Infoblox, GoDaddy back open DNS standards to verify AI agents online

When a software agent knocks on the door of another system and says it represents a legitimate business, how does that system know whether to believe it? Right now, that question does not have a clean answer. Infoblox and GoDaddy think the internet already built the foundation for one decades ago, and both companies are now pushing to extend it toward autonomous AI.

The two firms are backing separate but complementary open standards designed to give AI agents a way to identify themselves, verify their legitimacy, and find other agents across the public internet. Infoblox is advancing DNS-AID, short for DNS for AI Discovery, through the Internet Engineering Task Force. The framework lets agents publish metadata about their capabilities and endpoints through existing DNS mechanisms. GoDaddy, meanwhile, is helping develop Agent Name Service, which focuses on identity and cryptographic verification using DNS and public key infrastructure.

The logic behind both efforts is that agents operating inside controlled enterprise environments already benefit from established trust structures. Platform permissions, security boundaries, and internal policies handle the hard questions. That picture changes significantly once agents start crossing those boundaries to negotiate transactions, query external systems, or interact with third-party services without a human approving each step.

Both companies are steering deliberately away from any centralized registry or single-vendor control model. That position reflects a genuine concern shared across the industry. Whoever controls naming and verification for AI agents could exercise considerable influence over how autonomous software operates at scale. Given how closely regulators in Europe and elsewhere are already watching platform concentration in cloud and AI infrastructure, creating another centralized trust layer would likely attract immediate scrutiny.

The DNS angle is notable for a different reason. Domain names already function as identity primitives for the internet. Businesses own them. Certificates tie to them. Trust mechanisms already wrap around them. Extending those concepts toward AI agents is less a reinvention and more a careful expansion of something the internet has relied on for thirty years.

Whether open standards move fast enough to shape the market before major cloud platforms establish their own closed alternatives remains genuinely unclear. Commercial ecosystems tend to move faster than standards bodies. However, the alternative, a handful of vendors controlling how AI agents identify themselves across the web, raises questions worth taking seriously now rather than later.