Who will own the AI agent economy?
Who will own the AI agent economy?
https://mitsloan.mit.edu/ideas-made-to-matter/who-will-own-ai-agent-economy
Publish Date: 2026-07-13 12:04:00
Source Domain: mitsloan.mit.edu
-
Personal AI Agents: Ramesh Raskar envisions a future where every individual, like a pre-diabetic woman from rural India, will have their own personal AI agent that manages various aspects of their life, from booking travel and hotels to creating diets and event schedules tailored to their needs.
-
Project NANDA: Raskar leads Project NANDA, a research initiative that explores various aspects of the future AI agent systems, including identity, discovery, addressability, registries, verification, interoperability, and decentralized coordination.
-
Evolution of AI: Raskar compares the current phase of AI to a “mainframe era” evolving into a “PC era,” where AI agents become accessible to more individuals and organizations as computational costs decrease.
-
Agent-Centric Economy: Raskar argues that the future economy must shift from creating “agents for X” (discrete tasks or services) to creating “X for agents” – infrastructures and services that allow agents to perform their tasks efficiently, focusing on marketplaces, protocols, and services agents require.
-
Future AI Infrastructure: Raskar predicts several key sectors forming the infrastructure for this agent-driven economy, including agent identity systems, trust services, insurance and repair services, and stablecoin-based micropayments.
-
Maintaining Open AI Ecosystem: Raskar emphasizes the importance of keeping the AI agent ecosystem open to prevent monopolization by a few corporations, likening current risks to the limited interoperability seen in software like Microsoft PowerPoint and Apple’s Keynote.
-
Optimism and Efforts: Despite his concerns about a centralized future, Raskar is working through Project NANDA to establish a decentralized framework for the AI agent web before it’s too late.
-
Broad Research Focus: Raskar’s work at MIT Media Lab involves distributed AI agent architectures, machine learning, imaging for health and sustainability, and networked AI agents to create an “internet of AI agents.”
