Every time a solver fills a cross-chain intent — handing you USDC on Base while you have USDC sitting on Arbitrum — they are not just completing your trade. They are also creating a problem for themselves. Their Base inventory just shrank. Their Arbitrum inventory will swell once the protocol repays them. To stay solvent across a dozen chains simultaneously, a solver must constantly bridge funds back to wherever they ran short. That rebalancing process has a cost. And that cost, silently embedded in every spread a user pays, is the structural friction that Everclear was built to remove.
The Hidden Cost: Solver Inventory Imbalance
Solvers filling intents — as explained in Part 3 of this series covering cross-chain intent protocols — must front liquidity on the destination chain and wait to be repaid on the source chain. The gap between those two moments leaves them holding fragmented balances across many networks at once.
Each bridge rebalancing trip incurs fees, slippage, and latency. Everclear's own research quantifies the damage: solvers currently execute roughly 5x more bridging volume than the underlying net demand actually requires. That multiplier is the rebalancing tax — the overhead a solver absorbs just to stay balanced, before they earn a single cent of profit.
The economics of this tax are not neutral. Only large, well-capitalised solvers can afford to maintain permanent inventories on every chain. Smaller participants get squeezed out. The result is centralisation pressure in what should be a competitive market: fewer solvers competing for fills means wider spreads, slower execution, and worse prices for users — even if the intent protocol itself is technically decentralised.
The Netting Insight: What Traditional Finance Already Solved
The idea behind Everclear is old. Traditional clearing houses — Visa's settlement network, the Depository Trust & Clearing Corporation, interbank payment systems — do not settle every individual transaction gross. They aggregate all flows over a period, net the offsetting positions against each other, and only move the residual difference.
The intuition is simple. Imagine $80 worth of USDC flows from Arbitrum to Base every day. Imagine $80 also flows the opposite way. Those two streams cancel each other. Neither bridge nor gas is needed; every participant simply credits and debits on paper. Only deviations from parity — an imbalance of, say, $12 in one direction — need to physically cross a bridge.
Everclear applies this principle on-chain. The protocol's stated design target is that roughly 80% of daily inter-chain flows can be matched against opposing flows, leaving only about 20% as true residual that needs to move. One important caveat: the 80% is Everclear's objective, not a guaranteed result. During its December 2024 mainnet-beta phase, the protocol reported netting up to 50% of transactions at that early stage. The gap between target and achieved is real, and it narrows as the network grows.
Gross settlement moves every dollar every time. Net settlement moves only the difference. Traditional payment networks run thousands of transactions per second but settle them in batches after netting — dramatically reducing infrastructure strain. Everclear is that batch-netting layer for cross-chain solvers.
How Everclear Works: Intents, Queues, and the Epoch Cycle
Everclear is built as an Arbitrum Orbit rollup using an AnyTrust design. Cross-chain messaging runs over Hyperlane, and economic security is provided by EigenLayer. The rollup itself is the coordination hub; everything else is spoke contracts deployed on individual chains.
When a solver wants to rebalance — say, they are long USDC on Base and short on Arbitrum — they post a rebalancing intent to the Everclear spoke contract on Base. That intent flows into a three-stage queue on the hub:
- Intent Queue — incoming rebalancing requests arrive here.
- Invoice Queue — unmatched intents wait for a counterparty.
- Deposit Queue — matched pairs that are ready to settle.
Every epoch (currently 30 minutes), the hub scans the invoice queue and attempts to pair opposing invoices. A solver long USDC on Base is matched against a solver short USDC on Base. When a pair clears, each spoke settles the position locally — no bridge movement occurs. Only the net residual, if any, needs to cross.
When an intent finds no match after several epochs, a Dutch-auction fallback activates: the transfer amount is progressively discounted each epoch until an arbitrageur finds it profitable to step in as the counterparty. This ensures intents always eventually clear, even in thin markets.
At full mainnet launch in 2025, the spoke-hub model supported Ethereum, Arbitrum, Optimism, Base, Polygon, BNB Chain, and Gnosis, with Solana becoming its first non-EVM integration, bringing total chain coverage to 18.
The CLEAR Token and Protocol Economics
Everclear rebranded from Connext on June 3, 2024. The token migrated from NEXT to CLEAR on a 1:1 basis following a DAO vote approved in December 2024. CLEAR has a capped supply of 1 billion tokens; as of March 2025, roughly 312 million were circulating, with approximately 25 million staked.
Staking CLEAR for up to two years yields vote-bonded CLEAR (vbCLEAR), which grants two things: governance votes on which chains receive CLEAR emissions, and a share of protocol fees (on the order of a fraction of a basis point per settled transaction).
Solver incentives are denominated in CLEAR emissions and directed by vbCLEAR governance. Solvers settling more volume on a given chain receive more rewards, aligning protocol economics with network growth. Core contributors and early backers hold locked allocations that unlock over roughly 18 months from December 2024 — a schedule worth tracking for governance balance and potential supply pressure.
Why This Matters to Ordinary Users
Everclear is invisible to end users. It is pure business-to-business settlement infrastructure. No user ever interacts with a queue, an invoice, or an epoch. You never sign an Everclear transaction.
Its impact is indirect but real. When solvers can rebalance at a tiny fraction of a percent rather than paying multi-hop bridge fees and slippage, they can afford to quote tighter spreads on the intents users sign. Better-capitalised solvers — because they need far less inventory per chain — means a larger, more competitive solver market, which in turn means faster fills and better pricing.
The growth numbers from the beta phase reflect this dynamic. Volume grew from roughly $5M in January 2025 to about $125M by March 2025 as solver participation increased. This also illustrates the network-effects property: netting efficiency improves as opposing flows grow, so the protocol becomes more valuable as it scales.
For context on why permissionless smart contracts that replace intermediaries are the foundational layer this all rests on, or on how on-chain lending and liquidity protocols depend on similar arbitrageur incentives to remain solvent, the underlying mechanics rhyme.
Like any composable DeFi system, concentration of cross-chain flows through a single coordination point also raises the systemic questions covered in how interconnected protocols can fail together.
Risks to Understand Before You Participate
Smart-contract and upgrade risk. L2BEAT gives the Everclear hub a CRITICAL rating for upgradability: the team can push code changes without a timelock. No delay means no window for users to exit before a change takes effect. This is common at early-stage protocols, but it is a real trust assumption.
Centralised validation surface. The AnyTrust model relies on a Data Availability Committee and a small challenger set. A colluding subset could in principle finalise a malicious state transition.
Settlement hub as a single coordination point. Netting concentrates cross-chain flows through one on-chain system. A bug in the hub contract, a Hyperlane messaging failure, or an EigenLayer security outage could disrupt settlement across all connected chains simultaneously. A fully fragmented bridge ecosystem would not have this single-point failure mode.
Residual bridge risk is not eliminated. Netting removes the majority of bridging, but the unmatched residual still traverses a bridge. Solvers and users remain exposed to bridge risk for that 20% (or more, at current netting rates).
Protocol immaturity. Everclear's mainnet beta launched in September 2024 and full mainnet in 2025. The netting model and Dutch-auction fallback have not been stress-tested through a sustained period of extreme volatility or very low solver participation.
Token unlock dynamics. A large share of CLEAR supply is held by insiders on an approximately 18-month unlock schedule from December 2024. Large periodic unlocks could create selling pressure and shift governance participation as early allocations vest.
Key Takeaways
- Solvers powering the intent economy currently perform roughly 5x more bridging than net demand requires — a rebalancing tax that inflates spreads for users.
- Everclear applies the clearing-house netting model on-chain: match opposing flows first, bridge only the residual. Its stated design target is netting ~80% of flow; its December 2024 beta reported up to 50% in practice.
- The hub is an Arbitrum Orbit AnyTrust rollup using Hyperlane for messaging and EigenLayer for economic security; epochs run every 30 minutes with a Dutch-auction fallback for unmatched invoices.
- CLEAR (migrated from NEXT 1:1) governs emissions and earns protocol fees via vbCLEAR staking; insider unlocks over ~18 months from December 2024 are a supply and governance dynamic to monitor.
- Key risks: no upgrade timelock (CRITICAL per L2BEAT), AnyTrust committee trust assumption, single hub as a concentration point, residual bridge exposure, and limited battle-testing under stress conditions.
