Introduction
Vint Cerf, one of the co-architects of the TCP/IP protocols that underpin the open internet, has a new focus after leaving Google in July 2026 following a 20-year run. As TechCrunch reported on July 15, 2026, Cerf is now advising Innovation Labs, an organization proposing an open architecture for AI agents to identify themselves online. The proposed standard, called DNSid, would give each agent a verifiable identity anchored to an existing internet domain name.
What DNSid proposes
Innovation Labs is a subsidiary of Identity Digital, a DNS registry company that operates hundreds of top-level domains. Its bet is that domain-name infrastructure — the same system that maps example.com to a server — is a practical foundation for holding AI agents accountable.
Under DNSid, short for "DNS-Anchored Durable Identity for AI Agents," an agent is bound to a verifiable ownership record anchored in an existing domain name. The design combines three existing pieces — DNS, public-key cryptography (PKI), and a blockchain-based log — rather than inventing a new protocol from scratch. "By anchoring verifiable ownership records in DNS, this approach leverages a globally interoperable, unambiguous namespace that already operates at world scale," said Naveed Ihsanullah, CTO of Innovation Labs.
The goal is a shared way to answer basic but unresolved questions: which agent is this, what authority does it have, who granted that authority, and who is accountable when it acts.
This is further along than an internal experiment. Identity Digital launched DNSid publicly in April 2026 and submitted it as an Internet Draft to the IETF — the standards body that ratifies core internet protocols — on June 4, 2026, opening a public comment period. Interim CEO Allie Kline told TechCrunch the company is also trialing the standard with several unnamed hyperscalers and identity companies. Cerf, in other words, is joining an effort already in motion, not launching a new one.
Why agent identity is becoming a problem
Most AI agents today stay inside proprietary systems, calling internal resources for narrow tasks. But businesses are already planning for a world where agents operate far more autonomously across the open internet and interact directly with other agents — a shift that puts pressure on multi-agent systems to establish trust between parties that have never met.
The road block, Cerf argues, is the absence of a shared standard for identifying and auditing agents. And agents raise harder questions than domains do, because they are so much more active.
"This is largely triggered by the notion of AI agents and the question of what authorities they have, where they have derived those authorities, who is accountable for the behavior of an agent in this context, and where and how its identity is established, and why [you'd] trust it," Cerf told TechCrunch. He called the period ahead "fascinating — and at the same time maybe even exasperating."
A crowded field
DNSid is far from the only proposal. Identity is a different layer from the interoperability protocols that already define how agents talk — Anthropic's Model Context Protocol (MCP) for connecting agents to tools, and agent-to-agent (A2A) protocols for coordination between them. Those govern communication; DNSid tries to govern who is speaking. On that identity layer alone, the field is busy: Visa has a Trusted Agent Protocol for authorizing agent-initiated payments, blockchain communities are building on-chain agent registries, and a range of commercial "agent passport" ventures have launched since late 2025.
Cerf's view is that functionality, not mandate, decides which of these wins — the same dynamic that drove adoption of TCP/IP itself.
"Company X uses agent Y's technology, and company A uses agent C's technology, and then they don't interwork with each other," he said. "Nobody can do everything that you might want every agent to do… and so we're going to have to rely on the pressure coming from the users."
Kline framed neutrality as DNSid's edge: Innovation Labs says it does not plan to build other AI businesses on top of the standard or own the registration data. "I think there's a lot of organ rejection to a hyperscaler releasing [a standard] and having that proprietary data," she said.
That claim deserves scrutiny, because Identity Digital is not a disinterested party. It is a domain registry, and a world where every AI agent needs a domain-anchored identity is a world with a much larger market for the exact product it sells. "Neutral" here means not hoarding the agent data — not an absence of commercial stake in the outcome. It is a narrower promise than it first sounds, and worth reading as such.
Conclusion
DNSid is a live proposal — public since April 2026 and under review at the IETF — but not a ratified standard, and it competes with payment networks, blockchain registries and commercial passport schemes for the same job. Its bet is that the internet's existing naming system is the natural place to anchor accountability for software that increasingly acts on its own. Whether the agentic internet arrives on DNSid's terms — or at all — is unsettled. Cerf is characteristically measured about the destination: "I don't think it's inevitable," he said of an agent-driven economy. "But what I do think is inevitable is that people will try to do that."
Sources
- Vint Cerf is working on a plan to unleash AI agents on the open internet — TechCrunch, July 15, 2026
- Identity Digital Launches Neutral, DNS-Anchored Identity Standard for AI Agents — Identity Digital newsroom, April 27, 2026
- Innovation Labs by Identity Digital Submits DNS-Anchored Durable Identity Proposal for AI Agents to the IETF — GlobeNewswire, June 4, 2026
- DNSid — The Birth Certificate for AI Agents — official project site