an announced freeze block.
The rebuild on ZKsync OS
Oneiro abandons custom code and starts from a new genesis on ZKsync OS, ZKsync's operating system for layer 2s, whose official documentation describes an environment equivalent to the Ethereum virtual machine. Contracts compile with standard Solidity and the usual tools work without changes or adaptations.
For teams already building on the network, the message is full continuity in day-to-day work. Foundry, Hardhat, viem and ethers operate as before, so no one needs to rewrite deployments or learn a new dialect to keep working.
Rebuilding is not migrating, and the difference matters. Migrating would have meant dragging Libertas adaptations onto a new stack, with the risk of inheriting subtle flaws. Starting from zero allows applying from the outset what V1 taught.
Airbender proofs in every batch from day one
The technical heart of the announcement is the Airbender prover, which will generate a validity proof for every batch from the moment Virtus opens its doors. That proof cryptographically demonstrates that the state change is correct, so layer 1 can accept it without repeating all the computations.
The promise includes full discipline from the start. No grace periods with unproven batches and no permissive modes during the first weeks. Every batch will carry its proof and nodes will verify it independently.
This point deserves the caution of any corporate announcement. The prover's performance and timings come from the company itself and there is not yet a public network where they can be measured. Phase 1 must deliver the real prover before Virtus opens.
Data will live in Ethereum blobs
V2 will publish each transaction's information in blobs, the format Ethereum launched to lower the cost of layer 2s, and which Ethereum's official documentation presents as the path to scaling with rollups. Anyone will be able to rebuild the entire chain from layer 1.
That property has an ugly name and a nice consequence. It is called reconstructibility and it means that, even if Oneiro disappeared tomorrow, the history would remain available on Ethereum for anyone who wanted to reassemble it without asking permission.
The cost of that guarantee is paid by blobs, whose price fluctuates with demand across the rollup ecosystem. When many networks publish at once, blob gas gets more expensive. Oneiro does not yet detail its publication cadence or its cost per seat.
Farewell to Libertas and the EraVM stack
Libertas was the first version's testnet and is now on borrowed time. It runs on EraVM, ZKsync's original environment, which ZKsync itself will retire in favor of its new operating system. Staying there would have left Oneiro without support or audits.
The company chooses to document what it learned rather than hide it. The statement speaks of publishing what V1 taught and fixing every problem at the root. Fifty-one contract tests now pass on the internal devnet with the new stack.
For current users, the instruction is to do nothing. Libertas will be frozen at a block announced in advance and balances will be transferred according to published rules. Be wary of supposed migration forms already circulating.
250 million REVO with the same curve as always
The token economy is untouched. The published total supply and its curve remain at 250 million REVO over roughly ten years. The technical rebuild dilutes no one and does not accelerate the unlock schedule.
Programmable staking remains the piece Oneiro considers its true product. On the standard base, the network aims to offer flexible contracts and smart accounts so agents can operate with clear limits. The Cornerstone paymaster sponsors gas for certain operations.
It remains to be seen how value is split among validators, provers and delegators with real traffic. The documents do not detail fees or per-batch rewards. Virtus will serve to calibrate that machinery before serious money is involved.
| Item |
Libertas network (V1) |
Revolution Network V2 |
| Technical base |
EraVM, stack being retired |
ZKsync OS, environment equivalent to Ethereum |
| Validity proofs |
No announced full discipline |
Airbender in every batch from the testnet |
| Data |
Previous publication model |
Ethereum blobs, reconstructible chain |
| Tools |
Custom adaptations |
Standard Solidity with Foundry and Hardhat |
| Supply |
250M REVO over ten years |
No changes to total or curve |
| Status |
Testnet to be frozen |
Devnet with 51 tests, Phase 1 underway |
Oneiro and its bet on agents
Oneiro NA is a Delaware-registered company that presents its network as infrastructure for digital ownership and identity. Its director, Rob Frasca, sums up the thesis by saying its value lies in programmable staking and a trust layer for agentic commerce.
The bet on agentic commerce is no accident. More and more autonomous programs move funds and hire services from each other, and they need rails where they can operate with limited permissions. Smart accounts fit exactly there, with spending rules and time windows.
Half the industry says the same thing today and courts the same agents with similar promises. Whoever offers documentation, audits and tools that work will win. Oneiro seems to have understood this by rebuilding on open standards.
Programmable staking is where that thesis puts its money on the line. If the contracts allow delegating with fine-grained conditions and revoking permissions without friction, agents will have a real reason to stay. Without that piece, the network would be just another correct layer 2 among dozens.
The roadmap from Virtus to mainnet
Phase 1 concentrates all the visible work of this quarter. It includes the real prover, the verification nodes, the explorer, the bridge and Virtus itself, five pieces that must work together before thinking about mainnet. The internal devnet already validates the contracts with its 51 tests passing.
The order matters because each piece depends on the previous one. Without a prover there are no verifiable batches, without nodes no one checks them and without a bridge funds cannot enter or leave. The explorer closes the circle by letting anyone inspect blocks, proofs and settlements on Ethereum.
After Virtus will come what the statement does not date. External audits, a bug bounty program and exit criteria toward mainnet form the usual package in these launches. Oneiro has not yet published any of those milestones, so it is worth asking for them out loud.
- Real Airbender prover generating proofs in every Virtus batch
- Independent verification nodes checking every published proof
- Open explorer with traceability of batches, proofs and blobs
- Audited bridge to move funds between Ethereum and the testnet
- Virtus opening date with genesis and contract addresses
A big step to mainnet
Virtus is now the exam that decides everything. It will have to show that Airbender generates proofs at open-network pace, that the bridge moves funds without incidents and that the explorer withstands the curiosity of real developers.
Today the schedule remains open and any countdown circulating on social media is made up. The company has not given an opening date and Phase 1 is still underway. Follow the official channel and demand contract addresses and audits.
If Virtus delivers, Oneiro will reach mainnet with something rare among its rivals. It will not be the fastest network on paper, but one of the few that started from zero to avoid inheriting technical debt. In a sector fond of shortcuts, rebuilding slowly may be the fast track to trust.
The remaining question is whether the market will reward that patience. Layer 2s live on liquidity and developers, two audiences impatient by nature. Virtus has the floor, and the schedule, though open, favors whoever shows working code.