AI Swarm Company: IMD's Burn Flywheel and On-Chain Labor Experiment
Original author: @wmpeaster, Bankless Newsletter
Original compilation: AididiaoJP, Foresight News
A rapidly heating-up AI project has recently emerged in the Ethereum ecosystem: IMD (identity.md). Its goal is to build a company operated by AI agents and owned by the community. The project's mechanisms include NFT seats, agent swarms, IMD/sIMD, the Uniswap V4 burn Hook, and a Community Coins launchpad. After the agent network opened to NFT holders on September 20, the number of nodes grew from a few dozen to more than 370 within five days, and IMD rose over 200% in the past week.
The core logic of IMD is that swarm agents may become a labor force that other protocols are willing to continuously pay for, such as inference oracles or other broader uses. To understand this mechanism, it needs to be broken down into several modules.
Background
IMD founder Adam (@surfcoderepeat) previously launched the on-chain virtual pet game Fren Pet on Base, which went live in August 2023. In October 2025, he bridged FP to Ethereum via LayerZero. A few months later, FP was renamed VIBE; in May 2026 it was renamed again to IMD, along with a free mint of 2,000 identity.md NFTs.
The project has recently picked up pace. Liquidity migrated to Uniswap V4 last month; in September, the POOL4 burn Hook, sIMD staking, a bridge to Robinhood Chain, and the NFT-gated agent network launched successively. Subsequent progress has been disclosed by the team mainly on-chain: Adam posts updates in the transaction memos sent from address 0x200E710aCAA6A93bbc77146026328C40F1d60fB1, which can be read directly on Etherscan here.
How IMD Works
1) NFT Seats
There are 2,000 identity.md NFTs, each corresponding to a seat in the swarm. Holders must first register the NFT as an agent according to the ERC-8004 standard, which is Ethereum's standard for providing on-chain identity and reputation for AI agents. They then install the open-source worker client on a machine—a VPS is recommended—and connect their own Claude or Codex subscription.
One NFT authorizes only one active device, and compute costs are borne by the holder. High-value tasks, such as smart contracts and frontend development, are currently assigned only to seats running top-tier models, such as Claude Fable 5.1 at high effort.
2) The Swarm
The main orchestrator agent publishes tasks to the network, and seats claim, execute, and submit results. Validators rerun the tasks in isolated containers to confirm that only permitted files were modified. Other seats conduct adversarial review, and passed work is recorded on-chain for reputation.
Outputs are visible in the public explorer, including code, websites, oracles, audits, reports, and images. Current code output is mainly Uniswap V4 Hooks on testnet, websites are published to IPFS, and oracles consist of multi-agent independent answering panels.
Starting this week, external users can also purchase work. Anyone can open an order by paying 0.5 IMD, settled via x402. Deployment is currently still only on the Sepolia testnet, with mainnet deployment already on the agenda.
3) $IMD and sIMD
$IMD is the currency of this economy. Its total supply is fixed, distributed across Ethereum, Base, and Robinhood Chain, bridgeable 1:1, and the total can only decrease. Counting cumulative burns since the Fren Pet era, the initial 10 million has fallen to about 7.1 million, a reduction of roughly 29%.
Holders can deposit IMD into the StakedIMD vault in exchange for sIMD. sIMD is an ERC-4626 standard yield-bearing share; redemption yields more IMD, with no lock-up and no separate claim required. Currently about 2.3 million $IMD is staked, accounting for about 32% of total supply. On September 19, Adam renounced ownership of the staking contract, so admins can no longer execute emergency withdrawals on behalf of stakers.
4) POOL4
The protocol itself holds the main ETH/IMD pool on Uniswap V4, namely POOL4, run by CappedBurnHook. This Hook sets a cap on the amount of IMD in the pool. When selling pressure pushes the IMD in the pool past the cap, the Hook trims the excess once the trade settles.
Of the trimmed tokens, 85% is burned, 6% is reserved for the orchestration compute reserve, 4.5% is distributed to stakers, and 4.5% to NFT seats. The ETH freed up by the trim is then laid out as a buy wall below the current price. The cap is lowered by about 1,000 $IMD per day to prevent it from stalling. Selling pressure drives burns, and burns feed back to stakers and seats.
5) Community Coins
Community Coins is IMD's own launchpad. Anyone can issue a coin by paying gas, each with a total supply of 1 billion, priced by a bonding curve in $IMD rather than ETH. On the surface, trades are conducted in ETH, but underneath, every buy is essentially buying $IMD.
In each trade, 1% is distributed to ETH/IMD liquidity providers, 0.5% of the ETH side goes to the issuer, and 0.5% of the $IMD side is burned. These new coins share the same base pool, so buying any one lifts the others, and selling any one pushes the others down.
The Flywheel and the Swarm's Current State
Under the existing design, the ideal loop is: users trade IMD and Community Coins; POOL4 trims and burns, supply falls, and stakers and seats benefit; the orchestrator agent turns demand into shippable products; useful output attracts more people to issue IMD-priced coins, purchase work, and increase demand for seats; seat scarcity lifts the NFT, and token demand and burns lift $IMD.
The key question is whether the flywheel's cornerstone—the swarm—can actually do work today. According to IMD public API data from September 25, of the 2,000 seats at the time, about 380 were registered, with 372 agents online. Of these, 334 registered seats had submitted accepted work, 267 had more than 50 accepted records, and 193 had more than 100. The top 10 seats accounted for about 8.5% of accepted work, and the top 50 for about 32%.
Quality data is decent. Of about 50,700 attempts, 86% were accepted, direct rejections were only about 1%, and the rest were failures or queued. The pace is also accelerating, with about 29,600 accepted submissions in the past 24 hours. Paying customers have just begun to arrive: the public x402 channel has so far completed 115 orders at 0.5 IMD each, totaling about 57.5 $IMD, or roughly $560. Paid demand has emerged, but the scale is still early.
Ways to Participate
If you are watching the project, IMD is a relatively direct way to participate. It is a fungible token with a low trading threshold and can accommodate different amounts of capital; a seat costs about 1.99 ETH. If you judge that trading volume and burns are sustainable and that paid swarm demand can expand, IMD can be seen as exposure to the corresponding token sink. This is a bet on the token mechanism, not on ownership of the agents. You can later restake it as sIMD, but the yield depends on whether sell-side flow enters the POOL4 Hook.
The NFT side is better suited to more advanced participants: you can run a node yourself, or commit about 1.99 ETH (the floor price will change) to bet on seat capacity and scarcity. Seat income is currently limited, but it could change in the future and is worth tracking. It is akin to an option on future wages, while also occupying potential positions such as whitelists and airdrops. If you think the project's mechanism is complex, or judge that it remains only an on-chain experiment that is difficult to scale outward, choosing to stay on the sidelines is also reasonable.
Conclusion
Can IMD reach the hundreds-of-millions market cap of other leading agent projects, such as the Virtuals tier? For reference, even after drawing down about 85% from its all-time high, VIRTUAL still has a market cap of about $508 million. For IMD to catch up from its current position, it would need to rise roughly another 7x.
The two are not entirely comparable. Virtuals issues a separate token for each agent; IMD instead tokenizes a capped roster of agents together with the entire system. What happens next depends on how the market converges on IMD. If it truly becomes a labor swarm that more and more protocols are willing to pay for, further gains are not impossible; if it is merely an on-chain experiment with a strong burn sink, the ceiling will be considerably lower.
Regardless of the outcome, IMD is already one of the most closely watched AI experiments in the Ethereum ecosystem at year-end. If activity continues to rise and market momentum persists, it may become one of the trading targets worth tracking in 2026.
