How Nethermind is building the verifiable foundation for autonomous systems across Ethereum.

‍

At Devconnect 2025, Nethermind brings the Agentic Economy from theory to reality. Across Ethereum’s open ecosystem, our teams are showing how verifiable systems spanning infrastructure, security, identity, and AI form the foundation for this next era of autonomy.

‍

The next chapter of Ethereum is emerging where blockchain and artificial intelligence meet. Nethermind is building the engineering layers that allow autonomous agents to act, transact, and coordinate with cryptographic trust.

‍

The Agentic Economy envisions intelligent systems that do more than process data. They make decisions, execute onchain, and collaborate across decentralized networks. For this to work, agents need a verifiable foundation built on identity, secure execution, and transparent logic. That is the infrastructure Nethermind develops.

‍

Automation without verification is fragile. Even the most advanced AI systems depend on transparency and accountability. Blockchain provides the consensus layer and cryptographic guarantees for safe execution. Cryptography binds those guarantees to identity. Formal verification ensures that every autonomous action can be traced and audited. Nethermind’s work sits at the center of this intersection, developing infrastructure, security, and AI systems that operate with verifiable intent.

‍

‍

‍

Human identity in a trustless world

‍

At the base of agentic systems lies identity. Nethermind’s hardware initiative establishes a privacy-preserving link between humans and their digital agents. Through cryptographic credentials and on-chain attestations, we enable secure access to wallets, agents, and decentralized applications while protecting sensitive biometric data. It ensures authenticity without sacrificing privacy.

‍

As agents become extensions of human decision-making, verifiable identity becomes essential. Authenticity and accountability must remain anchored in human legitimacy, not data exposure.

‍

‍

‍

Scaling the rails of autonomy

‍

For autonomous systems to coordinate across Ethereum, scalability is critical. Nethermind’s engineering contributions to the Ethereum execution client directly address this requirement.

The Gas Benchmarks framework, developed initially by Nethermind, has become the standard used by Ethereum client teams and researchers to measure execution performance and validate safe gas limit increases. These benchmarks revealed key bottlenecks, guided optimizations across clients, and demonstrated readiness for agent-scale transaction volumes.

‍

Working with the Ethereum Foundation, we also operate specialized test networks that push mainnet parameters to their limits, exploring how state growth and computation interact under future conditions. This data-driven approach ensures that Ethereum’s infrastructure can support the computational intensity of agentic systems while preserving decentralization.

‍

‍

‍

From infrastructure to interaction

‍

The Agentic Economy extends beyond core infrastructure to how agents connect with real-world processes. Nethermind develops frameworks for decentralized finance, tokenization, and risk management that link blockchain logic with existing operational systems. These frameworks enable automated settlement, insurance, and compliance in transparent, auditable formats.

‍

Automation becomes not just efficient but accountable, translating onchain verifiability into off-chain reliability.

‍

‍

‍

Verifiable intelligence

‍

Nethermind’s AI work demonstrates what verifiable intelligence looks like in practice.

‍

ResultAI brings the idea of autonomous digital coworkers to life. Built from the automation of over twenty internal processes, it integrates directly with tools such as Slack, Notion, and GitHub. Each virtual coworker operates with its own identity, responsibilities, and measurable cost and performance.

‍

AuditAgent applies reasoning models to analyze Solidity code, finding vulnerabilities at speed and scale. AgentArena extends this through competitive collaboration, where independent agents contribute findings, broadening security coverage through collective reasoning.

‍

All of these systems connect through ChaosChain, a protocol for agentic accountability. Built around standards such as ERC-8004, A2A, and x402, ChaosChain introduces Proof of Agency, a verifiable record of agent activity. It enables causal auditing and collaboration through a decentralized knowledge graph. These systems transform AI from an opaque model into a transparent participant in the on-chain ecosystem.

‍

‍

‍

Securing autonomy

‍

As agents act more independently, their threat surface expands. They reason, adapt, and interact across networks and APIs. Traditional cybersecurity frameworks are not equipped for risks such as adversarial prompts, reward hacking, or memory poisoning.

‍

Nethermind Security extends formal verification and adversarial research into this domain. We verify smart contract agents, audit reasoning loops, and study the alignment between agent logic and policy. The goal is to secure not only the code that executes but the decision processes behind it. By embedding verification directly into system design, we ensure that increasing autonomy never means rising uncertainty.

‍

‍

‍

Research as infrastructure

‍

Every layer of this work is grounded in research that advances the boundaries of verifiable computation and secure coordination.

‍

Nethermind produced the first formal proof of an onchain zk-verifier for ZKsync using EasyCrypt and created Halva, a Lean-based framework for verifying Halo2 circuits that uncovered a critical bug in Scroll’s Keccak-256 implementation. Our Ethereum Foundation-funded research extends formal verification to Plonky3 circuits.

‍

In cryptographic theory, Nethermind published simplified security proofs for the FRI protocol, strengthening one of the core components of modern SNARK systems. At the protocol layer, our engineers authored six EIPs for Ethereum’s Fusaka upgrade, including EIP-7883 for ModExp gas repricing, derived directly from Gas Benchmark data. Our benchmarking methods have since been integrated into the Ethereum Execution Specs (EELS), becoming a shared reference for evaluating scalability across clients.

‍

This continuous research cycle ensures that every system we build, from clients to AI frameworks, rests on formal, verifiable foundations.

‍

‍

‍

An ecosystem in motion

‍

The Agentic Economy is not a distant idea. It is already taking shape across Ethereum. Builders, researchers, and networks are experimenting with agents that can reason, verify, and coordinate autonomously.

‍

At Devconnect, Nethermind is uniting these threads into one narrative. Our infrastructure powers Ethereum’s execution, our security teams formalize its trust, and our research drives the standards that guide its evolution. Together, they form the engineering base on which verifiable autonomy can run.

‍

The future of coordination is already here. It is transparent, cryptographic, and alive on Ethereum.