Polygon AggLayer Architecture and 2024-2026 Rollout · CDK + Unified Bridge + ZK Aggregation

TL;DR

Polygon’s strategic shift in 2024-2026 is a transformation from “Polygon PoS sidechain” to “AggLayer + CDK multichain protocol layer”: Polygon CDK lets any team build its own ZK rollup, while AggLayer provides a unified bridge + pessimistic proof + ZK proof aggregation across CDK chains, aiming to capture the “intra-stack interop” pole in Cross-chain Five-pole Comparison. The MATIC → POL migration (2024-09 mainnet) changed the token model from “Polygon PoS gas + staking” to “AggLayer full-stack gas + restaking,” and Polygon PoS itself was demoted from the strategic center to one of many AggLayer chains, 1 in number. Early CDK partner chains included Astar zkEVM, Manta Pacific (later withdrawn), Immutable zkEVM, X1 (OKX), Palm, IDEX, etc., but outflows became evident in 2025-2026, and the actual number of AggLayer v0.2 - v0.3 connections is far below the forecasts made at the beginning of 2024.

Wiki route

This entry sits under systems index. Read it against Cross-chain Five-pole Comparison Matrix to confirm AggLayer’s position among LayerZero / Hyperlane / CCIP / Cosmos IBC, and compare with Chain Abstraction Pattern Overview and Pectra EIP-7691 Blob Expansion to understand Polygon ZK rollups’ position in Ethereum L2 economics.

Polygon History Stages · PoS → zkEVM → AggLayer

Phase 1 (2017-2020): Matic Network: Plasma + PoS sidechain, EVM compatibility, low gas as the main axis, 2020-05 mainnet launch
Phase 2 (2021-2023): Polygon PoS golden age: DEX (QuickSwap, Uniswap V3) + NFT (OpenSea + Aavegotchi) + games (Sandbox) briefly pushed TVL above $5B
Phase 3 (2023-2024): multiproduct expansion: Polygon zkEVM (2023-03 mainnet) + Polygon Miden (STARK-based VM) + Polygon Avail (DA layer, later independent) + Polygon CDK (2023-08)
Phase 4 (2024-2026): AggLayer shift: 2024-02 AggLayer v0.1 launch, 2024-09 MATIC → POL migration, 2024-Q4 AggLayer v0.2 introduces pessimistic proof, 2025-2026 v0.3+ continues iterating, and Polygon PoS’s strategic position declines

AggLayer Architecture

AggLayer is not 1 chain, but a settlement + bridge protocol spanning multiple ZK rollups:

Unified Bridge: all AggLayer-connected chains share 1 canonical bridge contract (deployed on Ethereum L1 ); when users move assets from chain A to chain B, they go through the unified bridge and do not need an independent wrapped token
Pessimistic Proof (v0.2, 2024-Q4): prevents a particular AggLayer chain from withdrawing assets beyond its deposits (bridge insolvency protection), and is effective even if that chain’s ZK proof system has a bug
ZK Proof Aggregation: aggregates state transition proofs from multiple AggLayer chains into 1 L1 verification, greatly reducing single-chain L1 verification cost
State Sharing (roadmap): cross-chain asynchronous messaging + final state sharing (similar to Cosmos IBC + Ethereum L2 cross-rollup communication), on the roadmap
POL as gas + staking: POL can be restaked across multiple chains inside AggLayer; similar to EigenLayer, but limited to the Polygon stack

Difference from conventional Ethereum L2 s: each Optimism rollup or Arbitrum chain has an independent canonical bridge, and cross-rollup asset transfer depends on third-party bridges (LayerZero V2 / Hyperlane / CCIP). AggLayer tries to solve this “cross-rollup trust assumptions + bridge fragmentation” inside the stack.

Polygon CDK (Chain Development Kit)

CDK is the Polygon version of OP Stack / Arbitrum Orbit / zkSync ZK Stack:

Basic components: Polygon zkEVM prover (Plonk + KZG based) + sequencer + RPC + bridge contract template
Optional DA: Ethereum blob (default, expensive) / Polygon Avail / EigenDA / Celestia
CDK Erigon vs CDK OP Stack: since 2024 , Polygon Labs has maintained 2 CDK tracks at the same time — Erigon-based (original zkEVM route) + OP Stack-based (opened to teams that already selected OP Stack), reflecting Polygon’s tension between “protecting investment in its own prover” and “joining OP’s large ecosystem”
AggLayer connection is optional: CDK chains can choose not to connect to AggLayer, and vice versa (though AggLayer connection is the core route promoted by Polygon Labs)

Partner chains · List and Outflows

CDK / AggLayer partner chains announced by Polygon at the beginning of 2024:

Astar zkEVM (Japanese ecosystem, led by Astar Network, 2024-03 launch) — effectively wound down in 2024-Q4 due to insufficient user adoption
Manta Pacific (evaluated CDK at the beginning of 2024, later adopted OP Stack + Celestia, and did not reach deep practical integration with Polygon)
Immutable zkEVM (gaming L2, 2023-11 launch) — still operating in 2025, but traffic is below expectations
X1 / OKX Chain (OKX exchange L2, 2024-Q1 testnet) — renamed X Layer in 2025 and continues operating
Palm Network (NFT-focused, ConsenSys origin) — migrated to CDK
IDEX (DEX proprietary L2)
Gnosis Pay rail (some components): exploratory connection
Wirex / OKX and other wallet / exchange L2 projects

Reasons for outflow: OP Stack’s network effect (Base / OP Mainnet / Worldcoin / Zora) + the Arbitrum Orbit ecosystem (gaming chains such as Xai / Sanko) + the Celestia DA camp such as Manta have continued compressing CDK’s “RaaS / app-chain” market share in 2025-2026.

MATIC → POL Migration and Business Model

In 2024-09 the POL token officially launched on mainnet, replacing MATIC at 1:1 :

Expanded role of POL: from “Polygon PoS gas + staking” to “AggLayer multichain gas + restaking + governance”
Fate of Polygon PoS: it continues operating, but its strategic status is demoted to “one ordinary chain 1 inside AggLayer”; in 2024-2026, TVL fell from a peak of $5B+ to the ~$700M-1B range (DefiLlama basis)
Polygon zkEVM Mainnet Beta: since its 2023 launch, TVL has remained below $100M; in 2025 it effectively entered maintenance mode, and Polygon Labs shifted its strategic focus to CDK + AggLayer
POL single-token economic model: MATIC’s “fixed supply” was changed to annual 1% emission (2% allocated to validators / 1% to community treasury, etc.; exact ratios iterate), and the long-term inflation model is close to Cosmos ATOM, ETH, etc.

The business model changes from “selling Polygon PoS gas” to “selling CDK stack + AggLayer protocol-layer fees + POL value capture from staking.”

AggLayer vs Cross-chain Protocol Comparison

Viewpoint	AggLayer	LayerZero V2	Hyperlane	CCIP
Scope	Inside Polygon stack	Multiple ecosystems (EVM + Solana + Aptos, etc.)	Multiple ecosystems	Multiple ecosystems
Trust model	Unified bridge + pessimistic proof	DVN (Decentralized Verifier Networks)	ISM (Interchain Security Modules)	Chainlink DON
Asset model	Unified bridge (canonical)	OFT (native issuance across multiple chains)	warp route	standardized cross-chain token
ZK aggregation	Yes	No	No	No
Target	Rollups that chose the Polygon stack	Any L2 / L1	Any L2 / L1, modular	Institutional + regulated finance

AggLayer’s differentiation lies in “trust-minimized inside the stack + ZK aggregation”, but the cost is that it can only serve Polygon CDK chains. See Cross-chain Five-pole Comparison Matrix for detailed comparison.

Polygon’s Position in RWA / Institutional Scenarios

Polygon PoS RWA: RWA deployed on Polygon PoS in 2024-2026 includes Hamilton Lane / Securitize / Backed Finance / Ondo, but share continues to be taken by Ethereum L1 (starting from BlackRock BUIDL) + Stellar + Avalanche
Polygon ID: Iden3 -based ZK identity, deployed in a limited way in regulated DeFi and KYC scenarios in 2024-2026
Polygon Enterprise / Supernets concept: the enterprise-chain initiative presented in 2022-2023 was largely no longer promoted after 2024; compared with the enterprise sales strategies of Avalanche subnet / Cosmos appchain, Polygon clearly lags in the enterprise direction
vs institutional DLT: compared with Canton and Hyperledger Besu, Polygon did not enter the main battlefield of G-SIB tokenization

AggLayer v0.x Iteration Details

AggLayer is not 1 release, but a continuously iterated stack. The rhythm of major versions in 2024-2026 :

v0.1 (2024-02): Unified Bridge basic infrastructure + first wave of Polygon zkEVM + Astar zkEVM connections; essentially a “shared L1 bridge contract”
v0.2 (2024-Q4): Pessimistic Proof introduced — even if a specific AggLayer chain’s ZK proof system has a bug, that chain cannot withdraw assets from Unified Bridge beyond its deposits; this is AggLayer’s core commitment to “risk isolation inside the stack”
v0.3 (2025): some cross-chain messaging primitives + zkEVM prover performance improvements; actual connected chain count is lower than when v0.1 was announced
v1.0 (roadmap): production deployment of cross-chain atomic execution + ZK proof aggregation
State Sharing: the ability to share storage slots across chains, on the roadmap; competes with Optimism Superchain’s shared sequencer + Arbitrum BoLD route
POL restaking connection: in 2025-2026, Polygon tries to make POL the staking asset for “decentralized verifiers” inside AggLayer; this is similar to the position EigenLayer occupies in the ETH ecosystem, but actual mainnet connection is slow

Polygon zkEVM Prover Route · Plonk + KZG + Hardfork Rhythm

Plonky2 / Plonky3: Polygon Labs’ self-developed STARK + Plonk hybrid prover, iterated in 2023-2024, with a target prover time of < 1 minutes / block
Pessimistic Proof uses Plonk + KZG: depends on Ethereum L1 KZG precompile (EIP-4844), reducing verification cost
Hardfork rhythm: Polygon zkEVM hardforks every 3-6 months (naming style such as Dragonfruit / Etrog / Elderberry / Feijoa), asynchronous with the Ethereum L1 fork rhythm
vs zkSync / Scroll / Linea: zkSync Era uses Boojum (self-developed STARK), Scroll uses self-developed zkEVM + Halo2, Linea uses ConsenSys lattice-based prover, and Polygon does not have an absolute advantage in prover performance or Type-1 compatibility

Competition with Other RaaS / Multichain Protocols

Viewpoint	Polygon CDK / AggLayer	OP Stack / Superchain	Arbitrum Orbit	zkSync ZK Stack	Eclipse / Solana family
Main prover / consensus	ZK (Plonk) + Pessimistic	Optimistic + Fault Proof	Optimistic + BoLD	ZK (Boojum)	SVM + Celestia
Known deployed chain count	5-10 active	50+ (Base / OP / World / Zora / Mode, etc.)	20+ (Xai / Sanko / DeGen, etc.)	5-10	1-3
Flagship chain	(none)	Base	Arbitrum One	zkSync Era	Eclipse Mainnet
Unified Bridge	Yes (AggLayer Unified Bridge)	No (each OP chain independent)	No (each Orbit chain independent)	Hyperchain shared bridge (roadmap)	Single-chain centered
Shared sequencer	(none)	Planned (2026+)	(none)	(none)	(none)
Top RaaS suppliers	Polygon Labs / Caldera	Conduit / Caldera / Alchemy	Conduit / Caldera / Alchemy	Matter Labs	(in-ecosystem self deployment)

Polygon CDK’s AggLayer is the only approach that positions “unified bridge inside the stack” as its core differentiation, but the excessive strength of OP Stack’s network effect is the largest headwind.

Use Cases · Who Is Actually Running on Polygon CDK / AggLayer

Astar zkEVM (2024-03): Japanese ecosystem dApp connection; effectively wound down in 2024-Q4
X Layer (formerly X1, OKX): 2024-Q1 testnet → mainnet, still operating but activity is lower than OKX’s main chain
Immutable zkEVM: Immutable Games proprietary L2, 2023-11 mainnet; game GMV is lower than OP Stack gaming chains
Palm Network: NFT and Web3 IP platform; traffic is stable after CDK migration but small
Wirex Pay: Wirex crypto card / wallet backend
Aavegotchi (Gotchichain): game-specific chain
OKBC / some Gnosis Pay components: actual connection degree is limited
Total: in 2026-Q1 , AggLayer-connected mainnet chains are roughly 5-10 active, far below the 50+ active chains of OP Stack

Sources

Polygon AggLayer official documentation
Polygon CDK documentation & GitHub
Polygon Labs blog — AggLayer v0.1 / v0.2 announcements, POL migration
L2Beat — Polygon zkEVM, Astar zkEVM, Immutable zkEVM, X Layer tracking
DefiLlama — Polygon PoS / zkEVM / Astar TVL history
Messari “State of Polygon 2025”
The Block — Polygon 2.0 coverage 2024-2026