Nethermind

CMTAT-FIX is on the program at CMTA's Tokenization Seminar 2026 in Zurich on 4 June. The seminar pulls together finance, tech, and legal people working on the standards behind tokenised securities and digital financial assets. CMTAT-FIX puts FIX protocol descriptors onchain for tokenised securities issued under CMTAT. Co-designed by Nethermind with CMTA and Taurus SA. Second Nethermind/CMTA collaboration. The first was security tooling for the CMTAT codebase. This one takes the standard into capital markets infrastructure. Tomasz Kurowski (Head of Enterprise Business) and Swapnil Raj (Head of Innovation) will present.

Your team takes the break. We take the audit. Schedule the engagement for when people are off. Come back to findings, not to a backlog. A few slots open between July 1 and August 15. Adjusted rates. AuditAgent Pro included so the codebase is ready before we start. https://lnkd.in/djb85xwD

Cody Born, Principal Engineer at Uniswap, on bringing AuditAgent into their development cycle. Uniswap ran an evaluation across three months. A free AuditAgent scan first. Then a full AgentArena competition on UniswapX, where multiple AI security agents analyzed the contracts in parallel, an AI arbiter judged the findings, and human auditors validated them. The competition surfaced issues in core settlement logic, token transfer handling, Permit2 trust boundaries, and DCA execution design. They adopted the AuditAgent Business Plan and added it to CI. Two findings surfaced in development, one High severity. Both addressed before formal audit. AI in this engagement runs continuously during development and at adversarial milestones. Comprehensive audits and formal verification remain the deeper layers, with broader scope and mathematical guarantees where they're needed. Code that's been continuously analyzed reaches formal audit in better shape, freeing senior auditors and formal methods to focus on the harder problems.

Audit quality is bounded by what auditors can spend their time on. In a noisy codebase, that time goes to setup, dead code, missing documentation, and integration bugs that should have surfaced on testnet. Time spent on basics is time not spent on the bugs only auditors can find. Audit prep is about handing the audit team the largest possible window for the work that only they can do. Running AuditAgent before scoping clears common vulnerability patterns and dead code. Validating on testnet catches integration issues. Both let the comprehensive audit go deeper from day one. Audit slots are open through August 15 at reduced rates, with AuditAgent Pro included: nethermind.io/audit-slots

Your team takes the break. We take the audit. Schedule the engagement for when people are off. Come back to findings, not to a backlog. A few slots open between July 1 and August 15. Adjusted rates. AuditAgent Pro included so the codebase is ready before we start. https://lnkd.in/djb85xwD

Looking forward to joining the CMTA Tokenization Seminar 2026 in Zurich on 4 June. I will be presenting alongside Swapnil Raj, Nethermind’s Head of Innovation, on the FIX × Ethereum CMTAT standard connector, a solution designed and developed with CMTA and Taurus SA to connect traditional capital markets workflows with Ethereum-based tokenization infrastructure. This work is part of Nethermind’s broader focus on compliant blockchain infrastructure for regulated financial institutions. Looking forward to engaging with clients, partners, and Swiss financial market leaders on the next phase of tokenized capital markets infrastructure. https://luma.com/ko65zck3 #tokenization #onchain #standards #fix #blockchain #digitalassets

Arbitrum One's second execution client is now on mainnet. We built it for the Arbitrum Foundation, and 0.2.0 just shipped the validator side. Building a production execution client for an L2 is specific work. Each rollup has its own consensus, fraud-proof system, and protocol semantics. The execution layer has to interoperate with all of them, while remaining bit-equivalent to the reference for fraud proofs to function. None of it transfers from one L2 to the next. Programmable markets need verifiable infrastructure that doesn't depend on a single implementation. A rollup running on one execution client is one critical bug from a halt. Doesn't matter how good that client is. Arbitrum now has a second independent implementation, and a stronger path to permissionless validation.

Wholesale settlement is moving to public blockchain because that's where tokenized assets live. But regulators expect things public networks weren't built to provide: compliance enforcement, finality you can rely on, audit trails, operational resilience. The Bank of England and BIS Innovation Hub London Centre took this on in their DLT Innovation Challenge. The final report came out a few days ago, and it asked whether wholesale central bank money can settle securely on programmable ledgers the bank doesn't operate, in environments where trust isn't inherent. Nethermind contributed as part of a consortium with the Scottish Centre of Excellence in Digital Trust & DLT, Edinburgh Napier University, siccar, and TrackGenesis. The architecture we presented runs on Ethereum, with supervisory functions inside the protocol. As per their technical and product analysis, this works at the application and execution layers. Governance and resilience are harder. They still rely on processes happening offchain. Our compliant infrastructure work is focused on closing that gap.

ERC-8004 went live in January this year. It's three on-chain registries (identity, reputation, validation) that let agents transact across organisational boundaries without pre-existing trust. It closes the gap where agent communication protocols like don’t inherently cover agent discovery and trust An open marketplace where anyone can register and reputation follows. It's not just built for Institutional finance, but you and me (and our OpenClaw/Hermes agents), for a world where our finances are often managed, grown and accessed digitally. At Money20/20 last month, a speaker asked how much money people would trust their agents with - the response was deep thoughts, few tentative hands raised when USD1000 was suggested and sideways glances. The identity gives you control, reputation gives agents credibility and validation gives agents accountability and you reassurance. Looking forward to being on a panel with with Kevin Lepsoe (ETHGas) and Jaewook Lee (ShardLab), moderated by Nat Chittamai (Abstract) at SEA Blockchain Week this week to discuss the opportunity. This is what people are shipping: 1. An agent buying GPU time from another agent to run an inference job. Settles on-chain in seconds. Reputation is exactly enough, nobody needs to know who's behind either side. 2. An agent rotating a portfolio of tokenized T-bills on behalf of a vault. The asset itself encodes eligibility. The agent has to prove the end holder qualifies before it can even touch the position. 3. A perps trading agent operating across five venues. Its liquidation history and PnL track record should be portable. Its geographic eligibility shouldn't. Would love to know what you are building - come talk to me at SEABW 2026! And find out more about what we're shipping Nethermind! #SEABW2026 #AgenticFinance #DeFi #ERC8004

Banks and AI agents will pick infrastructure for the same reason: no single operator can take it offline. On Ethereum, that comes from client diversity. BlackRock, JPMorgan, Franklin Templeton are already there. The network holds $17B in tokenized real-world assets and $175B in stablecoins. Autonomous agents are next. Tomasz Kajetan Stańczak, Founder of Nethermind, at Paris Blockchain Week 2026, with Kartik Talwar, Co-Founder of ETHGlobal.