Polygon zkEVM Mainnet Beta Has Stopped Block Production on July 3
Polygon has confirmed that the Polygon zkEVM mainnet Beta sequencer sunsetted on July 3, 2026, and the network will no longer produce blocks.
Users can no longer bridge assets from zkEVM through the Agglayer; assets held in self-custody wallets (EOA) can be claimed on Ethereum via a dedicated Claims interface, which will open no earlier than July 13 and will remain available until December 31, 2027.
This move marks the official end of the zkEVM project, which Polygon acquired for $250 million, following low adoption rates and operational losses that led to an early halt in development, prompting the migration of remaining assets to the Ethereum mainnet.
Source: Public Information
ABAB AI Insight
Polygon Labs announced the sunset plan for zkEVM mainnet Beta in 2025, with the sequencer actually shutting down in July 2026. The CEO previously acknowledged that the team had stopped related work a year and a half early to focus on other scaling solutions.
In terms of capital and strategic direction, shutting down a low adoption, high-cost chain releases resources to shift towards Agglayer and payment infrastructure, motivated by loss mitigation and concentrating investment in higher-priority interoperability and practical application scenarios.
Similar to other L2 projects adjusting or merging due to underutilization, and the trend of Ethereum's ecosystem moving towards unified liquidity, we are currently in a phase of "multi-chain experimentation converging towards a core path."
This essentially represents a restructuring of the industry chain: when a single zkEVM cannot form network effects and revenue loops, resources shift from parallel chain construction to shared security and cross-chain layers, with pricing power moving from independent L2s to aggregation layers and Ethereum itself.
ABAB News · Law of Cognition
Chains with low adoption will eventually sunset.
Loss mitigation can release resources more effectively than persistence.
Aggregation layers are consuming parallel experiments.