Starknet, an Ethereum layer‑2 network that uses zero‑knowledge (ZK) rollups, is experiencing fresh mainnet disruption as the project enters 2026.
In an X post, the Starknet team said the network was facing downtime and that engineers were “actively investigating the issue and working to restore full functionality as quickly as possible,” without immediately disclosing a root cause. At the time of writing, the network had been experiencing downtime for just over two hours.
Starknet is a ZK‑rollup–based layer 2 that batches transactions offchain and posts cryptographic proofs to Ethereum, aiming to deliver higher throughput and lower fees for smart contracts, decentralized finance and gaming applications while inheriting Ethereum’s base‑layer security.
The project has also promoted a Bitcoin DeFi, or BTCFi, arc, pitching itself as infrastructure for bringing Bitcoin‑related financial applications into the Ethereum ecosystem. Despite the network disruption, the STRK token price held steady at the time of writing.
Related: Ethereum’s first ZK-rollup, ZKsync Lite, to be retired in 2026
Not the first time Starknet mainnet is down
The incident follows a series of outages in 2025 that have put Starknet’s reliability under closer scrutiny. In September, a major upgrade known as Grinta (v0.14.0) led to an extended mainnet disruption in which block production was halted and two chain reorganizations were required, reverting around an hour of activity and forcing users to resubmit affected transactions.
That episode followed an earlier multi‑hour outage in 2025 tied to sequencer issues, with external trackers logging multiple incidents of slow or halted block creation across the year.
A Starknet incident report on the September event said the Grinta‑related downtime lasted roughly nine hours and traced the problems to a sequence of issues, including failures in Ethereum RPC providers and bugs that affected sequencer behavior, prompting the team to commit to architectural changes and expanded monitoring.
A representative from Starknet told Cointelegraph that the team was working to repair the fresh incident.
Read the full article here