Nava raises $8.3 million in seed funding to keep AI financial agents from going off the rails

The crypto world has been buzzing in recent months about the potential for applying blockchain technology to agentic commerce, a term that describes AI agents undertaking purchases or other economic activity on a user’s behalf. Early innovations include Coinbase’s x402, a standard for AI agents to transact across the web, and Stripe-backed Tempo’s “Machine Payments Protocol.” 

But for this futuristic technology to actually catch on, humans will need to trust AI agents not to go rogue with their hard-earned money. That’s the goal of a startup named Nava, which has raised $8.3 million in seed funding to build the trust side of the autonomous payments equation. The funding round was co-led by the venture capital firms Polychain and Archetype. 

“These agents will organically, autonomously start executing, creating … transactions, and managing economic activity as we see more and more of these tools enter the market,” Nava CEO and co-founder Vyas Krishnan told Fortune. “We really want to position ourselves as a trustworthy system for handling real capital across financial workflows.”

Nava-n my watch

Krishnan and co-founder Brianna Montgomery previously worked together at EigenLayer, an Ethereum-focused startup. EigenLayer founder Sreeram Kannan, who has connections to Krishnan dating to before the crypto startup, is among the other investors backing Nava.

Nava’s specific approach to reining in rogue behavior by AI agents is the creation of an escrow service that holds onto funds until an agent proposes a transaction. Once the agent does, Nava uses a verification framework to determine whether the outcome of an agent’s transaction will match the user’s intent. If the transaction passes the check, the transaction is executed. If not, the funds remain in escrow. 

The reasoning that underlies any given decision by Nava to accept or reject a transactions will be posted on-chain, thereby creating a public ledger of decisions that other AIs can reference. Nava currently runs as a layer-3 blockchain built on Arbitrum, and it will have a parallel deployment on Tempo, Krishnan said. 

In the long run, Nava’s verification layer for autonomous commerce can create the necessary conditions for AI agent insurance markets to emerge, the project’s whitepaper argues. Nava plans to release a native stablecoin for “underwriting agent action through the protocol.”

Nava’s infrastructure is meant to serve both consumers and institutions, Krishnan said. “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.”

This story was originally featured on Fortune.com