NEAR Protocol lets users stake NEAR for monthly AI compute credits
The launch extends staking utility beyond validator rewards, tying locked tokens to AI usage and recoverable prepayment while preserving access to the principal after unstaking.
NEAR Protocol said users can access AI services by staking NEAR instead of linking a credit card. Under the system, staked NEAR generates monthly compute credits that can be used across 43 models on NEAR AI, including offerings from Anthropic, OpenAI and Google, while the principal remains withdrawable after unstaking. The amount of compute points varies with the staking amount, and users can increase or reduce stake based on usage. NEAR described the model as the first instance linking token staking with AI service usage and framed it as part of an "agent economy" in which the asset securing the network can also function as payment for machine-generated tasks. The protocol said broader adoption could reduce circulating supply by tying more tokens to computational work rather than free trading, though it gave no projection for how much supply might be locked. The feature follows earlier efforts to expand staking utility, including the Laser Digital NEAR Adoption Fund and a 2025 network upgrade that reduced token inflation.
- staking: Locking tokens on a blockchain to support network activity or gain access to related benefits.
- compute credits: Usage credits that pay for processing power and AI workloads on a platform.
- agent economy: A model in which software agents participate in digital markets and use network assets to pay for tasks.