AI Layer 1 Rapid Evolution, Building Decentralization AI Infrastructure

With the rapid development of AI technology, traditional blockchain architectures can no longer meet the high-performance computing and complex data processing requirements needed for AI applications. This has led to the rise of Layer 1 blockchain platforms optimized specifically for AI, which exhibit diverse characteristics in terms of technical architecture, application scenarios, and business models.

This article provides an in-depth analysis of five leading AI Layer 1 projects: Bittensor, Vana, Kite AI, Nillion, and Sahara.

Bittensor: Decentralization AI Network Infrastructure

Bittensor, as an early explorer in the blockchain AI field, is committed to building an open decentralized artificial intelligence collaboration network. Its goal is to break down the centralized barriers in traditional AI development, allowing more participants to contribute and benefit together.

Bittensor's technical architecture adopts a dual-layer structural design:

• Root Network (Mainnet): Responsible for the coordination, verification, and issuance management of TAO tokens for the entire system, serving as the hub for resource allocation in the entire network.
• Subnet Ecosystem: Each subnet acts like an independent AI laboratory, developing specialized solutions for specific AI application scenarios and proving its value in market competition.

This design allows Bittensor to balance the overall stability of the network with the expertise of various fields, providing a flexible infrastructure for the development of Decentralization AI.

Ecological Development Progress

• The number of subnets has expanded from the initial 32 to over 64, covering various AI application scenarios such as text generation, trading signals, and data annotation.
• The number of active users has reached 140,000, doubling compared to the previous year.
• The total market valuation of the subnet exceeds 100 million USD, with a daily trading volume maintained at around 45 million USD.
• Institutional participation has significantly increased, with well-known funds incorporating TAO into their Decentralization AI fund, adjusting its weight to 29.55%.

These data indicate that Bittensor is gaining recognition from an increasing number of market participants, and its ecosystem is entering a positive development track.

Bittensor recently completed the dTAO (dynamic TAO) system upgrade, which is an important innovation in its economic model. The core of this upgrade lies in optimizing the distribution mechanism of the token TAO, shifting from a resource allocation method that relies on validators' subjective judgments to a more market-oriented distribution mechanism, allowing resources to flow more precisely to those subnetworks that are truly competitive.

To address these issues, the dTAO upgrade introduces a dynamic resource allocation system based on market mechanisms. This system transforms each subnet into an independent economic unit, driven by users' actual needs for resource allocation. Its core innovation is the subnet token (Alpha token) mechanism:

• Operating Principle: Users can obtain Alpha tokens issued by various subnetworks by staking TAO. These tokens represent the user's support for specific subnetworks.
• Resource allocation logic: The market price of Alpha tokens serves as a signal for measuring the demand intensity of the subnet. Initially, the price of alpha tokens is the same, with only 1 TAO and 1 alpha token in each pool. As liquidity of the two types of tokens in the subnet increases, the price of alpha tokens will also change accordingly. The emission of TAO is proportionally allocated based on the price of subnet tokens among all tokens, with higher-priced subnets receiving more TAO allocation, thereby achieving automatic optimization of resource allocation.

This mechanism significantly improves the efficiency and fairness of resource allocation, solidifying the value of the TAO token and providing more avenues for ordinary users to participate in network governance.

The currently most active subnets include:

• Subnet Targon 4: Focused on AI inference services for text generation, characterized by fast response time and low cost.
• Subnet 64 Chutes: Provides API interfaces for various LLMs, allowing developers to build and deploy AI applications on the Bittensor network.
• Subnet No. 8 PTN: Focused on the financial sector, incentivizing miners to generate accurate trading signals through a reward mechanism, covering various financial markets such as foreign exchange and cryptocurrencies.
• Subnet 52 Dojo: Perform data labeling and encourage users to earn tokens through data labeling.

Vana: Data Sovereignty and Value Reconstruction Platform

The Vana project focuses on addressing a core issue in today's digital economy: the ownership of personal data and value distribution. Vana's innovation lies in establishing an ecosystem where users truly own and control their own data, while also being able to derive economic benefits from it.

As an EVM-compatible Layer 1 blockchain network, Vana's technical architecture consists of five core components:

1. Data Liquidity Layer: This is the core of the Vana network, achieving the incentivization, aggregation, and verification of data assets through the Data Liquidity Pool (DLP). Each DLP is a smart contract specifically designed to aggregate certain types of data assets, such as social media data, browsing history, etc.
2. Data Portability Layer: Ensures that user data can be easily transferred between different applications and AI models, enhancing the flexibility of data usage.
3. Universal Connectome: Tracking the real-time flow of data throughout the ecosystem, creating a data ecosystem map that ensures the transparency of the system.
4. Non-custodial data storage: One of Vana's important innovations is its unique data management approach. Users' raw data will not be put on-chain, but instead, users can choose their own storage location, such as cloud servers or personal devices, ensuring that users have complete control over their own data.
5. Application Ecosystem: Based on data, Vana has built an open application ecosystem where developers can use the data accumulated from DLP to create various innovative applications, including AI applications, while data contributors can receive dividend rewards from these applications.

This design enables Vana to create a fairer data value distribution mechanism while protecting user data privacy, providing an important data foundation for the development of Decentralization AI.

Latest developments

Vana's financing and collaboration expansion continues to advance:

• In February 2025, a certain laboratory announced a strategic investment in Vana, and a well-known figure joined as an advisor.
• In terms of ecological construction, Vana has built multiple data projects covering various fields from social media data to financial forecasting data, including: Finquarium (financial forecasting data), GPT Data DAO (ChatGPT chat data), Reddit Data DAO (Reddit user data), Volara (Twitter data), Flirtual (dating data), etc.
• Recently, Vana organized a hackathon during Eth Denver, offering a high reward pool to incentivize developers to build DataDAOs and AI applications based on Vana data, further expanding its ecosystem.

These developments indicate that Vana is actively building a complete ecosystem around data ownership and value realization, and its momentum is worth paying attention to.

Kite AI: Technical Breakthrough of AI-Native Public Blockchain

Kite AI is a native Layer 1 blockchain project focused on the AI field, built on the Avalanche framework. It aims to address the various challenges faced by traditional blockchains when handling AI assets, particularly how to achieve transparent rights and incentives for AI data, models, and agent contributions. Kite AI proposes four core technological innovations:

1. PoAI Consensus Mechanism: Proof of Attributed Intelligence is a consensus mechanism pioneered by Kite AI, which accurately tracks the value contributions of data, models, and AI agents through an on-chain verifiable contribution record system. The project has designed a dynamic reward pool mechanism that distributes earnings based on contribution proportions, effectively addressing issues such as "data black box" and "model plagiarism" in the traditional AI economy.

2. Composable AI Subnet: Kite AI adopts a modular architecture, allowing developers to build industry-specific AI collaboration ecosystems as needed. For example, in the healthcare subnet, patient data can be encrypted and selectively authorized for pharmaceutical companies to use for AI model development, with profits distributed among data subjects, model developers, and subnet maintainers in a certain proportion, creating a win-win ecosystem for multiple parties.

3. AI Native Execution Layer: Kite AI is building an on-chain AI native execution layer specifically for handling AI computing tasks such as inference, embedding, and fine-tuning/training. Users can authorize smart contract wallets to perform inference calls and interact with models through this layer. This execution layer not only supports blockchain transactions and state changes but also integrates confidential computing environments (such as Trusted Execution Environments TEE), ensuring data security and privacy protection during the computation process.

4. Decentralization Data Engine: Kite AI ensures that data creators receive fair rewards in the AI workflow. The platform has built-in compliance modules that meet global data privacy requirements, such as GDPR/CCPA, reducing compliance costs for developers.

These technological innovations enable Kite AI to create a fairer and more transparent value distribution environment for AI developers and data providers, promoting the Decentralization of AI technology.

Development Status

Kite AI launched its incentive testnet on February 6, 2025, which is the first AI-native Layer 1 sovereign blockchain testnet. The testnet performed impressively after going live:

• Less than 70 hours after the test went live, the number of connected wallets surpassed 100,000. As of now, a total of 1.95 million wallets have joined the incentive test network V1, with over 1 million wallets interacting with AI agents, and the total number of calls exceeding 115 million times.

• The project background is strong, built by an experienced Silicon Valley team, with co-founders who have deep technical leadership experience in the field of artificial intelligence and have worked at several top tech companies. Core team members come from several industry-leading companies and have academic backgrounds from top universities.

• In terms of capital support, the project has received investments from multiple top institutions and has established technical collaborations with several well-known projects.

• As a selection committee member of the Avalanche InfraBUILDL(AI) program, Kite AI plays an active role in promoting the development of the Avalanche artificial intelligence ecosystem. This collaboration aims to make Avalanche a leading blockchain for AI applications.

• With the global data economy expected to exceed 70 billion dollars by 2025, Kite AI is poised to become an important infrastructure for data rights confirmation and monetization, and its development potential is worth looking forward to.

Nillion: The Frontier Exploration of Privacy Computing

Nillion is redefining the way sensitive data is processed through its unique "blind computing" technology, opening new avenues for digital privacy protection in the future.

Nillion is a decentralized public network based on an innovative cryptographic primitive called Nil Message Compute (NMC), allowing network nodes to operate in a non-traditional blockchain manner. The project was founded in November 2021, led by several forward-thinking innovators, with the aim of creating a system that can securely handle high-value data without exposing sensitive details.

The core advantage of Nillion lies in its "blind computation" capability——a process that allows data to remain encrypted throughout its lifecycle of storage, transmission, and processing. Its technical architecture integrates various cutting-edge privacy protection technologies:

• Multi-Party Computation ( MPC ): Enables multiple nodes to collaboratively compute a function without disclosing their private inputs, achieving joint computation without data sharing.
• Fully Homomorphic Encryption ( FHE ): Allows direct operations on encrypted data, ensuring that data remains encrypted from start to finish, providing privacy protection throughout the entire process.
• Zero-Knowledge Proof ( ZKP ): Provides a method for verifying computations without disclosing any underlying data, enhancing the credibility of the system.
• Nada Language: This is a domain-specific language designed for creating secure MPC programs, simplifying the development process of privacy-preserving applications and lowering the learning curve for developers.

The network architecture of Nillion consists of three main layers: the processing layer ( is responsible for secure computation.