zerokn0wledge

Polygon zkEVM tech deep dive: Proof of Efficiency and plonky2

Thread1,092 words
Market since$MATIC−83%since this pieceHow it played out ↓

Technical walkthrough of Polygon zkEVM: Proof of Efficiency with sequencers and aggregators, bytecode-level EVM equivalence versus zkSync's language-level approach, and the recursive plonky2 proof system. Condensed from a 60 page report co-written with Louround.

Key claims
Polygon zkEVM is much closer to type 1 EVM equivalence than zkSync
plonky2 combines PLONK SNARKs with FRI and removes the trusted setup
Proof of Efficiency splits block production between sequencers and aggregators
0/ in July 2022 @zkSync, @0xPolygon & @Scroll_ZKP kicked off the a battle for zkEVM supremacy a new era of scaling begins & Polygon #zkEVM mainnet is just 3 days away a 🧵 providing a technology deep dive into @0xPolygon zkEVM based on 60 pages of pure alpha by me & @Louround_ 1/ together with @Louround_, I went down the zk rabbit hole. We came back with a 60 page research report on @0xPolygon’s zkEVM & the broader #zkEVM space We decided to share it with you guys for free because we love CT 🫶 2/ if you want to follow us down the rabbit hole, find the full report below if you're looking for an overview thread that also covers the non-tech part, check out @Louround_'s latest post louround.substack.com/p/zkevm 3/ but let's dive into the technology @0xPolygon #zkEVM is a bytecode-level #EVM equivalent zk-rollup, using a novel Proof-of-Efficiency (PoE) consensus mechanism to improve sequencer decentralization and built on a cutting-edge recursive proof system (plonky2) 4/ PoE employs a two-step model involving two types of participants: - Sequencers - Aggregators 5/ sequencers create batch proposals and submit them as L1 transactions They pay the gas fees for L1 transactions and an additional fee in $MATIC tokens for each proposed batch, which incentivizes sequencers to propose valid batches Batch fees vary depending on network load 6/ aggregators monitor proposed batches and initiate proof generation (via zkProver) The first valid proof that updates the #L2 state is accepted by the PoE smart contract, making the process highly competitive The design promotes decentralization and resilience against attacks 7/ the transaction validation and verification process in @0xPolygon #zkEVM is managed by a centralized zkProver component. The zkProver enforces the necessary conditions for a transaction to be valid through intricate mathematical calculations and generates a #plonky2 proof 8/ with regards to EVM compatibility, @0xPolygon zkEVM adopts an ambitious bytecode-level approach, eliminating the need for transpiler steps, retaining compatibility with the #EVM without the need to convert Solidity code into a separate language 9/ within Vitalik's EVM equivalence framework, the @0xPolygon #zkEVM is definitely among the ones with the highest degrees of EVM equivalence 10/ however, @0xPolygon's zkEVM uses a new set of assembly codes to express each opcode, introducing a minor intermediary step to the execution 11/ it is much closer to type 1 equivalence than competitor @zkSync though, that transpiles Solidity code into a language optimized for #zk proofs called Yul behind the scenes and can hence be referred to as a language-level compatible zkEVM 12/ but why does EVM equivalence matter you ask? When a development environment is #EVM equivalent, that is a big advantage for developers, as it means that basically any smart contract or developer tool from #Ethereum can be used on the EVM equivalent network too. 13/ it also simplifies migration for Ethereum based dApps and enables easier interoperability with other EVM dApps on L1 or other zkEVMs. Moreover, EVM-equivalence (Solidity & support for Ethereum dev tools) enables rollups to tap into the largest dev pool in Web3 14/ but what does it actually mean to be EVM equivalent? well, reaching EVM equivalence means reaching full bytecode-level compatibility (like @0xPolygon). This differs from mere compatibility, where devs are often forced to modify, or even reimplement certain lower-level code 15/ but now that we have an understanding of the network architecture, participants and VM, let's have a closer look at the #plonky2 proof system 16/ what are zero knowledge proofs (ZKPs)? #ZKPs are cryptographic techniques allowing one party to prove a statement's truth to another without revealing any information beyond the statement's validity 17/ In the crypto world, we care about non-interactive #ZKPs (NIZKPs), which don't require interaction between the verifier and prover. They rely on a common reference string (CRS), which is generated in a trusted setup ceremony that is publicly verifiable 18/ the CRS serves as the basis for the NIZKPs and allows for the verification of statements without any interaction. However, the reliance on a CRS for NIZKPs raises security concerns, as the CRS must be generated in a trusted setup ceremony to ensure system integrity 19/ but WTF are #SNARKs? SNARKs a.k.a. Succinct Non-interactive ARguments of Knowledge are a specific form of #ZKP and a method of verifying a statement's validity without revealing any information about the statement 20/ and what about #STARKs? STARKs or Scalable Transparent ARguments of Knowledge are another form of ZKP and offer benefits like eliminating the need for a trusted setup (like certain advanced SNARK systems), increased performance and transparency 21/ however, both #SNARKs and #STARKs come with limitations and trade-offs. While the former requires a trusted setup in many cases, it is primarily the computational overhead required to generate & verify proofs that can limit the use of STARKs. 22/ the #plonky2 proof system that Polygon #zkEVM uses is based on PLONK and combines the strengths of STARKs and SNARKs 23/ #PLONK, or "Permutations over Lagrange-bases for Oecumenical Noninteractive arguments of Knowledge," was introduced in 2019 and has seen various improvements, including fflonk, turbo PLONK, ultra PLONK, plonkup, & more recently, plonky2 24/ while not removing the need for a trusted setup entirely, #PLONK introduced a universal & updatable trusted setup however, with rapid innovation, more advanced extensions like plonky2 have even removed the need for a trusted setup. A significant leap for #zk proof systems 25/ so what is plonky2? it's a recursive zk proof system integrating PLONK-based SNARK proofs & FRI (polynomial commitment scheme used in STARKs), combining the best of STARKs with the benefits of SNARKs into whats likely the cheapest & most performant proving system on Ethereum 26/ a very important feature of #plonky2 is recursiveness This means that the protocol is scalable and capable of proving arbitrarily complex facts, making it a "general purpose" protocol that can be used for a wide range of zero knowledge applications (#zkApps) 27/ an important benefit that comes with that, is that proofs can verify other proofs, allowing for recursive rollup structures (e.g. application-specific L3s) on top of the generalized #L2 (i.e. the @0xPolygon zkEVM) 28/ recursive #L3 rollup structures are a critical part of how Polygon aims to add throughput to the #zkEVM. @0xPolygon aims to provide infrastructure and there are already multiple projects working on L3 implementations 29/ so in conclusion, @0xPolygon's zkEVM undoubtedly has some of the best tech out there and definitely is an ecosystem to watch in the coming months 30/ if you've made it this far, don't forget to check out the full report here: louround.substack.com/p/zkevm
First image of the post

8 more images on X

Open on X

How the covered assets moved · since 24 Mar 2023

Prices as of 23 Oct 2026

This explainer walks through how things work without making a call on the token, so price moves since publication are shown as market context.

Covered in this piece
$MATICMATIC (migrated to POL)to last trade 17 Oct 2025−83%−96% vs BTC
Since then

Zero has written 2 more pieces on $MATIC since this one.

Daily closes in USD from DefiLlama and Yahoo Finance. Only a bullish or bearish thesis is read as a call; explainers, maps and mentions are context. Not financial advice.

zerokn0wledge avatar
Working together

Research, advisory and builds for deep-tech teams.