The integrity of the input determines the integrity of the outcome. AI agents are only as good as their onchain data feed. Here’s what happens when you swap a delayed RPC for Subgraphs: 1️⃣ Speed → real-time decisions instead of stale data 2️⃣ Reliability → no single point of failure at machine speed 3️⃣ Accuracy → verifiable, indexed blockchain snapshots What’s one onchain data pain point you want solved for AI agents next?

Amp ingests any onchain data and any offchain data, from Canton, Solana, and Ethereum to legacy databases, KYC systems, CRMs, and ERPs. Every result is cryptographically verifiable before it reaches your analytics stack, whether that's BI tooling, operational dashboards, or compliance reporting.

The data a risk desk needs is already onchain: issuance events, redemption flows, collateral ratios, funding rates. The gap is infrastructure. Substreams move that data into a structured, queryable form; Amp makes the record verifiable and audit-ready for examination.

Big launch from @PropellerSwap: Turbine brings confidential, direct-to-solver swaps to Ethereum. Powering it is Tycho, their open-source liquidity indexer built on @graphprotocol's Substreams, the highest-throughput way to stream onchain data in real time. Congrats to the team!

First it was anonymous wallets testing the endpoint. Now it's real named agents. Volume is still low, but named, registered agents are paying autonomously for onchain data… and that is a signal. Track it live: x402-watch.vercel.app

The @EthereumRemix AI Bootcamp is live - but you can still join! 5 weeks with @graphprotocol, @ZKVProtocol, @chainlink, @base and @ensdomains. Real workshops, real builds. Past sessions (available to watch): - Intro to Remix: Building Real-World dApps with AI - Base: Smart Contracts for Advanced Agentic Payment Flows - Chainlink × Dynamic Oracles for Healthcare & Real-World Data Next sessions: - zkVerify: Verifiable AI and Zero-Knowledge Infrastructure - June 25 - ENS: Sovereign Identity and Human-Readable Infrastructure - June 29 - The Graph: Verifiable Indexing for Real-World and Clinical Data - July 1 Don't sit this one out ↓

“Stablecoins" isn't one asset. Fiat-backed, crypto-collateralized, delta-neutral synthetic, and algorithmic each carry different risk profiles, failure modes, and data needs. An institution that can't tell them apart in its risk system isn't ready for a regime that will ask it to.

AI agents read data, in milliseconds, across every chain at once. The infrastructure built for human apps wasn't designed for that. The Graph was.

MiCA is live. The GENIUS Act is moving. A legal definition of a compliant stablecoin is taking shape: fiat-collateralized, redeemable, auditable. The institutions that build stablecoin data infrastructure will now be ready when the rules land. The ones that wait won't⬇️ thegraph.com/blog/stablecoin…

"Where can I find the best stablecoin lending rates right now?" That single natural language query, running through the Graph Lending MCP, fans out across 40+ protocols simultaneously and returns comparable, structured results. That's what standardized Subgraph schemas on The Graph Network make possible.

The chain data you need, in your database, backed by a world-class backend. Getting on-chain data into your database is now the easy part via @graphprotocol's The Graph Market. Hosted Sinks is a fully managed Substreams-sink-as-a-service. You point it at a Substreams package and a database, click Deploy, and we run it for you, at scale, securely, with zero ops on your side. Chain data to your tables, live, in minutes, then query it with the SQL tools you already know. --- What it does Bring your Substreams package, pick PostgreSQL or ClickHouse, and hit Deploy. That's it. Fresh chain data starts streaming into your tables, and we keep it running, backed by our highly-available and highly-reliable distributed infrastructure. Running a different kind of sink? Let us know so we can prioritize it! Transactional data for your app, columnar analytics for your dashboards, your choice. Works on all chains available on The Graph Market: Ethereum, Base, Solana, and beyond. No new runtime on your infrastructure, no blockchain plumbing: your Substreams package maps protobuf straight into relational tables, and you read it back with plain SQL. You already have a database, let us connect and fill it for you. --- How your team benefits Spin up in minutes, not sprints. Deploy, start, stop, and delete on demand — all self-service from the portal. No infrastructure to provision. Always correct, fork-aware. Substreams tracks the canonical chain and signals reorgs automatically, so chain reorgs never leave sheared or stale data behind. Inserts, updates, and upserts keep mutable data current. No manual cleanup. Update without downtime. Edit your package or config in place and we roll it out gracefully. Need a clean slate? Reset and re-sync from your start block. Secrets stay secret. Your database credentials live in HashiCorp Vault, never in code or config. Injected at runtime, with built-in disaster-recovery export/import. Multi-tenant by design. Every deployment is isolated to your organization: strict, enterprise-grade separation out of the box. Full visibility. Live status, per-instance health, and a complete deployment history — created, scaled, updated, restarted — all in the dashboard. --- What does it cost Billing is the same as it currently is, pay only for the data coming out of your Substreams. For the next 6 months, we’ll manage the sink for free. After this initial period, each sink will incur a negligible cost per month, confirmed before billing begins. The bottom line Stop staffing a platform team to move blockchain data into a database. Hosted Sinks abstracts away the credential management, the scaling, and the re-sync headaches and gives you back a simple form and a Deploy button. Check the docs to get started today docs.substreams.dev/how-to-g… Your data. Your database. Our infrastructure. Live in minutes.

Team E&N shipping, shipping, shipping 🚢

ERC-8004 for identity. x402 for payments. Subgraphs for data. All three are live. All three are being used together, in production, on Base, right now.

The progression is worth understanding: MCP gives AI assistants real-time data access. Skills bundle the documentation and domain expertise into a single install. x402 lets agents pay per query autonomously, no API key setup required. Each layer removes a step between an agent and the data it needs to act.

Every new @paysponge agent inherits the same routing path to The Graph with zero new integration work. Michaelgent and Clarabotagent aren't one-off scripts. They're part of a financial platform built for the agent economy. That's a different growth pattern than hobbyist traffic, and it's the one worth watching.

Blockchain data has always been transparent. Accessing it hasn't been. MCP servers and new AI agent skills for both Subgraphs and Substreams mean anyone can now query live onchain data in plain natural language. No GraphQL. No SQL. No custom infrastructure. Here's what's available and what's coming next ⬇️ thegraph.com/blog/querying-b…