{
  "txid": "a0c1e1764d4c8376090211ad128499ab66cb4e81d74f531faf1dde21c309ee3e",
  "block_height": 0,
  "time": null,
  "app": "treechat",
  "type": "post",
  "map_content": "If you were ever curious about $TAO Subnets and how they work, read this post.\r\n\ud83d\udca5Bittensor $TAO Subnets: The Creation and Dissolution of a Decentralized Future of Machine Learning \ud83d\udca5\r\nThe launch of the Bittensor network has changed how we build, test, and distribute machine intelligence, moving from centralized corporate AI to a performance rewarded marketplace.\r\nAt the foundation of the $TAO ecosystem are subnets, which are self contained AI systems that produce specific digital commodities, like text generation, image generation, data storage, etc.\r\nSubnets are not just dApps on the Bittensor blockchain, they drive value creation, operating in direct competition where participants are rewarded in the $TAO token for providing verifiable utilities.\r\nSubnets and How They Are Designed\r\nBittensor is an open source protocol that houses a decentralized machine learning ecosystem, allowing the peer to peer exchange of intelligence and predictions.\r\nTo understand subnets, you have to look at it as a digital intelligence marketplace.\r\nBitcoin introduced a decentralized currency and Ethereum introduced decentralized smart contracts, but Bittensor introduces decentralized markets for digital commodities.\r\nThese markets are organized into individual subnets, each specializing in a specific task and governed by its own incentive mechanism.\r\nThe Four Roles in the Subnet Ecosystem\r\nThere are four roles in the subnet ecosystem that together determine the stability and productivity of a subnet. Subnet creators, miners, validators, and stakers\r\nThe creator designs, and manages the incentive mechanism and receives a portion of the subnet's emissions (daily $TAO distribution).\r\nThe miner produces digital commodities through computational work, such as training models, and earns rewards based on the performance of its output.\r\nThe validators evaluate miners' quality and weights it based on performance. They receive dividends based on a consensus with other validators or receive a penalty if they are too far outlying from the average.\r\nThe stakers are regular users. They stake their $TAO to subnets they want to support or see potential in, and receive a share of validator rewards based on each subnets APY percentage, which can fluctuate.\r\nThese roles combined create a competitive environment where miners are under constant pressure to optimize their models, and validators are incentivized to remain honest and accurate to maximize their consensus based rewards.\r\nUnderperforming models or dishonest actors are gradually phased out by the system's economic feedback loops i.e. clipping of outliers.\r\nArchitectural Layers and the Role of Subtensor\r\nThe Bittensor network operates in two layers.\r\nThe Subtensor blockchain,\r\nand,\r\nThe subnet layer.\r\nSubtensor acts as the ledger for the network, using the $TAO token to reward honest participation.\r\nWhile the heavy computational work is occurring off chain within the subnets, Subtensor performs several functions.\r\nIt logs the summaries of subnet activities, to ensure fair and transparent contributions.\r\nUsing the Yuma Consensus algorithm, Subtensor calculates and distributes $TAO rewards to all participants based on the subnet weighting submitted by validators.\r\nRecently, Ethereum Virtual Machine compatibility was introduced, allowing for complex dApps such as liquid staking, lending, and borrowing to exist on the Subtensor layer.\r\nIf you view Bittensor as a global, decentralized computer, the subnets represent specialized drives or regions, each dedicated to its specific cognitive or computational task.\r\nThis prevents the mixing of incomparable tasks.\r\nFor instance, a model doing image recognition is not evaluated against the same criteria as a model doing financial market prediction.\r\nMaking sure evaluations are task appropriate and that the network can adapt to new technological use cases without wrecking the entire protocol. A fish can't be compared to a cat on its ability to climb a tree.\r\nYuma Consensus Model\r\nThe biggest innovation of Bittensor is the Yuma Consensus mechanism or Proof of Intelligence.\r\nIn most blockchain networks, consensus is reached based on the validity of transactions such as double spending prevention.\r\nIn Bittensor, consensus is reached on the utility and quality of machine intelligence, hence Proof of Intelligence.\r\ndTAO and Alpha Tokens\r\nThe original distribution of emissions used to be determined by the Root Network (Subnet 0), which consisted of the 64 largest validators. (You can still stake into and receive a safe APY with no risk to your tokens.)\r\nHowever, the introduction of Dynamic $TAO or dTAO in February 2025 switched distribution to a new market driven approach.\r\nWith dTAO, each subnet operates as its own Automated Market Maker or AMM.\r\nEach subnet has its own currency, called an Alpha token.\r\nSubnets maintain a $TAO reserve representing the total stake and an Alpha reserve representing the subnet's own supply similar to a liquidity pool.\r\nThe exchange rate between $TAO and a subnet's Alpha token is determined by the ratio of the reserves, $TAO reserves vs Alpha Reserves.\r\nWhen a holder stakes $TAO into a subnet, they are purchasing Alpha tokens and adding $TAO to that subnet's liquidity pool. This counts as a vote for the subnet's value, telling the network that the subnet's commodity is in demand, but, the higher the demand, the more the rewards in APY are diluted.\r\nThe biggest change to staking on $TAO subnets versus other chains is that there's no minimum lock time or unstaking period. You can stake and unstake at will.\r\nFlow Based Emissions\r\nThe share of global $TAO emissions that a subnet receives is now determined by net $TAO flows, the difference between the amount of $TAO staked and unstaked in that subnet. Inflow vs outflows.\r\nSubnets that attract high user engagement and staking receive increased emissions, while those experiencing net outflows see their funding reduced or eliminated.\r\nThis loop forces subnet creators to maintain high project quality to attract stakers, while stakers are incentivized to identify and fund the most promising AI technologies early in their lifecycle.\r\nSubnet Creation\r\nCreating a subnet on the Bittensor network is a huge technical and financial undertaking, requiring a deep understanding of the Bittensor SDK, the CLI, and the specific incentive logic intended for the target commodity.\r\nBefore a developer can register a new subnet, they must establish a robust local infrastructure to ensure the subnet's initial stability.\r\nDevelopers must run a public Subtensor node synchronized with either the testnet or the mainnet. Lite nodes are often sufficient for basic interaction, while archive nodes are necessary for developers requiring full blockchain history.\r\nThree wallets are required for the initial setup. The subnet creator wallet, a validator wallet, and a miner wallet.\r\nRegistering a subnet requires a large balance of $TAO to cover the burn cost and fund the liquidity pool.\r\nRegistration is a competitive process, and is done through the Bittensor CLI.\r\nThe cost of registration, aka the recycle register or burn cost, doubles every time a subnet is registered but drops over time as blocks progress.\r\nAfter successful registration, the user is prompted to burn the required amount of $TAO.\r\nOnce the transaction is finalized, the subnet is assigned a unique ID.\r\nThe One Week Inactive Phase\r\nTo prevent the claiming of rewards before a subnet is fully functional, new subnets must wait in an inactive state for 50,400 blocks or roughly seven days, during which the subnet creator must:\r\nRecruit and register validators to ensure the work can be evaluated once emissions start.\r\nInvite high performing miners to join the subnet to begin producing the commodity.\r\nSet the operational limits such as tempo, immunity period using the sudo_set commands.\r\nThe incentive mechanism is the core logic of the subnet, maintained off chain in a code repository like GitHub. This code defines how miners and validators communicate and how work is scored.\r\nThe Deregistration and Pruning Lifecycle\r\nThe Bittensor network maintains a strict limit of 128 active subnets. When this limit is reached and a new applicant wants to register, the system WILL deregister an existing subnet to open a slot.\r\nPruning\r\nThe selection of a subnet for deregistration or pruning, is based on market performance.\r\nThe system identifies the subnet with the lowest Exponential Moving Average of its Alpha token price, as long as it has also finished its initial four month immunity period.\r\nThe EMA price calculation uses a standard formula that takes into account the subnet's age and historical price action.\r\nThis ensures that subnets are judged based on sustained market value rather than temporary price volatility.\r\nLiquidation\r\nUpon deregistration, the subnet's lifecycle ends with a standard liquidation process:\r\nAll liquidity in the subnet's Alpha vs $TAO AMM is dissolved.\r\nAll Alpha tokens associated with the subnet are permanently destroyed.\r\nThe $TAO remaining in the subnet's reserve is distributed proportionally to all Alpha token holders. These tokens are credited directly to the holders' balances as unstaked $TAO based on the Alpha tokens price upon dissolution.\r\nOwners of subnets registered after October 2025 do not receive a refund of their burn cost, since those tokens were already removed from circulation.\r\nHowever, they retain any owner emissions accumulated during the subnet's lifespan.\r\nChallenges of Decentralization\r\nAlthough its architecture is robust, the subnet model faces ongoing challenges that creators must navigate.\r\nShallow liquidity in new subnet pools can lead to large price swings, creating risk for long term stakers.\r\nRetail investors tend to gravitate toward highly marketed top subnets, potentially neglecting niche but superior projects.\r\nWhile Yuma Consensus is adversarially resilient, deals between owners, validators, and miners can still form to manipulate market demand.\r\nSubnet creators and the OpenTensor Foundation continue to work to address these through protocol upgrades such as Yuma Consensus 3 and dTAO, ensuring that Bittensor remains a transparent, auditable, and truly decentralized engine for the future of machine intelligence.\r\nBy mastering the technical, mathematical, and economic requirements of Bittensors subnet ecosystem as well as subnet creation and dissolution, developers can contribute to an ecosystem that will democratize and decentralize access to the most powerful technology of our lifetime.\r\nWritten and dictated by Mattyverse in conjunction with, and polished by AI.",
  "media_type": "text/markdown",
  "filename": "|",
  "author": "14aqJ2hMtENYJVCJaekcrqi12fiZJzoWGK",
  "display_name": "MattyV",
  "channel": null,
  "parent_txid": null,
  "ref_txid": null,
  "tags": null,
  "reply_count": 2,
  "like_count": 13,
  "timestamp": "2026-02-20T00:39:08.000Z",
  "media_url": null,
  "aip_verified": true,
  "has_access": true,
  "attachments": [],
  "ui_name": "MattyV",
  "ui_display_name": "MattyV",
  "ui_handle": "MattyV",
  "ui_display_raw": "MattyV",
  "ui_signer": "14aqJ2hMtENYJVCJaekcrqi12fiZJzoWGK",
  "ref_ui_name": "unknown",
  "ref_ui_signer": "unknown"
}