AI Agents Run Entire Hackathon as $30K in USDC Awards Distributed

 

By Muhammad Hassan // February 12, 2026 @ 10:11 AM
AI Agents Run Entire Hackathon as $30K in USDC Awards Distributed

Share

Points of Focus

  • AI agents handled submissions, peer review, voting, and prize allocation without human judges.
  • More than 200 projects competed for $30,000 in USDC across three structured tracks.
  • Format compliance and on-chain proof of work determined eligibility, not visibility.

 

Autonomous agents did not just build code between February 3 – 8, 2026. They organized a competition, reviewed each other’s work, enforced rules, and allocated capital. The USDC OpenClaw Hackathon, hosted on Moltbook and OpenClaw during that period, distributed $30,000 in USDC after agents submitted 204 valid projects, cast 1,352 votes, and generated more than 9,700 comments, according to Circle’s official recap.

 

 

The structure was strict, requiring each agent to submit a project in a predefined format and vote on at least five other unique submissions, with no human judges behind the scenes, while evaluation focused on completion, technical depth, creativity, usefulness, and clarity under shared criteria published at the start of the event.

 

AI agents run hackathon lifecycle without human judges

Unlike traditional hackathons where organizers curate finalists, this experiment delegated the full workflow to software agents. Submissions, peer review, and final voting occurred through public interaction on Moltbook. Circle confirmed that only entries meeting both submission and participation requirements advanced to final consideration.

The rule design shaped outcomes, as several projects attracted attention but failed eligibility because agents invented new track names or omitted required headers, showing that engagement alone did not secure prizes and that structure ultimately determined access to capital.

 

USDC payments power agentic commerce experiment

The Agentic Commerce track was won by ClawRouter, an open-source plugin that routes large language model requests to the lowest-cost capable model and settles each request in USDC on Base. According to its submission, the system uses wallet signatures instead of API keys and claims a blended cost of $3.17 per million tokens compared with $75 for Claude Opus, citing internal routing data.

The design shifts payment from a prepaid account model to per-request authorization using EIP-712 signatures, where the wallet signature functions as authentication and agents generate their own wallets to authorize spending directly, removing human-managed API keys from the loop.

 

Onchain governance tested through AI-only DAO

In the Most Novel Smart Contract category, MoltDAO deployed governance contracts on Base Sepolia, allowing only agents to create proposals and vote using USDC voting power. Humans funded the system, but agents executed decisions on-chain. Verified deployment transactions were available on Base Sepolia, grounding the experiment in auditable infrastructure.

Community discussion focused on quorum rules and Sybil resistance, as commenters questioned whether equal distribution among 20 AI founders could concentrate influence if identity controls remain loose, underscoring that governance design, not automation alone, determines resilience.

 

Process over popularity in capital allocation

Circle noted that projects with structured documentation and verifiable deployments were easier for agents to parse and evaluate, leading participants to optimize for machine-readable clarity rather than narrative appeal and resulting in a capital distribution process driven by compliance and proof rather than social momentum.

The hackathon offers a narrow but concrete example of automated capital markets at small scale. The system built, reviewed, voted, and settled rewards in USDC under published rules. Whether that model expands beyond controlled experiments will depend less on novelty and more on how future systems handle identity, verification, and economic incentives without a human referee.

Share

Muhammad Hassan

Muhammad Hassan is a tech writer with over 11 years of experience in the crypto space. He specializes in crafting data-driven strategic content that helps blockchain and fintech brands grow their organic reach. He has led editorial initiatives for global crypto media outlets, where his strategies and article series have reached millions of readers worldwide.

Latest Podcast

Mar 17 2026 / Length: 36:29
Mar 6 2026 / Length: 46:59
Feb 27 2026 / Length: 23:56
Feb 5 2026 / Length: 55:34
Wise Prize - Pulse by Alphawire

For this week’s episode of Pulse, Aldo…

Jan 26 2026 / Length: 45:05

Ad

Related Articles