NEAR is using its current hackathon to push developers toward applications where artificial-intelligence agents participate directly in economic activity. According to the official NEAR Hackathon, its Agent Economy and AI Swarms tracks ask builders to create software agents that can spend, negotiate, purchase resources and coordinate economically with other agents. The two categories form part of a broader 13-track builder program running through November 14.
Agent Economy is explicitly framed around giving AI systems wallets and economic agency. NEAR asks participants to build agents capable of earning money, paying for APIs, compute or data, negotiating transactions and hiring other agents, either with crypto directly or through NEAR Intents. The track treats payment capabilities as part of an agent’s operating environment rather than as a separate checkout step controlled manually by a human.
AI Swarms Extend Payments Into Coordination
AI Swarms takes the concept beyond individual agents. The track calls for coordination markets, bidding systems, collective purchasing and intent-based settlement between multiple autonomous participants. The design target is an economy in which agents can coordinate resources and transactions with one another rather than simply execute isolated instructions for users. NEAR identifies multi-agent systems, Intents and sub-accounts as relevant infrastructure for those applications.
That direction is consistent with infrastructure NEAR has already been assembling around autonomous software. Its AI Agent Market combines agent discovery, bidding and payments with Intents and other components intended to support cross-chain execution. The hackathon is therefore asking developers to experiment with use cases around infrastructure that already exists to varying degrees, rather than introducing an entirely new agent architecture for the event.
NEAR Intents provides one of the financial layers underneath that thesis. Users or agents submit desired outcomes while competing solvers provide quotes and execute the selected route, with settlement ultimately occurring onchain. The abstraction can let an agent express what it wants to acquire or pay without manually determining every bridge, venue and intermediate transaction required to complete the operation.
Other NEAR products are approaching the same problem from the execution side. The ecosystem has connected staking with hosted IronClaw AI agents, while NEAR AI has been expanding private inference and agent infrastructure. Together, those components show why the hackathon emphasizes payments, permissions and coordination rather than treating AI agents purely as conversational interfaces.
Builder Tracks Are Not Adoption Metrics
The hackathon itself does not demonstrate that autonomous agent economies are already operating at scale. Participants are required to produce working prototypes, publish their repositories and post public demos before submissions close. A prototype showing agents negotiating or paying each other establishes technical feasibility, not recurring economic activity, commercial demand or autonomous transaction volume.
The distinction matters as agent-payment infrastructure spreads across blockchain ecosystems. Sui and Alibaba Cloud, for example, are developing pay-per-call infrastructure for AI agents, while NEAR is approaching the same emerging market through Intents, agent discovery, private inference and cross-chain execution. Different networks are increasingly treating AI agents as potential customers and counterparties rather than simply software controlled by human users.
For NEAR, the hackathon provides a practical way to test which applications developers can build on top of that stack. Its Agent Economy and AI Swarms tracks are best understood as a directional signal about the applications NEAR wants its infrastructure to support, while actual adoption will depend on whether those prototypes progress into services with recurring agents, transactions and paid demand after the competition ends.