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

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

Bitensor, an early explorer in blockchain AI, aims to build an open, decentralized AI collaboration network. Its goal is to break down the centralized barriers in traditional AI development, allowing more participants to contribute and benefit collectively. Unlike traditional centralized AI systems (e.g., OpenAI), Bitensor creates an open peer-to-peer ecosystem where participants can earn rewards based on their contributions to the network.

Technical Architecture: Bitensor employs a dual-layer structure:

• Root Network (Mainnet): Manages system coordination, validation, and issuance of TAO tokens, serving as the central hub for resource allocation across the network.

• Subnet Ecosystem: Individual subnets function as specialized AI labs, developing tailored solutions for specific AI applications and proving their value through market competition.

This design allows Bitensor to balance overall network stability with specialization in various fields, providing a flexible infrastructure for decentralized AI development.

Ecosystem Development Progress:

• The number of subnets has expanded from an initial 32 to over 64, covering diverse AI applications such as text generation, trading signals, and data labeling.

• The active user base has reached 140,000, doubling from the previous year.

• The total market valuation of subnets exceeds $100 million, with daily trading volumes maintaining around $45 million.

• Institutional participation has significantly increased, with Grayscale including TAO in its Decentralized AI Fund, adjusting its weight to 29.55%.

These metrics indicate growing recognition from market participants and a positive trajectory for Bitensor's ecosystem.

dTAO System Upgrade: Bitensor's recent dTAO (Dynamic TAO) upgrade represents a significant innovation in its economic model. The core of this upgrade is to optimize the allocation mechanism of the TAO token, shifting from a resource allocation method reliant on validators' subjective judgment to a more market-oriented approach. This ensures resources are more accurately directed towards competitive subnets.

Key Issues in Bitensor's Original Economic Model:

1. Lack of Objective Assessment Mechanism: As the number of subnets increases, validators struggle to comprehensively and objectively assess the value of each subnet, leading to declining allocation efficiency.

2. Imbalanced Power Structure: Many validators are also subnet operators, creating potential conflicts of interest. Validators may favor subnets they are involved in, even engaging in private deals.

3. Participation Barriers: Ordinary TAO holders find it difficult to directly influence resource allocation decisions, with power overly concentrated among a few validators.

Solution: The dTAO upgrade introduces a market-based dynamic resource allocation system. Each subnet is transformed into an independent economic unit, with resource allocation driven by user demand. The core innovation is the subnet token (Alpha token) mechanism:

• Operation Principle: Users can stake TAO to obtain Alpha tokens issued by individual subnets, representing their support for specific subnets.

• Resource Allocation Logic: The market price of Alpha tokens serves as a signal for subnet demand intensity. Initially, Alpha tokens have the same price, with each pool containing only 1 TAO and 1 Alpha token. As liquidity for both tokens in the subnet increases, the price of Alpha tokens changes. TAO emissions are allocated proportionally based on subnet token prices, with higher-priced subnets receiving more TAO, thereby optimizing resource allocation automatically.

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

Most Active Subnets:

• Subnet 4 (Targon): Focuses on AI inference services for text generation, characterized by fast response times and low costs.

• Subnet 64 (Chutes): Provides APIs for various LLMs, enabling developers to build and deploy AI applications on the Bitensor network.

• Subnet 8 (PTN): Specializes in finance, incentivizing miners to generate accurate trading signals across various financial markets, including forex and cryptocurrencies.

• Subnet 52 (Dojo): Focuses on data labeling, encouraging users to earn tokens through data annotation. Yzi Labs has invested in its parent company, Tensorplex.

The Vana project addresses a core issue in today's digital economy: the ownership and value distribution of personal data. In the current internet ecosystem, user data is largely monopolized and controlled by large tech companies, with the actual data creators rarely benefiting. Vana's innovation lies in establishing an ecosystem where users truly own and control their data, earning economic rewards from it.

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

1. Data Liquidity Layer: The core of the Vana network, enabling the incentivization, aggregation, and verification of data assets through Data Liquidity Pools (DLPs). Each DLP is a smart contract dedicated to aggregating specific types of data assets, such as social media data or browsing history.

2. Data Portability Layer: Ensures seamless transfer of user data between different applications and AI models, enhancing data flexibility.

3. Universal Connectome: Tracks real-time data flows within the ecosystem, forming a data ecosystem map to ensure transparency.

4. Non-Custodial Data Storage: A key innovation of Vana is its unique data management approach. Users' raw data is not stored on-chain but is instead kept in user-selected storage locations, such as cloud servers or personal devices, ensuring complete user control over their data.

5. Application Ecosystem: Built on top of the data foundation, Vana's open application ecosystem allows developers to build innovative applications using data from DLPs, including AI applications. Data contributors earn dividend rewards from these applications.

This design enables Vana to protect user data privacy while creating a more equitable data value distribution mechanism, providing a crucial data foundation for decentralized AI development.

Latest Developments:

• In February 2025, Yzi Labs announced a strategic investment in Vana, with Binance founder CZ joining as an advisor.

• Vana has established data projects across multiple domains, including Finquarium (financial prediction data), GPT Data DAO (ChatGPT chat data), Reddit Data DAO (Reddit user data), Volara (Twitter data), and Flirtual (dating data).

• During Eth Denver, Vana organized a hackathon with a $50,000 prize pool to incentivize developers to build DataDAOs and AI applications based on Vana data, further expanding its ecosystem.

These advancements indicate that Vana is actively building a comprehensive ecosystem around data ownership and value monetization, with a development momentum worth watching.

Kite AI is an AI-focused native Layer 1 blockchain project built on the Avalanche framework. It aims to address challenges in traditional blockchains when handling AI assets, particularly how to achieve transparent attribution and incentivization of AI data, models, and agent contributions. Kite AI introduces four core technological innovations:

1. PoAI Consensus Mechanism: Proof of Attributed Intelligence is a consensus mechanism pioneered by Kite AI. It uses an on-chain verifiable contribution record system to precisely track the value contributions of data, models, and AI agents. The project designs a dynamic reward pool mechanism to allocate earnings based on contributions, effectively solving "data black boxes" and "model plagiarism" issues in traditional AI economies.

2. Composable AI Subnets: Kite AI employs a modular architecture, supporting developers to build industry-specific AI collaboration ecosystems as needed. For example, in a medical subnet, patient data can be encrypted and selectively authorized to pharmaceutical companies for AI model development, with earnings distributed among data subjects, model developers, and subnet maintainers, creating a win-win environment.

3. AI-Native Execution Layer: Kite AI is building an on-chain AI-native execution layer dedicated to handling AI computing tasks, such as inference, embedding, and fine-tuning/training. Users can authorize smart contract wallets to execute inference calls and interact with models. This layer supports blockchain transactions and state changes while integrating confidential computing environments (e.g., Trusted Execution Environments, TEEs) to ensure data security and privacy during computation.

4. Decentralized Data Engine: Kite AI ensures fair earnings for data creators in AI workflows. The platform includes regulatory-compliant modules (e.g., GDPR/CCPA) to meet global data privacy requirements, reducing developers' compliance costs.

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

Development Status:

• Kite AI launched its incentivized testnet on February 6, 2025, the first AI-native Layer 1 sovereign blockchain testnet. The testnet has performed impressively:

  • Within 70 hours of launch, the number of connected wallets exceeded 100,000. To date, 1.95 million wallets have joined the incentivized testnet V1, with over 1 million wallets interacting with AI agents and over 115 million total calls made.

• The project is backed by an experienced Silicon Valley team, with co-founders holding deep technical leadership experience in AI, having worked at top tech companies like Uber, Salesforce, and Databricks. Core team members come from industry-leading companies such as Google, BlackRock, Uber, and the NEAR Foundation, with academic backgrounds from top institutions like MIT and Harvard.

• In terms of funding, the project has secured investments from top-tier firms such as General Catalyst, Hashed, Hashkey, and Samsung Next, and has established technical partnerships with Eigenlayer, Sui, Avalanche, and AWS.

• As a selector for Avalanche's InfraBUILDL(AI) program, Kite AI actively promotes the development of Avalanche's AI ecosystem, aiming to make Avalanche a leading blockchain for AI applications.

With the global data economy projected to exceed $70 billion in 2025, Kite AI is poised to become a crucial infrastructure for data attribution and monetization, with significant potential for growth.

Nillion is redefining the processing of sensitive data through its unique "blind computing" technology, paving new ways for future digital privacy protection. 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. Founded in November 2021 by forward-thinking innovators like Alex Page and Andrew Masanto, Nillion aims to create a system that can securely process high-value data without exposing sensitive details.

Nillion's core advantage lies in its "blind computing" capability—a process that keeps data encrypted throughout its lifecycle, including storage, transmission, and processing. Its technical architecture integrates several cutting-edge privacy protection technologies:

• Multiparty Computation (MPC): Enables multiple nodes to collaboratively compute functions without revealing their private inputs, allowing joint computation without data sharing.

• Fully Homomorphic Encryption (FHE): Allows direct operations on encrypted data, ensuring data remains encrypted from start to finish, providing end-to-end privacy protection.

• Zero-Knowledge Proofs (ZKP): Offers a way to verify computations without disclosing any underlying data, enhancing system trustworthiness.

• Nada Language: A domain-specific language designed for creating secure MPC programs, simplifying the development process for privacy-preserving applications and reducing the learning curve for developers.

Nillion's network architecture consists of three main layers: the Processing Layer (responsible for secure computation), the Coordination Layer (NilChain, managing node-to-node communication), and the Connection Layer (acting as a gateway to external systems). This multi-layer architecture enables Nillion to provide robust computing capabilities while protecting data privacy, meeting the needs of various privacy-sensitive scenarios.

Latest Development Progress:

• Nillion's mainnet is scheduled to launch in March 2025. The total supply of Nillion tokens is 1 billion, with distribution planned to coincide with the mainnet launch.

• In terms of funding, Nillion completed a $25 million financing round led by Hack VC on October 30, 2024. Investors include well-known institutions such as HashKey Capital and Animoca Brands, as well as angel investors from projects like Arbitrum, Worldcoin, and Sei. This round brings Nillion's total funding to $45 million, providing ample financial support for long-term development.

• In terms of ecosystem expansion, Nillion has established integration partnerships with several mainstream public chains, including NEAR Protocol, Aptos, Arbitrum, Mantle, and Sei. Through its collaboration with NEAR Protocol, Nillion aims to enhance privacy tools, enabling developers to innovate more effectively in the DeFi space.

• In the AI ecosystem, Nillion has partnered with multiple AI-related projects, including Ritual, Crush AI, Skillful AI, and Virtuals Protocol. For instance, Virtuals Protocol, the largest multimodal AI agent protocol, collaborates with Nillion to leverage its secure computing infrastructure for private AI model training and inference, achieving a perfect integration of AI and privacy.

Project Development:
The core concept of Sahara AI is to build a "Human AI Collaboration Network," enabling ordinary users, developers, and enterprises to participate in the creation, deployment, and monetization of AI assets. Through this collaborative model, Sahara AI aims to lower the barriers to entry for AI, allowing every participant to share in the industry's growth. The project has successfully secured $43 million in funding led by Binance Labs, Polychain Capital, and Pantera Capital.

Technical Architecture:
The platform's architecture includes three key components:

1. Sahara Blockchain: Provides the foundation for secure, transparent transactions and efficient AI lifecycle management within the ecosystem.

2. AI Infrastructure: Supports distributed collaborative training and service capabilities for advanced algorithms and computing frameworks.

3. Sahara AI Marketplace: A decentralized trading center for AI assets.

These components together form a comprehensive ecosystem that supports the entire process from data collection and labeling to model training, deployment, and monetization.

Latest Development Progress:

• In December 2024, Sahara AI launched the beta version of its first-phase data service platform testnet, attracting over 780,000 user applications, with more than 10,000 waitlisted participants granted initial access. In this phase, participants can contribute to the AI ecosystem by completing data collection, optimization, and labeling tasks and earn rewards.

• In February 2025, Sahara AI initiated the second phase of its testnet, expanding the contributor base and introducing additional reward opportunities. This phase further enhances user participation in shaping the future of decentralized AI.

• Most recently, Sahara AI announced the launch of its public testnet, named "SIWA," on March 10, 2025. This is considered the final major test before the mainnet launch and TGE, and it may be the last opportunity for participants to earn airdrop rewards (referred to as "points").

Sahara AI has unveiled its 2024-2025 roadmap, featuring several key milestones:

• Q4 2024: Launched the data service platform and testnet, enabling users to earn rewards through data collection and labeling.

• Q1 2025: AI Marketplace went live, offering development tools and expanded data services to support model development, training, and deployment, with an early access program introduced.

• Q2 2025: Launch of the Sahara Studio toolkit, covering model training, deployment, and workflow management, further optimizing the developer experience.

• Q3 2025: Sahara Chain mainnet release, providing a secure, transparent on-chain infrastructure for large-scale decentralized AI, supporting the tokenization and trading of data and models.

On March 1, 2025, Sahara AI launched an incubator program aimed at identifying and supporting the most promising AI x Web3 innovation projects globally. The program focuses on two major tracks: AI infrastructure and AI applications, welcoming teams with MVP or higher maturity levels. Successful projects will have the opportunity to fully integrate into the Sahara AI ecosystem, gaining exclusive technical support, market expansion resources, and investment opportunities.

AI Layer 1 is at a critical stage of rapid evolution. This emerging field is reconstructing the underlying architecture of AI technology through decentralized infrastructure. From data attribution to computing resource allocation, from model training to application deployment, these platforms are breaking through the limitations of traditional centralized AI systems and building a more open, transparent, and efficient technical ecosystem. In the future, this track will continue to drive technological innovation, pushing artificial intelligence towards a more decentralized and collaborative direction.

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

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

Bitensor, an early explorer in blockchain AI, aims to build an open, decentralized AI collaboration network. Its goal is to break down the centralized barriers in traditional AI development, allowing more participants to contribute and benefit collectively. Unlike traditional centralized AI systems (e.g., OpenAI), Bitensor creates an open peer-to-peer ecosystem where participants can earn rewards based on their contributions to the network.

Technical Architecture: Bitensor employs a dual-layer structure:

• Root Network (Mainnet): Manages system coordination, validation, and issuance of TAO tokens, serving as the central hub for resource allocation across the network.

• Subnet Ecosystem: Individual subnets function as specialized AI labs, developing tailored solutions for specific AI applications and proving their value through market competition.

This design allows Bitensor to balance overall network stability with specialization in various fields, providing a flexible infrastructure for decentralized AI development.

Ecosystem Development Progress:

• The number of subnets has expanded from an initial 32 to over 64, covering diverse AI applications such as text generation, trading signals, and data labeling.

• The active user base has reached 140,000, doubling from the previous year.

• The total market valuation of subnets exceeds $100 million, with daily trading volumes maintaining around $45 million.

• Institutional participation has significantly increased, with Grayscale including TAO in its Decentralized AI Fund, adjusting its weight to 29.55%.

These metrics indicate growing recognition from market participants and a positive trajectory for Bitensor's ecosystem.

dTAO System Upgrade: Bitensor's recent dTAO (Dynamic TAO) upgrade represents a significant innovation in its economic model. The core of this upgrade is to optimize the allocation mechanism of the TAO token, shifting from a resource allocation method reliant on validators' subjective judgment to a more market-oriented approach. This ensures resources are more accurately directed towards competitive subnets.

Key Issues in Bitensor's Original Economic Model:

1. Lack of Objective Assessment Mechanism: As the number of subnets increases, validators struggle to comprehensively and objectively assess the value of each subnet, leading to declining allocation efficiency.

2. Imbalanced Power Structure: Many validators are also subnet operators, creating potential conflicts of interest. Validators may favor subnets they are involved in, even engaging in private deals.

3. Participation Barriers: Ordinary TAO holders find it difficult to directly influence resource allocation decisions, with power overly concentrated among a few validators.

Solution: The dTAO upgrade introduces a market-based dynamic resource allocation system. Each subnet is transformed into an independent economic unit, with resource allocation driven by user demand. The core innovation is the subnet token (Alpha token) mechanism:

• Operation Principle: Users can stake TAO to obtain Alpha tokens issued by individual subnets, representing their support for specific subnets.

• Resource Allocation Logic: The market price of Alpha tokens serves as a signal for subnet demand intensity. Initially, Alpha tokens have the same price, with each pool containing only 1 TAO and 1 Alpha token. As liquidity for both tokens in the subnet increases, the price of Alpha tokens changes. TAO emissions are allocated proportionally based on subnet token prices, with higher-priced subnets receiving more TAO, thereby optimizing resource allocation automatically.

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

Most Active Subnets:

• Subnet 4 (Targon): Focuses on AI inference services for text generation, characterized by fast response times and low costs.

• Subnet 64 (Chutes): Provides APIs for various LLMs, enabling developers to build and deploy AI applications on the Bitensor network.

• Subnet 8 (PTN): Specializes in finance, incentivizing miners to generate accurate trading signals across various financial markets, including forex and cryptocurrencies.

• Subnet 52 (Dojo): Focuses on data labeling, encouraging users to earn tokens through data annotation. Yzi Labs has invested in its parent company, Tensorplex.

The Vana project addresses a core issue in today's digital economy: the ownership and value distribution of personal data. In the current internet ecosystem, user data is largely monopolized and controlled by large tech companies, with the actual data creators rarely benefiting. Vana's innovation lies in establishing an ecosystem where users truly own and control their data, earning economic rewards from it.

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

1. Data Liquidity Layer: The core of the Vana network, enabling the incentivization, aggregation, and verification of data assets through Data Liquidity Pools (DLPs). Each DLP is a smart contract dedicated to aggregating specific types of data assets, such as social media data or browsing history.

2. Data Portability Layer: Ensures seamless transfer of user data between different applications and AI models, enhancing data flexibility.

3. Universal Connectome: Tracks real-time data flows within the ecosystem, forming a data ecosystem map to ensure transparency.

4. Non-Custodial Data Storage: A key innovation of Vana is its unique data management approach. Users' raw data is not stored on-chain but is instead kept in user-selected storage locations, such as cloud servers or personal devices, ensuring complete user control over their data.

5. Application Ecosystem: Built on top of the data foundation, Vana's open application ecosystem allows developers to build innovative applications using data from DLPs, including AI applications. Data contributors earn dividend rewards from these applications.

This design enables Vana to protect user data privacy while creating a more equitable data value distribution mechanism, providing a crucial data foundation for decentralized AI development.

Latest Developments:

• In February 2025, Yzi Labs announced a strategic investment in Vana, with Binance founder CZ joining as an advisor.

• Vana has established data projects across multiple domains, including Finquarium (financial prediction data), GPT Data DAO (ChatGPT chat data), Reddit Data DAO (Reddit user data), Volara (Twitter data), and Flirtual (dating data).

• During Eth Denver, Vana organized a hackathon with a $50,000 prize pool to incentivize developers to build DataDAOs and AI applications based on Vana data, further expanding its ecosystem.

These advancements indicate that Vana is actively building a comprehensive ecosystem around data ownership and value monetization, with a development momentum worth watching.

Kite AI is an AI-focused native Layer 1 blockchain project built on the Avalanche framework. It aims to address challenges in traditional blockchains when handling AI assets, particularly how to achieve transparent attribution and incentivization of AI data, models, and agent contributions. Kite AI introduces four core technological innovations:

1. PoAI Consensus Mechanism: Proof of Attributed Intelligence is a consensus mechanism pioneered by Kite AI. It uses an on-chain verifiable contribution record system to precisely track the value contributions of data, models, and AI agents. The project designs a dynamic reward pool mechanism to allocate earnings based on contributions, effectively solving "data black boxes" and "model plagiarism" issues in traditional AI economies.

2. Composable AI Subnets: Kite AI employs a modular architecture, supporting developers to build industry-specific AI collaboration ecosystems as needed. For example, in a medical subnet, patient data can be encrypted and selectively authorized to pharmaceutical companies for AI model development, with earnings distributed among data subjects, model developers, and subnet maintainers, creating a win-win environment.

3. AI-Native Execution Layer: Kite AI is building an on-chain AI-native execution layer dedicated to handling AI computing tasks, such as inference, embedding, and fine-tuning/training. Users can authorize smart contract wallets to execute inference calls and interact with models. This layer supports blockchain transactions and state changes while integrating confidential computing environments (e.g., Trusted Execution Environments, TEEs) to ensure data security and privacy during computation.

4. Decentralized Data Engine: Kite AI ensures fair earnings for data creators in AI workflows. The platform includes regulatory-compliant modules (e.g., GDPR/CCPA) to meet global data privacy requirements, reducing developers' compliance costs.

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

Development Status:

• Kite AI launched its incentivized testnet on February 6, 2025, the first AI-native Layer 1 sovereign blockchain testnet. The testnet has performed impressively:

  • Within 70 hours of launch, the number of connected wallets exceeded 100,000. To date, 1.95 million wallets have joined the incentivized testnet V1, with over 1 million wallets interacting with AI agents and over 115 million total calls made.

• The project is backed by an experienced Silicon Valley team, with co-founders holding deep technical leadership experience in AI, having worked at top tech companies like Uber, Salesforce, and Databricks. Core team members come from industry-leading companies such as Google, BlackRock, Uber, and the NEAR Foundation, with academic backgrounds from top institutions like MIT and Harvard.

• In terms of funding, the project has secured investments from top-tier firms such as General Catalyst, Hashed, Hashkey, and Samsung Next, and has established technical partnerships with Eigenlayer, Sui, Avalanche, and AWS.

• As a selector for Avalanche's InfraBUILDL(AI) program, Kite AI actively promotes the development of Avalanche's AI ecosystem, aiming to make Avalanche a leading blockchain for AI applications.

With the global data economy projected to exceed $70 billion in 2025, Kite AI is poised to become a crucial infrastructure for data attribution and monetization, with significant potential for growth.

Nillion is redefining the processing of sensitive data through its unique "blind computing" technology, paving new ways for future digital privacy protection. 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. Founded in November 2021 by forward-thinking innovators like Alex Page and Andrew Masanto, Nillion aims to create a system that can securely process high-value data without exposing sensitive details.

Nillion's core advantage lies in its "blind computing" capability—a process that keeps data encrypted throughout its lifecycle, including storage, transmission, and processing. Its technical architecture integrates several cutting-edge privacy protection technologies:

• Multiparty Computation (MPC): Enables multiple nodes to collaboratively compute functions without revealing their private inputs, allowing joint computation without data sharing.

• Fully Homomorphic Encryption (FHE): Allows direct operations on encrypted data, ensuring data remains encrypted from start to finish, providing end-to-end privacy protection.

• Zero-Knowledge Proofs (ZKP): Offers a way to verify computations without disclosing any underlying data, enhancing system trustworthiness.

• Nada Language: A domain-specific language designed for creating secure MPC programs, simplifying the development process for privacy-preserving applications and reducing the learning curve for developers.

Nillion's network architecture consists of three main layers: the Processing Layer (responsible for secure computation), the Coordination Layer (NilChain, managing node-to-node communication), and the Connection Layer (acting as a gateway to external systems). This multi-layer architecture enables Nillion to provide robust computing capabilities while protecting data privacy, meeting the needs of various privacy-sensitive scenarios.

Latest Development Progress:

• Nillion's mainnet is scheduled to launch in March 2025. The total supply of Nillion tokens is 1 billion, with distribution planned to coincide with the mainnet launch.

• In terms of funding, Nillion completed a $25 million financing round led by Hack VC on October 30, 2024. Investors include well-known institutions such as HashKey Capital and Animoca Brands, as well as angel investors from projects like Arbitrum, Worldcoin, and Sei. This round brings Nillion's total funding to $45 million, providing ample financial support for long-term development.

• In terms of ecosystem expansion, Nillion has established integration partnerships with several mainstream public chains, including NEAR Protocol, Aptos, Arbitrum, Mantle, and Sei. Through its collaboration with NEAR Protocol, Nillion aims to enhance privacy tools, enabling developers to innovate more effectively in the DeFi space.

• In the AI ecosystem, Nillion has partnered with multiple AI-related projects, including Ritual, Crush AI, Skillful AI, and Virtuals Protocol. For instance, Virtuals Protocol, the largest multimodal AI agent protocol, collaborates with Nillion to leverage its secure computing infrastructure for private AI model training and inference, achieving a perfect integration of AI and privacy.

Project Development:
The core concept of Sahara AI is to build a "Human AI Collaboration Network," enabling ordinary users, developers, and enterprises to participate in the creation, deployment, and monetization of AI assets. Through this collaborative model, Sahara AI aims to lower the barriers to entry for AI, allowing every participant to share in the industry's growth. The project has successfully secured $43 million in funding led by Binance Labs, Polychain Capital, and Pantera Capital.

Technical Architecture:
The platform's architecture includes three key components:

1. Sahara Blockchain: Provides the foundation for secure, transparent transactions and efficient AI lifecycle management within the ecosystem.

2. AI Infrastructure: Supports distributed collaborative training and service capabilities for advanced algorithms and computing frameworks.

3. Sahara AI Marketplace: A decentralized trading center for AI assets.

These components together form a comprehensive ecosystem that supports the entire process from data collection and labeling to model training, deployment, and monetization.

Latest Development Progress:

• In December 2024, Sahara AI launched the beta version of its first-phase data service platform testnet, attracting over 780,000 user applications, with more than 10,000 waitlisted participants granted initial access. In this phase, participants can contribute to the AI ecosystem by completing data collection, optimization, and labeling tasks and earn rewards.

• In February 2025, Sahara AI initiated the second phase of its testnet, expanding the contributor base and introducing additional reward opportunities. This phase further enhances user participation in shaping the future of decentralized AI.

• Most recently, Sahara AI announced the launch of its public testnet, named "SIWA," on March 10, 2025. This is considered the final major test before the mainnet launch and TGE, and it may be the last opportunity for participants to earn airdrop rewards (referred to as "points").

Sahara AI has unveiled its 2024-2025 roadmap, featuring several key milestones:

• Q4 2024: Launched the data service platform and testnet, enabling users to earn rewards through data collection and labeling.

• Q1 2025: AI Marketplace went live, offering development tools and expanded data services to support model development, training, and deployment, with an early access program introduced.

• Q2 2025: Launch of the Sahara Studio toolkit, covering model training, deployment, and workflow management, further optimizing the developer experience.

• Q3 2025: Sahara Chain mainnet release, providing a secure, transparent on-chain infrastructure for large-scale decentralized AI, supporting the tokenization and trading of data and models.

On March 1, 2025, Sahara AI launched an incubator program aimed at identifying and supporting the most promising AI x Web3 innovation projects globally. The program focuses on two major tracks: AI infrastructure and AI applications, welcoming teams with MVP or higher maturity levels. Successful projects will have the opportunity to fully integrate into the Sahara AI ecosystem, gaining exclusive technical support, market expansion resources, and investment opportunities.

AI Layer 1 is at a critical stage of rapid evolution. This emerging field is reconstructing the underlying architecture of AI technology through decentralized infrastructure. From data attribution to computing resource allocation, from model training to application deployment, these platforms are breaking through the limitations of traditional centralized AI systems and building a more open, transparent, and efficient technical ecosystem. In the future, this track will continue to drive technological innovation, pushing artificial intelligence towards a more decentralized and collaborative direction.