# Summary Treasure has launched an initiative that has been several months in production called the Autonomous Flywheel. The Flywheel is a series of protocols spanning DeFi, Social, and Consumer powered by autonomous AI. This proposal establishes the fee policy for these dapps. During current market conditions, revenue across Flywheel dapps will be converted into MAGIC and burnt with two exceptions: (1) burns will not occur if liquid runway for the DAO is below 18 months in USDC or highly-liquid assets like ETH, (2) the burn will not include the monthly opex costs for the DAO. These conditions ensure operational sustainability for the DAO. With a solid liquid runway, the focus should currently shift to boosting the value of the treasury through a programmatic buy-and-burn strategy. In a future proposal, token holders can decide to use earned fees to buy back MAGIC for a vested ecosystem fund if the price of MAGIC has recovered to a level where additional burns would be wasteful in an opportunity cost sense. # Proposal This proposal authorizes Treasure contributors to engage in the following actions with accumulated fees: - AI Frens Solana: this version of the platform will allow users to create 3D virtual characters through image uploads and text descriptions. Each character will have its own stream and chat interactivity immediately after launch. These AI Frens coins launch onto Pump.Fun, with TreasureDAO claiming the creator fees for each coin. Accumulated SOL will be converted to MAGIC and burnt by transfer to the zero address, with buy-burns reported in quarterly reports. - wizAI: this dapp is an X native token launcher for presales and NFT mints. No wallets required, built using autonomous AI. wizAI generates fees on the presale (1% of raised ETH) and 0.3% swap fees in Uniswap. Treasure will use raised ETH to buy and send MAGIC to the zero address. https://www.wizai.lol/ Both protocols have operational costs, specifically the Twitter API monthly cost and OpenRouter costs to finance chat and virtual character streaming. Burns will take place the month following revenue generation, first using accumulated fees to cover the operational costs of the programs then burning the remainder. Additional protocols will be released in coming weeks: Locket Proposal (options protocol with inventory management powered by autonomous AI); Halo Finance (ERC-8004 focused lending protocol). This proposal authorizes Treasure contributors to apply the same fee structure to each Autonomous Flywheel component. All protocol revenue, minus the operational cost of deployment and maintenance, will be burnt. ## Amendments from Comment Period - The first two flywheel dapps have since launched - A commenter observed that the buy-and-burn policy should not include monthly opex costs so as to remain sustainable. We agree and added to this idea that burns should not occur unless 18 months of stable runway is secured - The comment proposal discussed NEAR Chain Signatures for Wiz AI. The final implementation used Phala. NEAR will be used for future dapps but is not currently used