Nava, founded by former EigenLayer employees Vyas Krishnan and Brianna Montgomery, raised $8.3 million in seed funding co-led by Polychain and Archetype to build a trust verification layer for autonomous AI agent financial transactions. EigenLayer founder Sreeram Kannan also participated in the round, Fortune reported on April 14.
The product addresses a specific gap: AI agents that can spend money need a mechanism to verify that what they are about to do matches what the user actually intended.
The Escrow Mechanism
Nava works as an escrow service. When an AI agent proposes a transaction, the funds are held. Nava’s verification framework then checks whether the transaction’s outcome aligns with the user’s original intent. If intent matches, the transaction executes. If not, the funds stay in escrow.
“We really want to position ourselves as a trustworthy system for handling real capital across financial workflows,” Krishnan told Fortune.
Every decision, whether Nava accepts or rejects a transaction, gets posted on-chain. This creates a public ledger of AI financial behavior that other AI systems can reference. Nava currently runs as a Layer 3 blockchain on Arbitrum, with a parallel deployment planned on Tempo, the Stripe-backed Machine Payments Protocol.
Where Nava Fits in the Agent Payments Stack
The agentic payments infrastructure has filled in rapidly over the past two months. Coinbase’s x402 protocol and Tempo’s MPP handle how agents pay. Visa’s Intelligent Commerce Connect handles the rails. The Agentic Risk Standard (published last week by researchers from Google DeepMind, Microsoft Research, and Columbia University) provides a risk framework.
Nava occupies the trust layer: not how agents transact, but whether agents should be allowed to transact in any given instance. Krishnan framed the dual market: “Consumers are going to want to be able to protect assets from agent misbehavior, agent hallucination, but then institutions aren’t going to be able to onboard unless they have some sort of clarity of intent versus execution.”
Future Plans
Nava’s white paper argues that its verification layer could enable AI agent insurance markets. The company plans to release a native stablecoin for “underwriting agent action through the protocol.”
The founding team’s EigenLayer background (Ethereum restaking infrastructure) gives them direct experience building verification and security systems for on-chain protocols. Whether that expertise translates to the novel problem of verifying AI intent alignment is what the $8.3 million is designed to test.