Author: Chloe Zheng

According to a 2023 study by Sequoia Capital, 85% of developers prefer to fine-tune existing models rather than train from scratch. Recent trends have further validated this point: DeepSeek has open-sourced its model and introduced model distillation technology, transferring inference logic from a teacher model (large model) to a student model (small model) to optimize knowledge compression and performance retention. Similarly, OpenAI's ChatGPT O3 version also emphasizes fine-tuning and reinforcement learning. Bittensor provides an open, decentralized platform that supports the collaboration and sharing of AI models. In July 2024, Bittensor and Cerebras released the BTLM-3b-8k open-source Large Language Model (LLM), which has garnered over 16,000 downloads on Hugging Face, showcasing Bittensor's technical capabilities.

Although Bittensor was launched in 2021, it was almost absent during the AI Agent craze in Q4 2024, and the token price has remained stagnant. On February 13, 2025, Bittensor unveiled the dTAO upgrade, aimed at optimizing token issuance, increasing fairness, and enhancing liquidity. This change is similar to Virtuals Protocol launching the AI Agent LaunchPad, which led to a 50x increase in the market value of $VIRTUAL in 2024.

The report "The Evolution of dTAO and Bittensor: Reshaping the Open-Source AI Ecosystem through Market-Driven Incentives" delves into the impact of the dTAO upgrade completed on February 13 on the Bittensor ecosystem, focusing on its architectural innovation, economic model, and overall ecosystem dynamics.

The number of accounts in the Bittensor system has increased by 100%, growing from 100,000 in early 2024 to nearly 200,000.

1. Bittensor's Basic Architecture

The Bittensor system consists of the following three main modules:

1. Subtensor Parallel Chain and Its EVM Compatible Layer (tao evm): Subtensor is a Layer1 blockchain based on Polkadot's Substrate SDK, responsible for managing the blockchain layer of the Bittensor network. Its EVM compatible layer (tao evm) allows developers to deploy and run Ethereum smart contracts on the network, enhancing the system's scalability and compatibility. The Subtensor blockchain produces a block every 12 seconds, with each block generating a TAO token. Additionally, Subtensor records key activities in the subnet, including validator scoring weights and staked token amounts. Every 360 blocks (approximately 72 minutes), 64 subnets' token emissions are calculated through the Yuma consensus algorithm.

2. Subnets: The Bittensor network consists of 64 subnets, each focusing on a specific type of AI model or application scenario. This modular structure enhances the network's efficiency and performance, promoting the specialized development of different AI models. The incentive mechanism for each subnet is determined by the subnet owner, deciding how tokens are allocated between miners and validators. For example, Subnet 1 is operated by the Opentensor Foundation, with a focus on Text Prompting tasks. In this subnet, validators provide prompts similar to ChatGPT, miners respond to the prompts, validators rank the quality of miner responses, update weights regularly, and upload them to the Subtensor blockchain. The blockchain performs a Yuma consensus calculation every 360 blocks and releases tokens for the subnet.

3. Root Subnet: Serving as the core of the network, the root subnet is responsible for coordinating and managing the operation of all subnets, ensuring the overall coherence and stability of the network.

In addition, the Bittensor API serves as the conduit between subnet validators and the Yuma consensus on the Subtensor blockchain. Validators within the same subnet only connect to miners within the same subnet, and validators and miners from different subnets do not communicate or connect with each other.

This architectural design enables Bittensor to effectively integrate blockchain technology with artificial intelligence, creating a decentralized and efficient AI ecosystem.

The Subtensor EVM compatibility layer, tao evm, officially launched on December 30, 2024, eliminating the need to modify any Ethereum smart contracts to deploy and interact on the Subtensor blockchain. All EVM operations occur exclusively on the Subtensor blockchain without interacting with Ethereum. This means that smart contracts on Bittensor are confined to the Bittensor network and are independent of the Ethereum mainnet. Currently, tao evm is in a relatively early stage, with ecosystem project TaoFi planning to develop AI-based DeFi infrastructure, including the first stablecoin supported by TAO, a decentralized exchange, and a TAO token liquidity staking version.

1.1 Account System

1.1.1 Coldkey-Hotkey Dual Key System

The account system of dTAO adopts a Coldkey-Hotkey dual-key mechanism to ensure higher security and flexibility. When users create a wallet, they can choose to generate the wallet through a Chrome extension or locally. The wallet created through the Chrome extension is used for storing, sending, and receiving TAO. The system will generate a coldkey (48 characters, usually starting with 5) and a 12-word mnemonic. The wallet created locally will generate a hotkey in addition to the coldkey. The hotkey is used for participating in subnet creation, mining, and validation operations.

The main reason for adopting the Coldkey-Hotkey dual-key system is that the hotkey is frequently used in the daily operation of the subnet and faces potential security threats, while the coldkey is mainly used for storing and transferring TAO, effectively reducing the risk of TAO loss. This dual protection mechanism ensures the security and flexibility of account operations.

In terms of binding relationships, one hotkey can be bound to one coldkey of the same subnet, but it can also be bound to a coldkey of a different subnet (not recommended). One coldkey can be bound to multiple hotkeys.

1.1.2 Subnet UID System

1.1.2.1 Subnet UID Generation

After paying a registration fee of at least 100 TAO, the system will generate a Subnet UID and bind it to your hotkey. This UID is a necessary credential for participating in subnet mining or validation. To become a miner, you only need to have a hotkey, coldkey, and Subnet UID, and then run Bittensor to participate in mining.

1.1.2.2 Requirements to Become a Validator

To become a subnet validator, you must stake at least 1000 TAO, and in each subnet, the staking amount must rank in the top 64. It is important to note that validators can hold multiple UID slots simultaneously, allowing them to validate in multiple subnets without requiring additional stake (similar to the concept of restaking). This mechanism not only reduces the risk of validator misconduct but also increases the cost of misconduct, as staking a large amount of TAO (at least 1000 TAO) significantly raises the cost of misconduct. To enhance their competitiveness in the subnet, each validator strives to build a good reputation and performance record to attract more TAO delegated stakes, ensuring a stable position in the top 64.

1.1.2.3 Subnet Structure and Capacity Limit

Subnet 1: There are a total of 1024 UID slots in this subnet, with a maximum of 128 validators; the total number of validators and miners is capped at 1024.

Other Subnets: Each subnet has 256 UID slots, accommodating up to 64 validators; the total number of validators and miners in each subnet does not exceed 256.

1.1.2.4 Subnet Competition and Incentive Mechanism

Within each subnet, validators allocate tasks to miners, and after all miners complete their tasks, they submit the results to the respective validators. Validators evaluate and rank the quality of tasks submitted by each miner, and miners receive TAO rewards based on the quality of their work. At the same time, validators are incentivized to ensure that high-quality miners receive better rewards, thus driving continuous improvement in the overall subnet quality. This series of competitive processes is automatically executed by the subnet creator's designed code incentive mechanism, ensuring the fair and efficient operation of the system.

Each subnet has a 7-day protection period (immunity period), starting from when a miner registers the subnet UID. During this period, miners accumulate their rewards. If new miners register within the protection period and the current subnet's UID slots are full, the miner with the lowest cumulative reward will be eliminated, and their UID will be reassigned to the newly registered miner.

1.2 Subnet Constructs a Multilayer Ecosystem

The Bittensor subnet has built a multilayered ecosystem where miners, validators, subnet creators, and consumers each play their roles in driving the generation of high-quality AI services.

Miners: As the core computing nodes of the network, miners host AI models and provide inference and training services. They compete in peer-to-peer ratings by minimizing the loss function to earn TAO rewards. A miner's success depends on the quality and performance of the services they provide.

Validators: Validators are responsible for evaluating the task results submitted by miners, building a trust matrix, preventing collusion and cheating, and ensuring that high-quality miners receive higher rewards. They rank based on the quality of miners' responses, with more accurate and consistent rankings leading to greater rewards for validators.

Subnet Creators: Subnet Creators design custom subnets according to the specific needs of application domains (such as natural language processing, computer vision), building independent consensus mechanisms, task flows, and incentive structures. They take on the role of network administrators, with the authority to allocate rewards through their respective subnets.

Consumers: Consumers invoke AI services by paying TAO tokens to perform tasks such as querying APIs, accessing training data, or utilizing computing resources for model training purposes. They are the end users of AI models provided by Bittensor.

The overall process is as follows: Subnet validators generate a problem and distribute it to all miners. Miners generate answers to the task and return them to the validators. Validators score the answers based on quality and update miner weights, then regularly upload the weights to the chain. Through intense competition and a survival-of-the-fittest mechanism, continual progress in AI model technology and ecosystem optimization within the subnet is driven.

1.2.1 Miner Layer

Miners in the Bittensor network play a crucial role as core computing nodes, with their main responsibilities including:

Hosting AI models, providing inference or training services: Miners host local machine learning models to offer prediction services to client applications. When a client needs a prediction, they send a request to the Bittensor network, which routes it to a miner registered as a service provider. The miner processes the request and returns the prediction results to the client.

Earning TAO tokens as computational incentives by competing in P2P rankings: Miners compete in peer-to-peer rankings based on their model's performance and contribution to the network to earn TAO token rewards. This incentive mechanism encourages miners to continuously optimize their model performance to provide high-quality AI services to the network.

Ensuring high-quality AI model contributions: Miners strive to provide high-quality AI models to meet network demands and ensure service quality. This not only helps them achieve higher rankings and rewards in the network but also enhances the overall performance and reliability of the entire Bittensor network.

By fulfilling these responsibilities, miners make significant contributions to the efficient operation and development of the Bittensor network.

1.2.2 Consumer Layer

In the Bittensor network, a consumer refers to end users or businesses who access artificial intelligence (AI) services provided by miners by paying TAO tokens. This model allows consumers to access AI capabilities within the network without owning or maintaining their own AI models, thus reducing AI computing costs.

Specific use cases for consumers include:

· Developer Querying AI API: Developers can invoke the AI interface provided by Bittensor to obtain the required intelligent services for application development or feature integration.

· Research Institutions Accessing AI Training Datasets: Research institutions can leverage resources within the network to access and analyze large AI training datasets to support research projects and experiments.

· Enterprises Utilizing Bittensor's Compute Resources for AI Model Training: Enterprises can utilize Bittensor's decentralized computing resources to train and optimize their proprietary AI models, enhancing their business's level of intelligence.

Through this approach, Bittensor provides consumers with a flexible and efficient path to access AI services, promoting the dissemination and application of artificial intelligence.

1.2.3 Stake-based Consensus Mechanism

Bittensor's stake-based consensus mechanism primarily addresses the following issues:

1.2.3.1 Smooth Handling and Density Evolution

To avoid abrupt adjustments leading to system instability, the adjustment function utilizes "smooth handling." We define the stake-weighted mean absolute deviation as:

This process is akin to statistical smoothing on a large dataset. After multiple rounds of "smoothing," the true ranking of each participant is revealed. Importantly, density evolution can significantly compress outlier weights (i.e., excessively high weights from malicious players) while minimally affecting honest players.

1.2.3.2 Weight Trust Mechanism and Zero-weight Vulnerability Prevention

dtao Upgrade has been made to address the above issues, including:

· Optimization of Iteration and Smoothing: Increase iteration count η, fine-tune the smoothing parameter α or δ to reduce zero-weight vulnerabilities, and prevent overcorrection.

· Enhanced Weight Trust Mechanism: More accurately detect non-zero weights and apply stricter thresholds so that only nodes recognized by the majority can receive full rewards.

· Reduced Computation Overhead: By algorithmic optimization to reduce computing costs, making it adaptable to blockchain computational constraints without compromising theoretical accuracy.

Bittensor's Proof-of-Stake consensus mechanism combines mathematical models and game theory tools. Through updating formulas, weighted average consensus, iterative correction, density evolution, and other methods, the system can automatically calibrate abnormal weight deviations to ensure fair final reward distribution.

This process is similar to an intelligent balancing system or a reputation mechanism, capable of continuous self-calibration to ensure fair scoring, incentivize excellent contributors, and prevent malicious collusion and vote manipulation.

Building on this foundation, the dtao Upgrade adopts finer smoothing control and an improved weight trust strategy, further enhancing the system's robustness and fairness. Therefore, in adversarial environments, honest contributors can always maintain a competitive advantage, and overall computing resource consumption is also optimized and reduced.

2. Yuma Consensus: Dynamic Programmable Incentives and Consensus

Bitcoin has built the world's largest peer-to-peer computational network, where anyone can maintain a global ledger through contributing local computing power. Its incentive rules were fixed at the design phase, leading to ecosystem development in a relatively static manner.

In contrast, Yuma Consensus (YC) is a dynamic, programmable incentive framework. Unlike Bitcoin's static incentive mechanism, YC integrates the objective function, staking rewards, and weight adjustment mechanism directly into the consensus process. This means that the system does not rely solely on fixed rules but dynamically adjusts based on the actual contributions and behaviors of nodes, thereby achieving a more equitable and efficient reward distribution.

YC consensus algorithm runs continuously on the Subtensor blockchain and operates independently for each subnet. Its main workflow includes the following components:

· Subnet Validator's Weight Vector: Each subnet validator maintains a weight vector where each element represents the validator's rating weight assigned to all subnet miners. This weight is based on the validator's historical performance and is used to rank miners. For example, if a validator's rating vector is , the resulting ranking reflects the validator's assessment of each miner's contribution level.

· Staking Amount Impact: Every on-chain validator and miner stakes a certain amount of tokens. YC consensus combines the weight vector and staking amount to calculate reward distribution. In other words, the final reward depends not only on the rating weight but also on the staked amount, forming a "Staking → Weight → Reward" loop.

· Dynamic Subjective Consensus: Each participant assigns a local weight to their machine learning model. These local weights are adjusted through a consensus strategy and then aggregated on the blockchain as a global metric. In essence, YC is able to achieve large-scale consensus even in adversarial environments and dynamically adapt to changes in node behavior.

· Reward Computation and Distribution: Subnet validators collect their respective ranking results and submit them as collective input to the YC algorithm. Although different validators' rankings may arrive at different times, Subtensor processes all ranking data approximately every 12 seconds. Based on this data, the system calculates rewards (in TAO) and deposits them into the wallets of subnet miners and validators.

This comprehensive mechanism enables YC to continuously and fairly distribute rewards in a decentralized network, dynamically adapt to the quality of contributions, and maintain the overall network's security and efficiency.

2.1 Knowledge Distillation and Expert Mixing (MoE): Collaborative Learning and Efficient Contribution Evaluation

2.1.1 Knowledge Distillation (Digital Hivemind)

Bittensor introduces the concept of knowledge distillation, similar to the collaborative work of neurons in the human brain, where nodes collectively learn by sharing knowledge, exchanging data samples, and model parameters.

During this process, nodes continuously exchange data and model parameters, forming a network that self-optimizes over time to achieve more accurate predictions. Each node contributes its knowledge to a shared pool, ultimately enhancing the overall performance of the entire network, making it faster and more suitable for real-time learning applications such as robotics and autonomous driving.

Importantly, this approach effectively mitigates the risk of catastrophic forgetting, a common challenge in machine learning. Nodes can integrate new insights while retaining and expanding their existing knowledge, thereby enhancing the network's robustness and adaptability.

By distributing knowledge across multiple nodes, the Bittensor TAO network becomes more resilient to interference and potential data leaks. This robustness is particularly important for applications handling highly secure and privacy-sensitive data, such as financial and medical information.

2.1.2 Expert Model Mixing (MoE)

Bittensor employs a Distributed Expert Model (MoE) to optimize artificial intelligence predictions. By leveraging the collaborative work of multiple specialized AI models, it significantly enhances the accuracy and efficiency of solving complex problems. For example, when generating Python code annotated in Spanish, a multilingual model working in concert with a code-specialized model can produce much higher-quality results than a single model.

The core of the Bittensor protocol consists of parameterized functions, commonly known as neurons, distributed peer-to-peer. Each neuron records zero or more network weights and evaluates the value of neighboring nodes through mutual ranking to train the neural network, thus cumulating the ranking scores onto the digital ledger. Nodes with higher rankings not only receive monetary rewards but also gain additional weight, establishing a direct link between node contribution and rewards, thereby enhancing the network's fairness and transparency. This mechanism builds a marketplace enabling other intelligence systems to price information over the internet in a peer-to-peer manner and incentivizes nodes to continually enhance their knowledge and expertise. To ensure a fair reward distribution, Bittensor borrows the Shapley value from cooperative game theory, providing an efficient method to allocate rewards among parties based on node contributions. Under the YC consensus, validators rate and rank various expert models and allocate rewards fairly based on the Shapley value principle, further enhancing the network's security, efficiency, and continuous improvement capabilities.

3.dtao Upgrade

The Bittensor project faces the following key issues in its resource allocation and economic model design:

1. Resource Overlap and Redundancy: Multiple subnets focus on similar tasks, such as text-to-image generation, text prompting, and price prediction, leading to duplication and waste in resource allocation.

2. Lack of Real-World Use Cases: Some subnets (such as price prediction or sports event outcome prediction) have not yet proven their practicality in real-world scenarios, which may result in a mismatch between resource allocation and actual demand.

3. "Gresham's Law" Phenomenon: High-quality subnets may struggle to obtain sufficient funding and development opportunities. Due to only having a seven-day protection period, subnets that fail to garner adequate support from root validators may be eliminated prematurely.

4. Validator Centralization and Insufficient New Subnet Incentives:

· Root validators may not fully represent all TAO holders, and their assessments may not reflect a broad range of perspectives. Under the Yuma consensus, top validators hold a dominant position in the final scoring, but their evaluations are not always objective. Even if biases are identified, they may not be corrected immediately.

· Additionally, validators lack incentives to migrate to new subnets because transitioning from a high issuance old subnet to a low issuance new subnet may result in an immediate loss of rewards. The uncertainty of whether new subnets can eventually match the established subnets' token issuance further reduces their migration willingness.

Main Issues with the Economic Model:

One key problem in Bittensor's mechanism design is that although all participants receive TAO, in reality, no one pays TAO, leading to sustained selling pressure. Currently, the questions miners answer are not posed by real users but rather provided by subnet owners—either simulating real user queries or based on historical user demands. Thus, even if miners' answers are valuable, this value is captured by subnet owners. Whether the miners' answers help subnet owners improve their model algorithms or are directly used by subnet owners for model training to enhance their products, the value generated by miners and validators is owned by subnet owners. In theory, subnet owners should pay for this value.

Furthermore, subnet owners not only incur no costs but also enjoy 18% of the subnet's token issuance. This means that the Bittensor ecosystem is not tightly interconnected—participants maintain loose connections based on development and collaboration. Projects on the subnet can exit at any time without incurring any losses (as subnet registration fees are refunded). Currently, the primary mechanism for token recapture in the Bittensor system is the registration fees paid by subnet miners and validators; however, these fees are minimal and insufficient to support effective value capture. Despite staking becoming a primary mechanism, the amount of TAO reclaimed through blockchain transaction fees and registration fee refunds is still limited.

Staking comes in two forms:

1. Validator Staking: Participants stake TAO to support network security and earn rewards, accounting for 75% of all issued TAO. Validators currently receive approximately 3,000 TAO per day, with an annualized return rate of over 15%. However, after the first halving, this allocation will decrease to 1,500 TAO per day, reducing the attractiveness of staking, weakening its effect on balancing token supply and demand.

2. Subnet Registration Staking: The addition of new subnets significantly impacts TAO supply. This poses a challenge as the total issuance of TAO is fixed; the increase in the number of subnets will dilute the rewards of all subnets, making it difficult for existing subnets to sustain operations and potentially causing some subnets to exit the network.

These issues indicate that Bittensor's resource allocation and economic model design need further optimization to ensure the network's sustainable development and fair incentives.

3.1 What is dTAO

dTAO is an innovative incentive mechanism proposed by the Bittensor network to address the inefficiency of resource allocation in decentralized networks. It abandons the traditional manual voting by validators to determine resource allocation and instead introduces a mechanism based on market dynamics adjustment. It directly links the distribution of TAO issuance among subnets to the market performance of subnet tokens. Through an embedded liquidity pool design, it encourages users to stake TAO in exchange for subnet tokens to support high-performing subnets.

Simultaneously, adopting a fair distribution model ensures gradual allocation of subnet tokens, incentivizes teams to receive token shares through long-term contributions, and balances the roles of validators and users. Validators rigorously evaluate team technology and market potential like venture capitalists, while users further drive the formation of subnet value through staking and market trading.

3.1.1 Core Mechanism of dTAO

3.1.1.1 Strong Binding of Validators and Teams to the Ecosystem: Investment in subnet tokens is required to earn rewards

dTAO's design is based on dual drivers of market and technology, with each subnet having a liquidity pool consisting of TAO and subnet tokens. When $TAO holders (validators and subnet owners) stake, they effectively use $TAO to purchase corresponding dTAO, which can be exchanged for a certain amount of dTAO following the formula:

Unlike Uniswap V2, the $dTAO liquidity pool does not allow direct liquidity addition. Except when the Subnet Owner creates a Subnet, all newly injected liquidity comes entirely from the allocated $TAO and 50% of the total $dTAO minted. In other words, the newly minted $TAO allocated to each Subnet is not directly distributed to that Subnet's Validator\Miner\Owner; instead, it is all injected into the liquidity pool for backing. Simultaneously, 50% of the newly minted $dTAO is also injected into the liquidity pool, while the remaining 50% is distributed to the Validator\Miner\Owner of the Subnet according to the Subnet's own agreed-upon incentive mechanism.

This approach prevents teams from conducting rapid sell-offs through an initial large coin holding, encouraging ongoing contributions and technical iterations by the team. Validators are required to play a role similar to that of a venture capitalist, rigorously assessing the technology, market potential, and actual performance of the subnet.

Staking\Unstaking will not change the size of K, while liquidity injection will increase K to K'.

3.1.1.2 The Subnet Token with the highest market price will receive the most $TAO emissions

In the previous scheme, the proportion of additional $TAO emissions that each Subnet could receive was determined by the Root Network's Validator. This scheme exposed some potential issues. For example, due to the concentration of power in the hands of a few Validators in the Root Network, even if Validators conspired to allocate the additional $TAO emissions to low-value Subnets, they would not face any punishment.

The Dynamic TAO eliminates the privilege of the Root Network and transfers the power to decide how the additional $TAO emissions should be allocated to all $TAO holders. The specific approach is to adopt the brand-new Yuma Consensus V2, which performs a softmax operation on the prices of each Subnet Token to obtain the corresponding release proportion, as follows:

Softmax is a commonly used normalization function that converts each element in a vector to a non-negative value while preserving the relative size relationships between elements and ensuring that the sum of all elements after conversion is 1.

Where P is the price of $dTAO relative to $TAO, calculated as the amount of $TAO in the liquidity pool divided by the amount of $dTAO.

According to the formula, the higher the price of a Subnet Token relative to $TAO, the higher the percentage of newly minted $TAO that can be released.

3.1.1.3 Delegating Incentive Mechanism Setting Power to Each Subnet

Previously, the $TAO incentives received by Subnets were distributed to Validator/Miner/Owner at a fixed ratio of 41%-41%-18%.

Dynamic TAO gives each Subnet the power to issue its own "Subnet Token" and specifies that, in addition to 50% of the inflationary supply being injected into the liquidity pool, the remaining 50% will be distributed to Validator/Miner/Owner based on a mechanism decided by the Subnet participants themselves.

This mechanism simultaneously ensures that only subnets that continuously improve their product and attract users can receive more incentives, preventing a Ponzi-like short-term profit-driven pattern.

3.1.2 Example Analysis

After the Dynamic TAO network upgrade, all Subnets have now minted their respective $dTAO, where the genesis quantity of $dTAO equals the amount of $TAO locked by the Subnet Owner when creating the Subnet. 50% of the $dTAO is injected into the Subnet's liquidity pool, while the remaining 50% is allocated to the Subnet Owner.

Assume Subnet #1's Owner had locked 1000 $TAO, then the genesis quantity of $dTAO is also 1000. Of this, 500 $dTAO and 1000 $TAO are added to the liquidity pool as initial liquidity, and the remaining 500 $dTAO is allocated to the Owner.

Next, when a Validator comes to Subnet #1, registers, and stakes 1000 $TAO, the Validator will receive 250 $dTAO, leaving 2000 $TAO and 250 $dTAO in the liquidity pool.

Assuming Subnet #1 receives 720 $TAO block rewards per day, the liquidity pool will automatically receive 720 $TAO per day. As for the daily amount of $dTAO injected, it depends on the inflation rate set by the Subnet itself.

3.2 Impact of dTAO

The introduction of dTAO fundamentally reshapes the distribution and staking mechanism of TAO. Firstly, the newly minted TAO is no longer allocated solely by a few Validators, but rather determined indirectly by all TAO holders through market behavior, making staking TAO more akin to "buying into" a Subnet's Token rather than a simple dividend. Under this mechanism, the short-term impact of staking and unstaking on the price of dTAO far outweighs the effect of the actual TAO received by the Subnet, introducing uncertainty to staking rewards.

The benefit is that top Validators lose absolute control over block reward distribution, significantly raising the cost for potential attackers to mount a staking attack on the network. Additionally, newly emerged high-quality Subnets have a greater opportunity to stand out, with early Validator support offering the potential for significant returns, possibly even multiples of the principal. Furthermore, intensified competition among Subnets will drive stakers to become more rational investors, meticulously selecting the most promising Subnets through thorough due diligence.

Overall, the implementation of the dTAO mechanism will propel the entire ecosystem towards a more efficient, competitive, and market-oriented direction.

3.3 How Will the Bittensor Ecosystem Evolve After the dTAO Upgrade?

To analyze the impact of the dTAO upgrade, we need to focus on two key questions:

1. How will Subnet demand transform into demand for Subnet tokens?

2. Can the introduction of Subnet tokens create a "Summer of TAO" to accelerate innovation within the TAO ecosystem?

3.3.1 How Will Subnet Demand Transform into Demand for Subnet Tokens?

Initially, all Subnet tokens have the same price, and each Subnet's liquidity pool contains only a small amount of TAO and dTAO tokens. Therefore, any trading activity could lead to significant price fluctuations.

To participate in a Subnet and receive rewards, users must first purchase dTAO Subnet tokens and stake them to Validators, driving up the price of dTAO within that Subnet. As the price of dTAO rises, the total value of dTAO in the liquidity pool increases, and the system automatically allocates more TAO rewards to that Subnet, enabling miners and stakers to earn higher returns.

This has created a positive feedback loop: users buy dTAO, driving up the price--price increase leads to the subnet receiving more TAO issuance--more rewards attract additional users to join--further driving up the price of dTAO

Conversely, if users start mass selling dTAO, its price decreases, causing a reduction in the TAO issuance received by the subnet, thereby decreasing user participation. Overall, the fluctuation in subnet token price is mainly influenced by market supply and demand, liquidity pool size, and the system's automatic incentive mechanism.

This mechanism bears similarities to the AI Agent Launchpad model, where users first need to purchase platform tokens to invest in AI agent tokens. In the AI Agent Launchpad ecosystem, once an AI agent token's price rapidly rises, creating a wealth effect, a large number of users flock in, further increasing the demand for platform tokens.

However, there are some key differences between the dTAO mechanism and the AI Agent Launchpad:

· In the AI Agent Launchpad ecosystem, users typically only use platform tokens to purchase these AI agent tokens when the market value of the AI agent tokens is low (i.e., in the project's internal market).

· Once an AI agent token reaches a certain valuation, users can sell it for ETH/SOL to realize profits, and new users can also directly purchase AI agent tokens using ETH/SOL.

In contrast, in the dTAO system:

· When the dTAO price rises, users looking to cash out or migrate to another subnet with higher potential can only exchange dTAO for TAO.

· This process can lead to significant fluctuations in the price of dTAO in the liquidity pool.

Currently, users can trade the dTAO token on Backprop Finance to provide secondary market liquidity for subnet tokens.

3.3.2 Unique Issuance Mechanism of the dTAO Ecosystem

Another key aspect of the dTAO ecosystem is its unique token issuance mechanism. As shown in the diagram below, after the dTAO upgrade, the issuance is highly concentrated in the initial few subnet projects. The first five subnet projects currently receive 40% of the total issuance.

Currently, 7,200 TAO is distributed daily, and based on the TAO price on February 18, 2025, this means that the first five Subnet projects each individually receive approximately $1 million worth of TAO per day.

If the development trajectory of the dTAO ecosystem is similar to the Virtual Ecosystem, where certain projects receive significant market attention, then high-market-cap Subnets will capture the majority of the newly minted TAO supply.

For new projects to stand out in competition, they must demonstrate strong potential to attract stakers, miners, and validators. This usually means:

· Participants need to migrate from other Subnets, exchanging their TAO for the new Subnet's dTAO.

· This may involve selling Subnet tokens in existing liquidity pools to increase the new Subnet's market cap.

This competitive model may encourage more activity in Subnet token markets and further drive innovation and development within the entire TAO ecosystem.

3.4 Does dTAO Solve the Issues in the Bittensor Subnet Model?

3.4.1 Mechanism Issues Still Exist

The dTAO upgrade ties the TAO issuance to the market performance of Subnet tokens, shifting resource allocation decision-making from a few root validators to a market-driven approach aimed at incentivizing broader user participation and interaction. While this mechanism partially alleviates the inefficiencies caused by resource overlap and ensures that only high-performance Subnets with strong token price performance receive more TAO rewards, it does not fundamentally address the following key issues:

· Resource Overlap and Redundancy: If multiple Subnets focus on similar tasks (such as text generation, image generation, or price prediction), even with a market-driven adjustment, resource duplication and underutilization still remain fundamentally unresolved.

· While all participants can earn TAO, there is no external user payment for the contributions of miners and validators. This results in TAO continually facing selling pressure, as rewards are continuously distributed without a sustainable demand mechanism to support the TAO price.

· Some subnets may have issues with a fake model and incomplete evaluation criteria: Bittensor is evolving into an “outsourcing layer” within the AI tech stack, where token incentives rapidly attract resources and drive the allocation of specific AI tasks. For example, Kaito AI outsources search engine development to a subnet, leveraging collective intelligence to reduce costs. However, this incentive-driven model may attract developers in the short term, but long-term success still depends on real demand and quality assurance. When testing the Cortex.t subnet, it was found that its answers came directly from the OpenAI API rather than being generated by Bittensor miners. This indicates that some subnets are simply “packaged apps” and do not fully utilize Bittensor's decentralized AI computing power. Subnet validators rely on OpenAI results for comparison, which may pose centralization risks, while the accuracy of some price prediction subnets is low, making them difficult to apply in practice.

Improvement direction: Enhance usability and transparency:

· Miners should submit intermediate data or proof of work to verify their model training and inference processes, ensuring that the output indeed comes from the Bittensor network rather than an external API.

· Standardized test datasets should be established for different types of subnets (such as prediction models, generative AI models) to conduct benchmark tests.

· Regularly publish benchmark test rankings to promote healthy competition among subnets and improve model quality.

3.4.2 dTAO Still Faces Adoption, Lack of Use Cases, and Decreasing Staking Rate Issues

Currently, dTAO is mainly limited to the Bittensor network and has not yet gained enough adoption in the broader crypto market. While dTAO has introduced EVM compatibility, it has not generated the same level of buzz on social media as the Virtual Ecosystem’s AI Agent token. At the same time, hardly any projects have incorporated dTAO into their core tokenomics, resulting in dTAO still lacking real-world application demand. Currently, purchasing subnet tokens is more like a one-time investment, and when users choose to cash out, it may trigger significant price fluctuations. This issue is particularly evident in AI infrastructure outsourcing subnets, such as Kaito, where the dTAO token is almost unrelated to its core business, causing the token to lack market value support.

However, dTAO still has certain advantages over the AI Agent Launchpad. According to the dTAO economic model, 50% of newly issued dTAO must be injected into the liquidity pool, while the remaining 50% is allocated by subnet participants (including validators, miners, and subnet owners). This mechanism ensures that only subnets that continuously improve the product and attract users can receive more rewards, thereby avoiding the proliferation of low-quality AI agents and driving technological innovation in the TAO ecosystem. However, as the dTAO ecosystem is still in its early stages, its audience has not yet expanded, and there is a lack of large-scale application scenarios, resulting in its market acceptance remaining relatively low.

Currently, the expansion speed of the Bittensor ecosystem has not been able to match the demand for tokenomic growth. According to the latest data, the staking rate of TAO has decreased from a peak of 90% to 71%. This indicates that some token holders lack confidence in the network's long-term incentive mechanism and may turn to other more financially attractive DeFi or AI ecosystem projects.

3.5 Focus on Subnet Projects Closely Integrated with the Bittensor Ecosystem and Having Practical Use Cases

The healthy development of the Bittensor ecosystem depends on its ability to attract and support high-quality subnets. Evaluating a subnet's long-term potential requires a focus on its application scenarios, incentive mechanisms, team background, and the actual use of the token.

Firstly, a subnet must have a clear and practical application scenario. A successful project should not only solve real-world problems but also receive feedback from real users. The technical architecture needs to be robust and innovative, capable of supporting distributed AI model training and inference. In addition, the subnet should leverage on-chain data and adopt a transparent evaluation mechanism to demonstrate its contribution to the Bittensor ecosystem.

Secondly, a reasonable incentive mechanism is key to sustaining the long-term operation of a subnet. The incentive structure should be fairly distributed to miners, validators, and subnet owners, avoiding market sell pressure due to a lack of sustained application demand. Subnets need to be able to self-sustain through a business model, rather than relying solely on TAO issuance for incentives.

Furthermore, a successful subnet project often has a strong team background, ecosystem integration capabilities, and community support. Focusing on Bittensor-native subnets rather than purely AI outsourcing subnets can ensure the long-term stability of the entire ecosystem. For outsourcing projects, the key is whether their subnet token is truly integrated into the core tokenomic model and not just used as an incentive tool.

Finally, the actual use of the subnet token is crucial in determining its long-term value. Currently, almost no project truly incorporates the subnet token into its operational system, and dTAO is still in its early stages. If the subnet token can be used for payments, access to AI services, governance participation, or providing additional incentives, then real market demand can be established to ensure long-term value and ecosystem health. Otherwise, the subnet token remains merely a speculative asset, prone to market fluctuations, and ultimately struggles to attract long-term users and developers.

4. Economic Model

All TAO token rewards are newly minted, similar to Bitcoin, Bittensor's TAO employs the same tokenomics and issuance curve as Bitcoin, with a total supply cap of 21 million, halving every 4 years.

Bittensor adopts a fair launch approach, with no pre-mine or ICO, where every circulating token must be earned through active participation in the network. Currently, the network mints 7,200 TAO per day (1 TAO per block, approximately every 12 seconds), following a programmatic issuance schedule: once half the total supply has been distributed, the issuance rate halves, a process that occurs approximately every 4 years and continues at each halving point until all 21 million TAO are in circulation.

While the issuance curve of TAO is similar to Bitcoin, the introduction of a burn mechanism, based on taostats token burn data, anticipates a delayed halving date for the Bittensor network (launched on January 3, 2021) to December 2025.

About HTX Research:

HTX Research is the dedicated research arm of HTX Group, responsible for in-depth analysis of a wide range of areas including cryptocurrency, blockchain technology, and emerging market trends, compiling comprehensive reports, and providing expert evaluations. HTX Research is committed to offering data-driven insights and strategic foresight, playing a key role in shaping industry perspectives and supporting informed decision-making in the digital asset space. With a rigorous research methodology and cutting-edge data analysis, HTX Research consistently stays at the forefront of innovation, leading industry thought and promoting a deep understanding of the evolving market dynamics.

This article is contributed content and does not represent the views of Blockbeats

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia