The long-awaited transition of Ethereum to Proof-of-Stake (PoS) was finally completed on September 15, 2022. At that point, the PoW miners who had contributed years of computational power to Ethereum’s development collectively went offline.
According to data from whattomine, the Ethereum network’s hash rate was still 840 Th/s on September 14, with over a million mining machines—most of them GPU-based—operating diligently just before the Merge. Although these machines could no longer mine Ethereum, it didn't mean they had to leave the industry entirely. Beyond selling the equipment directly, miners had several potential paths forward, including:
- Switching to other public chains that support GPU mining, such as ETC
- Participating in other Web3 protocols that require computational power
- Transitioning to become data center operators
- Engaging in ZK mining
Among these, ZK mining—once a relatively niche option—has gradually entered the spotlight.
Understanding ZK Mining
ZK mining refers to the process where machines use computational power to generate zero-knowledge proofs (ZKPs) based on Zero-Knowledge Proof algorithms. In an article titled “Hardware Acceleration for Zero Knowledge Proofs” published by Paradigm this year, the importance of ZK mining was highly emphasized. The article suggested that the scale of ZK mining could eventually rival that of Proof-of-Work mining.
ZK mining finds application in several scenarios, including decentralized storage, Layer 1 (L1) blockchain networks, and Layer 2 (L2) scaling solutions. Below, we explore these scenarios along with representative projects and their underlying principles.
Decentralized Storage: Filecoin
Filecoin announced the introduction of a GPU solution for zero-knowledge proofs as early as its testnet phase in 2019 and began using zk-SNARKs on its mainnet in 2021. Protocol Labs, the team behind Filecoin, even launched a research website called “zk-SNARKs for the World” to document their work on implementing zero-knowledge cryptography. They claim that Filecoin is one of the largest deployed zk-SNARK networks, generating 6–7 million zk-proofs daily, primarily using GPUs for parallel proof generation.
Filecoin’s storage proof mechanism utilizes zk-SNARKs to compress both Proof of Replication (PoRep) and Proof of Spacetime (PoSt). Storage miners run PoRep once when a storage deal is first established to prove they are storing a physically unique copy of the data. Each on-chain PoRep consists of 10 separate SNARK proofs. PoSt is run repeatedly to prove continued storage of the same data over time, with the final stage compressing these proofs into a single zk-SNARK.
Although creating zk-SNARKs for Filecoin is computationally expensive, it effectively reduces costs related to block storage and network bandwidth. Compared to the original proofs, zk-SNARKs are very small, making them efficient to store on the blockchain. For example, a proof that would have taken hundreds of KB on the Filecoin chain can be compressed to just 192 bytes using zk-SNARKs. This minimizes the storage pressure on each node in the Filecoin network, avoids consuming excessive network bandwidth, and allows validators to complete the verification process quickly and cheaply.
In April 2022, the team announced Lurk, a programming language for recursive zk-SNARKs. Lurk can support the development of the Filecoin protocol in multiple ways. For instance, its integration with the Filecoin Virtual Machine (FVM) will enable the network to run smart contracts that can be zero-knowledge verified.
Clearly, zk-SNARKs will play an increasingly important role in the Filecoin network. As stated in their blog: “Zero knowledge has long been part of the Filecoin network, and it will continue to prove vital to the evolution of the network in the future. As more and more data is stored, maintained, and accessed on Filecoin, the ecosystem needs to continue delivering efficient, cost-effective, and secure solutions to verifying. Zero knowledge is a game-changer for the Filecoin network, reducing complex verification processes to fractions of their original size without compromising security, trust, or confidence.”
Layer 1 Blockchain: Aleo
Aleo is a privacy-focused Layer 1 blockchain that supports smart contracts. It separates program execution and state maintenance through zkCloud, an off-chain, trustless execution environment, thereby increasing transaction throughput and meeting blockchain’s needs for privacy and programmability. On-chain validators are responsible for verifying and producing blocks, while off-chain provers handle computation. This design incentivizes validators to stay active by producing blocks and also encourages provers to contribute proving power to the Aleo ecosystem.
Since Aleo transactions are processed off-chain, network nodes do not perform computations but only verify, allowing Aleo to support larger transaction volumes. To some extent, Aleo’s design can be likened to a combination of Ethereum and zkRollup. To apply privacy to any application, Aleo provides the necessary infrastructure and tools, including Zexe (a specific zero-knowledge solution), Leo (a programming language for writing zero-knowledge applications), Aleo Studio (an IDE tool), and the Aleo package manager.
In August 2022, Aleo launched Testnet 3, introducing a hybrid consensus model called AleoBFT that combines instant finality from validators with the computational power of provers. In this model, provers need to solve core components of zero-knowledge proofs—specifically, multi-scalar multiplication (MSM) and fast Fourier transform (FFT)—on the network. They generate Proof of Succinct Work (PoSW) proofs and submit them to PoS stakers in exchange for a portion of each block’s rewards. PoSW is a consensus protocol unique to Aleo; it is a SNARK-based variant of Bitcoin’s PoW algorithm, with the key difference that the underlying computation is not an arbitrary hash function but a zero-knowledge proof. The rewards provers can earn depend on the speed at which they generate proofs: the more zero-knowledge proofs generated per second, the greater the rewards.
As an open platform, Aleo attracted over 10,000 nodes during Testnet 2, with proof generation reaching up to 20,000 per second. In Testnet 3, the project encourages developers to create and open-source faster GPU proving programs, where GPUs are expected to play an even more critical role.
Layer 2 Scaling: Scroll
Scroll is an EVM-equivalent zkRollup designed to scale Ethereum. It supports generating zero-knowledge proofs while maintaining compatibility with the EVM, allowing Ethereum smart contracts to be deployed and run without modification. Scroll’s technical architecture consists of three main components: Scroll Node, Roller Network, and Rollup/Bridge Contracts.
The Rollup/Bridge Contracts are deployed on Ethereum and Scroll to verify zkEVM validity proofs, provide data availability for Scroll transactions, and enable asset movement between Ethereum and Scroll. Rollers in the Roller Network act as provers, generating zkEVM validity proofs and submitting them to the Roller Network. A Coordinator assigns tasks to the Roller Network and relays transaction information and proofs back to the Rollup contract on Ethereum.
In the operational principle of zkRollups, a Sequencer collects, orders, and batches transactions before publishing them to Ethereum. Then, a Prover generates proofs for these transactions. The transactions are only finalized once the proofs are submitted and verified. Currently, major zkRollups like zkSync and StarkWare have centralized Sequencers and Provers controlled by the project teams. However, Scroll differs significantly in that it aims to create a decentralized proof market through its Roller Network, where Rollers can outsource proof generation to miners providing GPU, FPGA, or ASIC machines.
In a competitive market, this decentralized computational market will likely give rise to mining pools similar to those in Bitcoin or Ethereum. Capable developers or pools will optimize various algorithms to improve proof generation efficiency and reduce costs to earn more rewards. Ordinary miners can then join these pools and receive rewards based on their contributed computational power.
Although Scroll is still in the testnet phase, it has already attracted considerable market attention. The Scroll team is also collaborating with the Ethereum Foundation’s Privacy & Scaling Explorations team to advance this field.
The Future of Ethereum Miners and ZK Mining
Let’s briefly analyze the potential paths for Ethereum miners and estimate how much hash rate each option might attract, including the potential scale of ZK mining.
ETC, as the original chain of Ethereum, can seamlessly support ETH mining machines. However, on September 14, ETC’s hash rate was less than 10% of ETH’s. After the Merge, ETC’s hash rate exceeded 300 Th/s, but without a corresponding price increase, mining profitability dropped sharply, and the hash rate began to decline. Even considering other GPU-supported projects and emerging ETH forks, their capacity to absorb new hash rate is limited.
Other projects requiring GPU computational power—such as Livepeer Network (a decentralized video streaming service providing video transcoding), Render Network (a decentralized GPU rendering solution), or Akash (a GPU cloud computing platform)—are growing but have not seen explosive demand for computational power. If a large amount of GPU power from Ethereum miners floods in, it could lead to an oversupply situation.
Among the ZK mining options in decentralized storage, L1 blockchains, and L2 scaling, decentralized storage—specifically Filecoin—is currently the most likely to absorb graphics cards and other machines. However, Filecoin mining requires not only GPUs but also storage servers and other hardware. For Ethereum miners, switching to Filecoin mining would require additional investment in other machines and architectural designs.
Although ZK mining in L1 blockchains and L2 scaling has enormous potential, related projects are still in the testing phase, and many are not open to GPU miners, meaning proof generation is primarily handled by project-controlled servers. Among projects open to GPUs, Aleo is still in its early stages, and rewards from testnet mining cannot yet be transferred—making it more akin to a form of early investment. Scroll is also in the testnet phase and has not yet launched incentivized test mining activities.
Therefore, in the immediate aftermath of the Merge, most Ethereum mining machines may remain idle or be sold directly. According to tracking by Tom’s Hardware, GPU prices have been declining over the past six months, reflecting selling pressure from Ethereum miners on the secondary market. Nvidia’s CFO recently stated that graphics card prices have almost returned to normal levels, marking the end of the hot market driven by Ethereum mining.
However, as ZK technology develops and more ZK-related projects launch, the computational power required for ZK proofs is expected to grow exponentially. Vitalik Buterin mentioned in his blog, “Type 1 ZK-EVM is what we ultimately need to make Ethereum Layer 1 itself more scalable.” Such a fully Ethereum-equivalent ZK-EVM would undoubtedly require substantial computational power to generate zero-knowledge proofs. Perhaps one day, Ethereum mining machines will find their way back to mining on Ethereum—only this time, they will be computing zero-knowledge proofs instead of random numbers.
👉 Explore advanced mining strategies
Frequently Asked Questions
What is ZK mining?
ZK mining involves using computational resources to generate zero-knowledge proofs (ZKPs), which are cryptographic methods that allow one party to prove to another that a statement is true without revealing any additional information.
How does ZK mining differ from traditional mining?
Traditional mining, like Proof-of-Work, involves solving complex mathematical puzzles to validate transactions and create new blocks. ZK mining focuses on generating zero-knowledge proofs to verify computations off-chain, which can enhance privacy and scalability.
Which projects currently utilize ZK mining?
Prominent projects include Filecoin for decentralized storage, Aleo for privacy-focused Layer 1 blockchain, and Scroll for Layer 2 scaling on Ethereum. Each leverages ZK proofs to improve efficiency, security, or privacy.
Can existing Ethereum GPU miners switch to ZK mining?
Yes, but it may require additional hardware or software adjustments. For example, Filecoin mining needs storage servers alongside GPUs, while Aleo and Scroll are still in test phases and may have specific requirements.
What are the potential rewards for ZK miners?
Rewards vary by project and network conditions. In Aleo, provers earn based on proof generation speed. Filecoin miners earn block rewards and transaction fees for providing storage and generating proofs.
Is ZK mining profitable?
Profitability depends on factors like hardware efficiency, electricity costs, and network demand. While still emerging, ZK mining could become lucrative as adoption of zero-knowledge technology grows.