The Graph (GRT) is a decentralized indexing and querying protocol that organizes blockchain data so that applications can retrieve it quickly and reliably. Think of it as the Google of blockchain — not a search engine for the web, but an infrastructure layer that makes on-chain data queryable without each developer having to build their own data pipeline from scratch.
Background
To understand why The Graph exists, consider what it takes to build a decentralized application. A smart contract can store and emit data on a blockchain, but raw blockchain data is not designed for fast lookups. If a DeFi app wants to display a user’s full transaction history, the price history of a token pool, or the governance votes in a DAO, it cannot simply run a database query. Historically, developers had two bad options: run their own centralized server to index the blockchain (which defeats the purpose of decentralization), or query the chain directly in real time (which is slow and expensive in terms of gas and fees).
The Graph solves this by introducing a shared, open protocol where independent operators called Indexers process and store blockchain data in organized structures called subgraphs. Any application can then query those subgraphs using GraphQL, a widely used query language, getting structured answers in milliseconds instead of parsing raw blocks.
The practical result is that decentralized applications — particularly in DeFi and NFTs — can deliver fast, responsive user interfaces without sacrificing decentralization at the data layer.
History
The project was conceived around 2017 and 2018 by a team of engineers frustrated by the difficulty of building usable decentralized applications on Ethereum. Yaniv Tal, Brandon Ramirez, and Jannis Pohlmann had worked together at a software company and shared a conviction that the missing piece for Web3 was a reliable data layer.
The Graph launched its hosted service in 2019, allowing developers to deploy subgraphs for free on infrastructure run by the team itself. This was a centralized stopgap — useful for proving the concept, but not the final vision. It quickly attracted major protocols. Uniswap, Compound, Aave, and others relied on The Graph’s hosted service to power their front-end dashboards.
The mainnet of the decentralized network launched in December 2020, bringing with it the GRT token and the economic machinery that lets independent participants run the protocol. The team announced the eventual sunset of the hosted service to push developers toward the decentralized network — a transition that unfolded gradually over subsequent years as the tooling matured.
A key milestone in the project’s evolution has been the expansion beyond Ethereum. The Graph has progressively added support for additional blockchains, including Solana, Arbitrum, and others, moving toward its stated goal of indexing all public blockchain data.
Technology
The Graph’s architecture involves four distinct roles, each with a specific economic incentive.
| Role | What they do | How they earn |
|---|---|---|
| Indexers | Run nodes that index subgraphs and respond to queries | Query fees and indexing rewards |
| Curators | Signal which subgraphs are valuable by staking GRT | A share of query fees from signaled subgraphs |
| Delegators | Stake GRT to Indexers without running infrastructure | A portion of the Indexer’s earnings |
| Developers | Deploy subgraphs and pay for query fees | N/A — they pay into the system |
A subgraph is a configuration file that tells Indexers which smart contract events and calls to watch for, how to transform that raw data, and how to store it in a queryable schema. Once a subgraph is deployed, Indexers process the relevant blockchain history and serve queries against it.
Queries are made using GraphQL, which lets developers specify exactly the shape of data they need. This is more efficient than generic REST APIs because the server only computes and returns what was asked for.
The protocol uses proof of stake mechanics to secure both the quality of indexing and the honesty of query responses. Indexers must stake GRT as collateral. If they serve incorrect data or behave maliciously, their stake can be slashed — destroyed as a penalty. This mechanism aligns financial incentives with protocol integrity.
The Curation mechanism is particularly clever. Curators essentially bet GRT on which subgraphs will attract the most query traffic. Because they earn a cut of fees from subgraphs they curate, they are incentivized to identify useful, high-quality subgraphs early. This creates a market-driven signal for Indexers about where to focus their resources.
The Graph demonstrates a broader pattern in Web3 infrastructure: using token-based incentives to coordinate a decentralized workforce of node operators without any central employer. The same pattern appears in Chainlink’s oracle network and Filecoin’s storage market.
For developers, The Graph’s tooling integrates with standard Web2 development workflows. Subgraphs can be written and tested locally, and the GraphQL interface is familiar to anyone who has built modern web applications. This deliberate focus on developer experience has been central to the protocol’s adoption.
Tokenomics
GRT is the native token of The Graph Network and serves three functions: staking collateral, curation signal, and the medium of exchange for query fees.
The total supply of GRT is capped at ten billion tokens. At mainnet launch, a portion of tokens was allocated to the founding team, early backers, and a foundation treasury, with vesting schedules designed to prevent immediate selling pressure. The protocol also has an annual issuance rate that rewards Indexers for their work — this inflation is the mechanism that distributes new GRT to active participants rather than passive holders.
A counterweight to inflation is query fee burning. A portion of the fees paid by developers to use the network is burned — permanently removed from supply — rather than paid out. This creates deflationary pressure that scales with network usage. The balance between issuance rewards and burn is a deliberate design choice: as query volume grows, the burn rate increases, and the net inflation to the supply decreases. This is a common design in token economics intended to tie token value to actual protocol utility.
Delegators who stake GRT to Indexers receive a share of earnings but are subject to a delegation tax — a small percentage of staked GRT is burned when delegating. This discourages rapid in-and-out movement of delegated stake and rewards longer-term participation.
Because Curators bond GRT to subgraphs using a bonding curve, early curation of a successful subgraph is more profitable than late curation. This creates a mild first-mover incentive for curators to identify valuable data sources before they become widely used.
In summary
The Graph addresses a real and unglamorous problem — getting data out of blockchains efficiently — that every serious decentralized application has to solve. Its protocol replaces ad-hoc, centralized indexing servers with a shared, incentivized network of Indexers, Curators, and Delegators. Whether that network achieves sufficient decentralization and competitive performance relative to centralized alternatives is a question the market continues to evaluate. For anyone building or studying Web3 applications, understanding how data indexing works is essential context for understanding how the whole stack hangs together.
Last reviewed January 1, 2026.