Building the Internet of AI Agents: MIT Project NANDA @ Capgemini
As we enter the era of agentic AI, Project Nanda—incubated at MIT—is building the foundation for an “Internet of AI Agents.” Just as the early web needed DNS and HTTP to scale, Nanda is reimagining core building blocks to support trillions of autonomous agents that must discover, verify, and collaborate in real time.
At the heart of the project is a key question: Can we upgrade the current internet—or do we need a new one?
📅 PROGRAM
Opening Remarks
John Robins, Head of AI & Data Business, Capgemini Invent – Synapse
Keynote
“The Infrastructure Challenge for the Agentic Web: Can today’s internet handle billions of autonomous AI agents?”
Dr. Ramesh Raskar, Associate Professor, MIT Media Lab; Head, Camera Culture Group
Fireside Chat
Dr. Ramesh Raskar & Andreas Sjostrom, VP, Applied Innovation Exchange, Capgemini San Francisco
ABOUT PROJECT NANDA
The internet wasn’t built for AI Agents to talk to each other autonomously. DNS, IP, and digital certificates were designed for human-centric use, with limited metadata and propagation speeds measured in hours, not milliseconds.
Nanda introduces a modular, decentralized stack for the agentic era:
Decentralized Registry: Like DNS for agents—enables discovery, authentication, and verifiable interaction.
Agentic Protocols: Built for advanced A2A communication, distributed discovery, and semantic search.
Verifiability & Trust: Native identity, traceability, and behavioral records.
Open Ecosystem: Host your own registry, plug into APIs, or contribute to open source.
🔗 Learn more: nanda.mit.edu