Vint Cerf’s DNSid Project Could Expand the AI Infrastructure Trade
Listen to the audio version of this article (generated by AI).
AI agents have learned how to work.
They just haven’t learned how to leave the office.
Right now, most of the inference demand driving the AI infrastructure boom is happening inside company walls. A bank runs agents on internal risk data. A logistics company uses agents to optimize its own supply chain. A retailer deploys agents to manage procurement inside systems it already controls.
That demand and spending are real. And the trade around chips, memory, networking, storage, and power is already playing out.
But it is still mostly contained.
And Vint Cerf just started working on the layer that could let it break out.
Why Vint Cerf’s DNSid Project Matters for AI Agents
Cerf helped build the modern internet. He’s one of the architects behind TCP/IP – the foundational protocols that made the internet possible.
After more than two decades at Google, he is now turning to his next project: identity infrastructure for AI agents operating on the open internet.
As TechCrunch reported, Cerf is now advising Innovation Labs on something called DNSid – essentially a passport system for AI agents. The idea is to link each agent to a verified domain name and use cryptographic proof to show where it came from, who authorized it, and who is responsible for what it does.
Cerf helped solve the internet’s first coordination problem. Now he’s working on the next one.
The Trust Problem Holding Back the Agentic Web
The agent economy is still mostly trapped inside company walls.
Take a retailer’s procurement agent, for example. It operates inside the retailer’s own systems – searching its own inventory databases, working within its own supplier relationships, accountable to its own IT team.
Most enterprise AI agents in production today work this way, contained within a single organization’s limits. And that’s why the current infrastructure demand, as large as it already is, may represent only a fraction of what’s coming.
The biggest use cases – and the ones that would generate the most compute demand – involve agents crossing organizational boundaries.
- A procurement agent that negotiates directly with a supplier’s agent.
- A financial agent that transacts in real time with a bank’s agent.
- A logistics agent that coordinates across a dozen different carriers’ systems simultaneously.
- A healthcare agent that pulls verified records from multiple hospital networks to inform a treatment decision.
All require agents to operate across the open internet – interacting with unfamiliar systems, on behalf of humans who are not watching every step.
But that world has a trust problem.
When an agent shows up somewhere on the internet today, there’s no reliable way to verify who sent it, what it’s authorized to do, or who’s accountable if something goes wrong. Without a solution, the highest-value agentic use cases simply can’t safely deploy at scale.
That is the missing layer.
What Happens When AI Agents Can Work Across the Open Internet
Before shared internet protocols, computer networks were islands. Each organization ran its own system. Those systems had value, but they could not easily talk to one another.
Once a shared standard let those networks communicate, the internet became a global market. Every person and institution suddenly needed to connect. The demand for routers, cables, servers, and all the physical infrastructure underneath became essentially limitless.
The agentic web is approaching a similar moment. Today’s enterprise agent deployments look a lot like those isolated networks of the 1980s: valuable, growing, but fundamentally contained. Once a shared identity standard lets agents operate across organizational boundaries, with accountability built in, the addressable market for agentic infrastructure will likely expand dramatically.
Every cross-enterprise workflow becomes a potential agent-to-agent interaction. Every government service, financial transaction, and logistics chain becomes a candidate for agentic automation – each requiring inference compute, memory, networking, and storage that currently sits outside the demand projections most investors are working from.
The infrastructure thesis doesn’t change. The size of it does.
The Investment Implication: The Inference Market Gets Bigger
The physical infrastructure stack – accelerators, high-bandwidth memory, optical networking, power, cooling, storage – remains the core of the trade. That demand keeps growing, and the companies supplying it are reporting it in earnings quarter after quarter.
Cerf’s work adds a new layer. If agents get a trusted way to identify themselves online, the demand story moves beyond internal enterprise workflows and onto the open internet. That would create a larger market than most investors are modeling.
Cloudflare (NET) may be the cleanest public-market way to play that identity-and-routing layer. It already sits in the flow of internet traffic, security, authentication, and developer infrastructure. And its tools are increasingly being built for a world where agents need to discover services, prove who they are, and transact across the web. If the agentic web moves beyond the enterprise firewall, Cloudflare could become one of the trust-and-routing layers underneath it.
Beyond that, the broader infrastructure names supplying the physical substrate that every agent workload runs on are the same names that benefit directly when the agentic economy expands. More agents operating across more boundaries means more inference calls, more memory consumption, more networking traffic, more power draw.
The market understands enterprise agents.
It has not fully priced internet agents – or their much larger workload.
The Bottom Line: AI Agent Identity Could Unlock the Agentic Web
Agents are already working inside companies.
The bigger opportunity begins when they can work between companies.
That requires identity and trust – a way to know who sent the agent, what it is allowed to do, and who is responsible if something breaks.
Vint Cerf is working on that layer now.
If it works, the inference supercycle expands onto the open internet.
The physical infrastructure that powers that market is already being built. The companies supplying it are already reporting the demand. And the names best positioned for the next leg of this expansion – the ones the market hasn’t fully found yet – are exactly what we’ve been tracking.
That’s the trade. And it just got bigger.
