Internet of Agents @Wilson Sonsini presented by MIT Nanda
Reimagining Internet in the age of AI Agents: Ventures and Investments
Thursday May 16th, 5:30-7pm
Reception: 530pm-6pm, Presentation-Discussion 6pm-7pm
Can our existing web infrastructure handle billions of autonomous AI agents? Does this remind you of the days before Netscape? Do we have to reinvent or rebuild http/url/dns/ICANN/Verisign/Godaddy and more.?
Much like how we once realized dialup couldn't support the evolving internet, our current DNS, IP addressing, and certificate systems weren't designed for agent-to-agent interactions at massive scale. These systems operate on human timescales, with limited metadata capabilities and propagation times measured in hours—potentially inadequate for agents requiring real-time verification and discovery.
As we stand at the threshold of a profound digital transformation—the rise of the agentic web— the MIT team at Project Nanda: Architecting the "Internet of AI Agents" is reflecting on a critical question: Upgrade or Switch?
The solutions such as MCP from Anthropic and A2A from Google show great promise for Agent to Agent (or agent to tools) communication. But we need to create a whole new architecture for tomorrow’s agent ecosystems. What are emerging research areas, startups, protocols and investment opportunities?
At MIT, Prof. Ramesh Raskar and his team have been pioneering the next wave of decentralized intelligence. Their initiative, 10 years in development — NANDA (Networked Agents + Decentralized AI) — introduces the critical infrastructure needed to scale distributed agent intelligence.
Deep Dive into the NANDA-MCP architecture for the 'Internet of AI', and how it will change the venture and startup ecosystem.
Prof Ramesh Raskar, MIT, Director of Project Nanda, https://nanda.mit.edu
Deep Nishar, General Catalyst
Gunjan Sinha, OpenGrowth Ventures
Follow Nanda at https://nanda.mit.edu
https://www.linkedin.com/feed/update/urn:li:activity:7317906490458058753/
Papers: https://github.com/aidecentralized/nandapapers/tree/main