Boundless: Cross Chain
Apr 18 · $292M drained · 20+ OFT bridges frozen

Strengthen your bridge with ZK verification.

Most cross-chain systems still rely on offchain observers to say an origin-chain event happened and is final. Boundless adds a cryptographic verification layer to that model, starting with out-of-the-box LayerZero DVN support today.

Alpha status. Live LayerZero testnet example. Not ready for production use yet.
Security model
Independent
verification layer
Packaged today
LayerZero
DVN support
More coming soon
Integration
1 config
change on supported routes
Shared trust assumptions

Most cross-chain systems diversify operators. They still share similar trust assumptions.

The shared model

LayerZero calls them DVNs. Wormhole calls them Guardians. Others use validators or DONs. The pattern is similar: offchain actors observe an origin-chain event, decide it is irreversible, and attest to it on the destination chain.

Apr 18 · what the KelpDAO exploit made clear
$292M
drained from KelpDAO
$196M
bad debt on Aave
20+
OFT bridges frozen
1
failure mode exposed

The KelpDAO exploit made the failure mode plain. An attacker poisoned the RPC nodes a configured LayerZero DVN relied on for source-chain data. The attestation looked valid to the destination system because the observer had already been fooled.

That is not unique to one bridge. Most modern cross-chain systems still reduce to operator honesty and infrastructure integrity at the moment they decide what happened on the source chain.

Boundless changes the evidence, not just the operator set. A ZK-backed verifier path reduces to validator signatures, proof verification, and circuit soundness. It is slower than the fastest trust-based path, but it is materially harder to spoof in the same way.

What layer diversity buys you

The point is not just more walls. The point is walls whose holes do not line up.

1 verifier
One DVN leaves a clean path.
SOURCEDESTINATIONRPCKEYOPSDVN 1
Attack reaches end

If the main hole is poisoned observation, there is only one wall to get through.

Operator diversity
More trust-based DVNs help, but shared holes can still line up.
SOURCEDESTINATIONRPCKEYRPCOPSRPCKEYDVN 1DVN 2DVN 3
Shared path remains

Three walls are better than one. If all three still depend on similar observation infrastructure, one path can remain.

Add ZK layer
A different layer breaks the aligned path.
SOURCEDESTINATIONRPCKEYRPCOPSRPCKEYBUGCRYDVN 1DVN 2DVN 3ZK LAYER
Attack stopped

The trust-based walls can stay. Once the last wall reduces to proof verification and validator signatures, the same RPC path no longer makes it through end to end.

RPC = poisoned observationKEY = key compromiseOPS = operator failureBUG / CRY = circuit or cryptographic break

The goal is not to remove every trust-based path. The goal is to build a system where high-value routes can require stronger evidence, while faster routes still keep practical latency. Boundless-DVN is the current LayerZero packaging of that model.

How it works

Same packet flow. Stronger evidence at the verification step.

01
Your bridge flow runs normally.
The source application emits its packet or message as usual. On LayerZero routes today, SendLib assigns the verification job to Boundless without OApp code changes.
02
Boundless requests the proof tier that fits the route.
Helios, Signal, Steel, and similar verification paths cover different source chains and security targets. That gives the system tiers of assurance instead of one fixed proof model.
03
The destination verifier checks the proof on-chain.
The destination contract verifies the seal against the expected image ID and only then passes verification upstream. Invalid proof, invalid delivery.
04
Existing verifier paths stay in place.
On LayerZero, Boundless slots into the DVN set through config. The same source-proof plus destination-verifier pattern can be adapted to other bridge integrations without replacing the bridge itself.
Alpha end-to-end example · Sepolia ↔ Base Sepolia

Out-of-the-box LayerZero support is live in alpha today.

This Sepolia ↔ Base Sepolia corridor proves the packaged LayerZero path today: packet emission, proof request, destination verification, and delivery. Boundless is the broader cross-chain verification product; this alpha shows the current LayerZero implementation surface.

Ethereum Sepolia
source · eid 40161
OApp.send()
PacketSent
SendLib.assignJob
Boundless
ZK proving · marketplace
ProofRequest · signal-eth
Prover fulfills seal
Seal ready · R0 Groth16
Base Sepolia
destination · eid 40245
BoundlessDVN.attest()
ReceiveUln302.verify()
OApp.lzReceive()
Current step
Ready
Click Send to run the current LayerZero implementation path.
adapter.log
connected
Waiting for job…
Alpha status. This is a live end-to-end alpha, not a mock, and it is not ready for production use yet. It demonstrates the current LayerZero packaging on a real Sepolia ↔ Base Sepolia corridor today.
Assurance tiers with real costs

Safe cross-chain systems should be tiered: faster paths for routine flow, heavier proofs for higher-value routes.

The mistake is to talk about ZK as if it were one fixed latency profile. It is not. Trust-based paths remain the speed tier. Helios, Steel, and similar systems are the fast cryptographic tier. Full finality is the premium tier. The numbers below come from measured or benchmarked paths we use for this product.

Fast cryptographic tier
Helios
Ethereum consensus light-client proof. Faster than full finality, materially stronger than plain RPC observation.
1.19Bcycles
Measured against live mainnet · 2026-04-21 · slot 14,164,555
~$0.04
At the current ~$0.03 / 1B cycles marketplace rate, a 1.19B-cycle proof is about ~$0.04 before destination-chain verify gas. Helios proofs can land in under a minute.
Good default when a route needs cryptographic verification without full-finality latency.
Highest-assurance tier
Signal-Ethereum
Casper FFG finality over the full validator set. The premium path for Ethereum-origin value.
~37Bcycles
Per finality proof at the current measured Signal-Ethereum path
~$0.37
Signal-Ethereum currently clears closer to ~$0.01 / 1B cycles, so a ~37B-cycle finality proof is about ~$0.37 before destination-chain verify gas. Slower than Helios, but appropriate when a route secures $20M+ or when counterparties want full-validator slashable security.
Use when the highest-assurance path is worth a few extra minutes of latency.

The right system mixes these modes instead of forcing one proof depth everywhere. Fast routes can keep trust-based attestation. Higher-value routes can require a Helios-class proof. The routes that justify it can step up again to full finality. That is how the system becomes both safer and practical.

How the tiers fit together
Tier
Primitive
Latency today
Cost today
When to use it
Speed tier
Existing trust-based verifiers
Near-instant
No proof cost
Low-value or latency-sensitive flow
Fast cryptographic tier
Helios / Steel / blobstream0
Under a minute for Helios; lower on lighter routes
~$0.04 per Ethereum L1 proof at today's rate; lower on lighter paths
Default upgrade when evidence quality matters
Highest-assurance tier
Signal-Ethereum
Extra minutes today
~$0.37 per finality proof at today's rate
Treasury moves, issuance routes, transfers securing $20M+

Destination-side on-chain verify() remains roughly ~35K gas on L2s, with L1 routes quoted separately. That cost sits underneath every proof tier.

The batching multiplier

One proof covers every message in its window. Batch your outbound flow and per-message cost collapses toward the verify gas.

A Helios proof covers a ~27-hour sync-committee window. A Signal proof covers a finality event on Ethereum mainnet. A blobstream0 proof can cover 1,000+ Cosmos headers at the same cycle cost as one. You don't pay per-message — you pay per proof, and the proof is shared across every message the verifier path attests in that window.

1 msg / window
$0.04
/ msg
10 msgs / window
$0.004
/ msg
100 msgs / window
$0.0004
/ msg
1,000 msgs / window
< $0.0001
/ msg
Illustrative, using the Helios baseline (~$0.04/proof) at the current ~$0.03 / 1B cycles rate. Destination-chain verify gas is additional. Steel and blobstream0 batching follow the same logic.
Why the economics work.
Every proof request runs a reverse auction on Boundless. Provers compete for the work, which keeps proof pricing tied to actual market supply instead of a fixed operator markup.
$1.00 → $0.03 per 1B cycles
Marketplace rate · 2024 → 2026
Why this scales.
Not every route needs the heaviest proof, and not every message needs its own proof. Match evidence depth to value at risk, batch where you can, and the system stays practical as coverage expands.
Volume & SLA terms
Quoted by conversation

These are current benchmark-backed ranges, not promised sticker prices. Actual route pricing depends on proof tier, message frequency, destination-chain gas, and whether the route warrants dedicated capacity.

Why this is deployable now

The proof systems are audited, the proving market is live, and the integration model already works.

Sigma Prime
Signal-Ethereum audit
Full Ethereum finality is already provable.
This is the high-assurance tier for Ethereum-origin routes where the value at risk justifies the extra latency and cost.
a16z
Helios audit
Fast light-client proofs already exist.
Helios is why ZK verification is now practical for more than the rarest treasury move. It gives a faster cryptographic tier for Ethereum-origin routes.
74.1T
Signal cycles / week
Real proving volume, not a lab exercise.
These systems already run at production scale on the Boundless marketplace. The cost and operational model are not hypothetical.
3,000+
Provers competed to date
Competitive proving instead of one operator.
Every proof request runs through a reverse auction, so the verifier path does not collapse to one prover or one piece of infrastructure.
Current packaged path

One LayerZero setConfig() tx. No bridge replacement.

Boundless is the broader cross-chain verification product. Boundless-DVN is the packaged LayerZero integration today. On supported LayerZero routes, you add it as an enforced verifier layer without replacing your bridge or touching OApp code. Snippet below is illustrative — exact values depend on your OApp's DVN set and route set.

Live alpha
Ethereum Sepolia → Base Sepolia
Packaged today
LayerZero OFTs on enabled routes
Route rollout
Corridor by corridor, based on proof path
Pricing
Proof-tier dependent, marketplace pass-through
SLAs / volume terms
By conversation
// 1. Boundless-DVN is already deployed for you on the destination chain.
// Share the route you want to secure and we provide the current addresses.
 
address constant BOUNDLESS_DVN = 0x0000_...9AF2; // Base mainnet
address constant LZ_LABS_DVN = 0x589D...5BE8; // keep your existing required DVN
 
UlnConfig memory cfg = UlnConfig({
confirmations: 15,
requiredDVNCount: 1,
optionalDVNCount: 1,
optionalDVNThreshold: 1, // message cannot commit without ZK proof
requiredDVNs: toArray(LZ_LABS_DVN),
optionalDVNs: toArray(BOUNDLESS_DVN)
});
 
SetConfigParam[] memory params = new SetConfigParam[](1);
params[0] = SetConfigParam({
eid: BASE_EID,
configType: 2, // ULN config
config: abi.encode(cfg)
});
 
// 2. One tx, OApp-owner signed.
ILayerZeroEndpointV2(ENDPOINT_V2).setConfig(oapp, receiveLib, params);