Recap of The Graph's participation at ETHGlobal NYC 2023

The ETHGlobal Lisbon hackathon begins July 24 at Pavilho Carlos Lopes, running for 36 hours with over $100,000 in total prizes. **The Graph Prize Tracks ($15,000)** - Best AI Tooling for The Graph - Best AI Use Case of The Graph - Best Use of Composable or Standardized Graph Data Products Participants can build AI agents, MCPs, and SKILLs using live on-chain data from 15,000+ Subgraphs. The Graph will have an on-site booth for support. Pragma Lisbon follows on July 25 at the same venue. [Hackathon resources](http://thegraph.com/blog/hackathon-resources/)

**AI agents operate at machine speed**, making thousands of decisions per minute in the crypto economy. Unlike humans who browse block explorers, these agents require **structured, normalized, and queryable blockchain data** to function effectively. - Agents assessing DeFi risk cannot work with raw transaction logs - They need data formatted for rapid querying and decision-making - The infrastructure enabling this agent economy operates behind the scenes This represents a fundamental shift in how blockchain data must be organized and accessed. As AI agents become more prevalent in web3, the demand for **properly structured data infrastructure** will intensify dramatically. The agent economy isn't about human-readable interfaces—it's about machine-optimized data pipelines that can support thousands of automated decisions every minute.

**Agent0 Subgraphs are now operational across five major blockchain networks** - Base, BNB Chain, Ethereum, Monad, and Polygon - using a single GraphQL schema. **Key features:** - Real-time indexing of every ERC-8004 agent across all chains - No code rewrites needed when switching between networks - Already processed over 1 million queries - Accessible via [Graph Explorer](https://thegraph.com/explorer?search=agent0) **What this enables:** - Developers can query agent data with a single GraphQL request instead of scanning thousands of blocks - AI agents can discover, hire, and interact with each other using structured blockchain data - Machine-scale data access for agents that operate continuously The infrastructure indexes agent identities, capabilities, reputation, and validation data in milliseconds. Developers building ERC-8004 agents can now access comprehensive onchain data without building custom indexers.

A new blockchain data tool called **graph-aave-mcp** now connects Claude AI with live Aave protocol data across seven chains. **Key capabilities:** - Simulate health factors during price crash scenarios - Track liquidation events in real time - Compare supply and borrow rates across different deployments - Monitor governance proposals through conversation The tool is powered by The Graph Protocol's indexed data and requires just one npm install to access every major Aave deployment. **Who benefits:** - Risk analysts can model scenarios conversationally - DeFi researchers gain real-time market insights - Builders can prototype dashboards by describing requirements - Content creators can query live data while writing The tool eliminates manual GraphQL queries, unit conversions, and stale data snapshots—users simply ask questions in plain English.

The Graph is rolling out three institutional-grade tools for real-time blockchain data: **Substreams** delivers high-performance data streaming for DeFi, DePIN, AI, and traditional finance applications. Horizon integration scheduled for Q2-Q4 2026. **Amp** introduces a SQL-native blockchain database with verifiable data designed for enterprise scale. Developer preview is currently live, with full SQL platform launching Q4 2026. **Tycho** provides real-time onchain liquidity indexing specifically built for trading systems and solvers. Private MVP expected Q1 2026, public beta Q2 2026. These tools aim to bridge the gap between blockchain data and enterprise applications, offering familiar interfaces like SQL alongside specialized streaming capabilities for time-sensitive use cases.