NEAR apps and AI
Application contracts and confidential inference

NEAR applications use smart contracts to hold assets, trade tokens and manage staking positions. Token contracts define the assets and their rules, with individual IDs distinguishing NFTs and multi-token assets. NEAR AI Cloud adds model-hosting services with privacy terms tied to the model and request route.
Application tokens have their own contracts
NEAR’s token standards define shared interfaces for fungible tokens, NFTs and multi-token assets. Wallets and applications use those interfaces to query balances and transfer assets. A fungible token is identified by its contract account; NFTs and assets within a multi-token contract also need their individual token IDs. Check the contract, any applicable token ID and the network, particularly when bridged versions use similar names.
Trading uses application liquidity
A decentralised exchange uses contracts and liquidity pools to exchange tokens. NEAR’s developer documentation includes Rhea Finance integrations and explains wrapped NEAR for token trading. Wrapped NEAR is a contract-based representation used by applications; it should be distinguished from a native NEAR balance. Pool liquidity and the size of a trade affect the executable quote.
NEAR decentralised exchange documentation.
Staking services have different designs
The NEAR documentation lists Meta Pool, Rhea Finance and LiNEAR as liquid staking providers. Each service has its own contracts, fees, validator allocation and redemption process. Depositing a liquid staking token into a lending market adds that market’s contract and liquidation rules to the position. Compare the underlying stake, the token’s redemption terms and any further application that uses it.
AI infrastructure is another layer
NEAR AI Cloud provides model inference infrastructure, including models hosted in trusted execution environments. TEEs isolate computation within supported hardware, and attestation helps users verify the environment. Privacy claims should be checked against the specific model’s hosting route and verification procedure. An AI service is a separate system from the blockchain validators that process NEAR transactions.
Product revenue can fund several economic uses
NEAR’s 24 February 2026 economics update describes Intents revenue and a governance framework with discretion over tools such as token buybacks and staking acquired tokens. The update says no particular mechanism is guaranteed. A buyback transfers tokens to a buyer, staking commits them to network security, and burning removes them from supply. Product revenue reaches holders through the mechanisms actually implemented and their rules.
Volume, value locked and revenue measure different things
Trading volume counts activity over a period. Value locked measures assets deposited at a point in time. Fees measure charges paid, while protocol revenue depends on how those charges are distributed. A liquid staking position can appear in more than one application’s figures, so adding headline totals can count the same underlying assets repeatedly. Cross-chain trading volume also includes assets delivered to other networks.