After a year of hard work, the zkEVM ecosystem has achieved a performance leap, with Ethereum block proof time dropping from 16 minutes to 16 seconds and costs decreasing by 45 times. Participating zkVMs can now complete 99% of mainnet block proofs within 10 seconds on target hardware.
On December 18, the Ethereum Foundation (EF) announced the implementation of the real-time proof mechanism and the removal of performance bottlenecks, but emphasized that the real challenge lies in reliability. Speed without security support will become a burden, and several mathematical conjectures relied upon by zkEVM based on the STARK algorithm have been disproven in recent months, leading to a decrease in security level.
EF previously set a comprehensive goal for real-time proof in July, covering multiple dimensions such as latency, hardware, and security, and has now passed the EthProofs benchmark test.
Its core shift lies in moving from pursuing throughput to provable security. It is clear that L1-level zkEVM must meet the 128-bit security standard and align with mainstream encryption specifications. Because forging proofs may lead to fatal risks such as token forgery and L1 state tampering, the security margin is non-negotiable.
EF simultaneously released a three-phase security roadmap:
- By the end of February 2026, all zkEVM teams must integrate their proof systems with EF's soundcalc security assessment tool to unify security calculation standards.
- The Glamsterdam standard will be achieved by the end of May, realizing transitional goals such as 100-bit provable security.
- The final goal of H-star is to be completed by the end of December, achieving 128-bit provable security, and providing a formal security demonstration of the recursive topology.
To achieve its goal, EF mentions core technology tools such as WHIR and JaggedPCS, which can improve efficiency by optimizing proof generation and avoiding wasted computing power, while also reducing the size of proofs by using techniques such as recursive topology.
However, multiple challenges remain: real-time proofs have not yet been implemented on the blockchain, and the actual effectiveness of validators is questionable; security parameters need to be dynamically adjusted as mathematical conjectures are disproven; it is unknown whether some teams can meet the targets on schedule; formal verification projects with recursive architectures are still in their early stages, and the ecosystem is developing unevenly.
It is worth noting that once the zkEVM meets the requirements, it can support Ethereum in increasing gas limits, thereby increasing block capacity while ensuring the feasibility of staking, promoting L1 to become a trusted settlement layer, and blurring the boundaries between L2 and L1 execution.
Now that the performance sprint is over, the core issue of the zkEVM ecosystem has shifted to achieving sufficiently reliable security proofs to support assets worth hundreds of billions without relying on fragility conjectures. The security race has officially begun and will be the main theme of Ethereum in 2026.
