By Jonathan King

TL;DR

• Zero-Knowledge Proofs (ZKPs) and the resulting tech are a breakthrough area of cryptography that is largely seen as the endgame for blockchain design concepts.

• Today, ZKPs are increasingly becoming a promising solution set for unsolved problems in web3, including 1) blockchain scalability, 2) privacy-preserving applications, and3)  trustless interoperability.

• In 2023, just over ~$400M was invested into ZK tech, with a predominant focus on scalability of Ethereum L1/L2 protocol layers and emerging infrastructure and developer tools

• The ZK landscape can be divided into 3 layers: 1) Infrastructure (i.e., tooling/hardware  for building protocols/applications on top of ZK primitives), 2) Networks (i.e., L1/L2 protocols that leverage ZK proving systems), and 3) Applications (i.e., end-user products that leverage ZK mechanisms).

• While still in its early stages, the rapid development of the ZK ecosystem promises to usher in a new era of secure, private, and scalable blockchain solutions.

Zero-Knowledge Proofs (ZKPs) and the resulting tech have largely been seen as the endgame for blockchain design, especially as it relates to offering a solution for onchain applications to verify information with minimal trust assumptions. At its core, a Zero-Knowledge Proof is a cryptographic technique that allows one party, known as the prover, to demonstrate to another party, the verifier, that a computation is valid without exposing any of the underlying data used in creating the computation. Originating in 1985, ZKPs have evolved from theory to practical utility, overcoming decades of lag through recent advancements in software tooling and hardware. 

Today, ZKPs offer promising solutions for web3’s biggest challenges, including: 

1. Blockchain scalability: one of the biggest challenges facing Ethereum L1 is scalability. However, L2 networks have emerged to enable faster and cheaper transactions without compromising Ethereum’s security or decentralization. While optimistic rollups remain dominant given a higher degree of EVM compatibility and developer friendliness, ZK rollup adoption is steadily increasing. ZKPs help to summarize complex computations offchain, thus enhancing L2 designs for rapid and cost-effective onchain verification and settlement.

2. Privacy-preserving applications: to date, work on privacy in the blockchain context has been mostly limited to obfuscating transactions. However, researchers are progressively working towards enabling full transactional anonymity and confidentiality on public blockchains. Importantly , novel privacy-preserving concepts leveraging ZKPs  are emerging that aim to break perceived trade-offs between preserving user privacy and enabling compliance (i.e., deterring illicit activity). 

3. Trustless interoperability: existing blockchain interoperability protocols rely on trusted systems (e.g., multisigs or incentivized validator sets).  ZKPs can help replace crypto-economic trust assumptions with cryptographic guarantees, opening avenues for more secure and robust cross-chain communication. However, among the primary applications of ZKPs, interoperability is the most nascent. 

According to Messari’s deal screener, over ~$400M was invested into the ZK landscape in 2023, emphasizing scalability of Ethereum L1/L2 layers and emerging ZK developer infrastructure. Despite the relative nascency of ZK, its rapidly accelerating ecosystem foresees convergence on best practices for more secure, private, and scalable blockchain applications. With this framing in mind, let’s take a closer look at the ZK landscape layer-by-layer to explore key players and emerging concepts.

Infrastructure

Any form of a ZKP must be written in arithmetic circuit language, which has limited expressions, and it’s complex to translate most blockchain functions to circuit form. Limitations in developer tooling and advanced hardware meant real-world use cases of ZK were slow until recently. Today, we are seeing an array of systems and tools emerge that empower developers to build protocols and applications on top of ZK cryptographic primitives.

Programming Frameworks & Tools: Domain Specific Languages (DSLs), such as Leo, Noir, Cairo,  and o1js are programming frameworks for developing ZK-provable programs within specific L1/L2 ecosystems (e.g., Aleo*, Aztec*, Starkware*, and Mina* respectively). Additionally, generalizable frameworks such as Elusiv* and Hinkal, are emerging with the aim of allowing developers to define specific criteria for how transaction data can be shielded onchain but verified using ZKPs. Growing adoption is expected, meeting potential developer and end-user demand for ZK-powered applications.

ZK Coprocessors: ZK coprocessors provide developers with cost-effective and trustless offchain computing capabilities, while eliminating the need for developers to engage with complex ZK-related components in their tech stack. Teams like RiscZero, Axiom, and Herodotus provide verifiable computing platforms that generate a proof attesting to the execution and validity of arbitrary programs or enable smart contracts to store, access, and verify historical onchain data without imposing additional trust assumptions. Over time, ZK coprocessors are poised to become essential for increasingly advanced onchain applications.

Proof Networks / Markets: Today, the majority of ZK networks and protocols rely on a centralized proving process. Subject to ZK adoption growing over time, we expect that teams will seek to decentralize their proving layer to improve their liveness and censorship resistance. Emerging proof networks and markets, such as those provided by =nil; Foundation, RiscZero, Gevulot, and Lumoz, aim to allow applications to outsource their proving mechanisms to third-party operators, thereby lowering the overhead for operating ZKP infrastructure.

Hardware Acceleration: ZKPs are expensive and computationally-intensive to produce given the large number of mathematical operations required. However, we are seeing significant advancement in the usage of specialized hardware like Field Programmable Gate Arrays (FPGAs) and Application Specific Integrated Circuits (ASICs), which are helping to improve proof generation and verification times. Specialized hardware providers like Ingonyama, Cysic, and Fabric, are at the frontier of providing FPGAs and ASICs for ZK proof systems and we expect to see increasing innovation and investment in ZK hardware design going forward.

App-chain Infrastructure: Rollup-as-a-service (RaaS) providers like Spire, ProtoKit, and Lumoz provide low-code tooling for developers to build, test, and deploy general-purpose or app-specific L2/L3 chains that leverage ZK proving mechanisms. Sequencers, such as Espresso*, Radius, and Madara, provide infrastructure for accepting transactions from users, determining their order, and posting blocks to the L1 consensus and data availability layers. We believe the next-generation of Ethereum scalability is poised to be powered by modular L2 rollup stacks, which may create demand for these providers in the short-to-medium term.

Interoperability & Bridging: Bridging systems become more trust minimized as they remove the need for users to rely on humans (e.g., multisigs or incentivized validator sets) and replace trust with code (e.g., Light Clients, Relays and ZKPs). Teams such as Polyhedra, Lambda Class, and Polymer Labs* are exploring this topic. Among the primary applications for ZKPs, interoperability is the most nascent, but we expect to see more innovation in bridging design concepts as access to ZK primitives accelerates.

ZK Machine Learning (ZKML): ZKML, a frontier field of cryptography, focuses on proving the correctness of onchain machine learning (ML) model inferences using ZKPs. By adding ML capabilities, smart contracts can be made more autonomous and dynamic, allowing them to make decisions based on real-time onchain data and adaptable to various scenarios, including those that may not have been anticipated when the contract was initially created. Teams like Modulus Labs, Giza, Zama are pioneering unique ZKML use cases, which may offer a promising synergistic balance at the intersection of AI and crypto.

Networks

Some blockchains face limitations in processing high transaction volumes, leading to slower transaction times and increased costs during peak demand. Additionally, popular blockchains, such as Bitcoin, Ethereum, and Solana are built on open, public ledgers, but the lack of privacy raises concerns for mainstream participants that are likely to require full transactional confidentiality and anonymity. New L1 and L2 networks are emerging with ZK proving infrastructure to  solve the problems related to blockchain scalability and  onchain privacy.

Privacy-focused L1s: Emerging L1 networks like Aleo, Mina, and IronFish offer privacy-first smart contract capabilities powered by ZKPs, providing application-level privacy for dapps within their respective ecosystems. L1 networks such as Fhenix and Inco employ fully homomorphic encryption (FHE) to make it possible for developers to write private smart contracts and perform computations on top of encrypted data, thus enabling complete transactional anonymity and confidentiality. Given that many of the L1s above are undergoing incentivized testnets and require developers to learn new  programming languages, signs of mass adoption and value capture are likely 1-2 years out. 

ZK-EVMs: ZK-EVMs leverage zero-knowledge proofs to make cryptographic proofs of execution of Ethereum-like transactions. There are different types of ZK-EVMs such as zkSync Era*, Polygon zkEVM*, Linea, Scroll, and Taiko, that each have varying design tradeoffs between EVM compatibility and  performance (i.e., proof production times). We expect ongoing innovation in this segment for scaling Ethereum and Ethereum-based ZK-rollups.

ZK-Rollups: A zero-knowledge rollup is a L2 scaling solution that moves computation offchain and proves state changes onchain using ZKPs. ZK-rollups like Aztec* provide a “privacy engine on top of Ethereum'', which aims to encrypt transactional data while ensuring costs remain low. Zeko is an upcoming ZK-rollup stack built on top of Mina that enables apps to recursively verify and compose with each other, while ImmutableX* and LayerN are app-specific ZK-rollups for gaming and high-performant DeFi use cases, respectively.  While optimistic-based rollups command roughly ~90% of the total L2 market share, ZK-rollups are poised for increased demand as the underlying tech becomes more accessible.

Applications

Atop the ZK infrastructure and network layers sits an emerging crop of end-user applications that leverage ZKPs for onchain payments, identity, private yet compliant DeFi, and consumer use cases. 

Teams like Elusiv*,  provide user-friendly interfaces for private payments and DeFi transactions through shielded addresses, while also employing compliance mechanisms to decrypt transactions by identified illicit actors. In identity, zCloak*, ZKPass, and zkp-ID employ ZKPs to allow users to prove verifiable data to third parties without exposing personal information. 

DeFi protocols like Lumina and Panther focus on building private yet compliant decentralized exchanges. Renegade employs both multi-party computation (MPC) and ZK to offer dark pool trading, an onchain trading venue that conceals the orderbook and allows large institutional or whale traders to execute orders without alerting the wider market to their activity. 

Consumer apps like Sealcaster and Dark Forest utilize ZKPs in social and gaming apps, shielding user identities and gaming strategies from other onchain participants.

The future of ZK

The future of ZK involves novel zero-knowledge proof designs prioritizing speed, reduced hardware requirements, improved developer tooling, and support for decentralized proof generation. While both Optimistic and ZK scaling solutions serve to verify rollup transactions, each with associated design tradeoffs between security, latency, and computational efficiency, we see a convergence of the two stacks, in the medium-to-long term, to accommodate a versatile range of onchain applications.  Lastly, the ZK app layer is nascent today, but likely poised for growth as end-user demand for privacy-preservation on public blockchains grows over time. Additionally, it’s worth noting that ZK research is primarily explored in the Ethereum context. However, emerging concepts such as Solana’s Token22 program with Confidential Transfers (i.e., a privacy feature that utilizes ZKPs to encrypt token balances and transfer amounts for SPL tokens), showcase the adaptability and potential of ZK beyond specific ecosystems.

In conclusion, the transformative potential of ZK is unfolding, promising a future marked by heightened security, privacy, and scalability in blockchain solutions. Within the ZK landscape, Coinbase Ventures is investing in emerging ZK developer infrastructure (e.g., coprocessors, proof markets, app-chain infra)  and applications (e.g., private payments and DeFi) that unlock new forms of onchain utility and led by teams with top ZK cryptography talent (a rare/small talent pool). If you’re building in these areas, we would love to hear from you - JK’s DMs are open!

— 

Disclosures and footnotes

*The following Coinbase Ventures portfolio companies appear in the above landscape: Aleo, Anoma, Aztec, Consensys, Espresso, Elusiv, Mina, Polygon, Polymer Labs, Starkware, Sunscreen, zCloak, zkLink, zkSync