The rapid advancement of Machine Learning (ML) and Artificial Intelligence (AI) has revolutionized various industries, providing new capabilities and efficiencies. From healthcare to finance, these technologies have transformed how we process and analyze data, offering valuable insights and automating complex tasks. However, with the rise of AI and ML, there is a growing concern about the privacy and security of personal data. As data collection and processing increase, so does the risk of unauthorized access, data breaches, and misuse of information.

Privasea AI Network addresses these concerns by integrating advanced privacy and security measures into its framework. By employing Fully Homomorphic Encryption (FHE) and decentralized protocols, Privasea ensures that data is securely processed and protected. This innovative approach allows users to benefit from AI capabilities without compromising their privacy, providing a secure solution for the digital age.

What is Privasea?

Privasea is a Decentralized Physical Infrastructure Network (DePIN) designed to enhance data privacy and security throughout the AI computation process. This powerful system leverages advanced cryptographic techniques, specifically Fully Homomorphic Encryption (FHE), which allows computations on encrypted data, producing results identical to those performed on unencrypted data. This ensures that the privacy of the data is preserved throughout the entire workflow, including AI model training and evaluation.

Vision and Mission of Privasea

Founded in 2022 by David Jiao, Privasea’s vision is to create a secure digital landscape where users can confidently share and utilize their data without fear of privacy violations. Its mission is to develop and implement cutting-edge solutions, safeguarding personal information while enabling seamless access to AI-driven insights.

Key Features of Privasea

• Fully Homomorphic Encryption Machine Learning (FHEML): Privasea utilizes FHE, a cutting-edge encryption method that allows computations to be performed on encrypted data without needing to decrypt it first. This ensures that sensitive information remains private, even during complex AI tasks, making it a game-changer for confidential computing.
• Decentralized AI Network: By leveraging distributed computing resources, Privasea eliminates reliance on centralized servers, reducing the risk of single points of failure or data breaches. This decentralized infrastructure supports secure and scalable AI operations.
• ImHuman App: This application allows users to prove their humanity (e.g., for verification purposes) without revealing personal details. It provides a privacy-preserving alternative to traditional identity verification systems, ensuring a bot-free environment while safeguarding user anonymity.
• Support from Advanced Technology: Privasea integrates tools like ZAMA’s Concrete ML, enhancing its ability to execute sophisticated FHE-based AI tasks efficiently. This technical backbone strengthens its capacity to deliver secure and practical solutions.
• Community-Driven Approach: Privasea allocates a portion of its token supply to its community, fostering engagement and incentivizing participation in its ecosystem. This reflects its commitment to decentralization and collective ownership.

Privasea AI Network Architecture

Privasea’s AI network architecture is built around Fully Homomorphic Encryption (FHE), a groundbreaking cryptographic method that enables computations on encrypted data without decryption. This ensures data privacy throughout the entire process, from submission to model training and predictions. The architecture integrates a decentralized computation network, user-friendly interfaces, and blockchain-based incentives, creating a secure, scalable, and privacy-preserving ecosystem for AI applications.

FHE Infrastructure

The FHE Infrastructure is the cryptographic backbone of Privasea, designed to perform secure computations on encrypted data. It incorporates libraries like TFHE-rs and ConcreteML, offering developers tools for operations such as addition, multiplication, and machine learning model evaluation—all without exposing raw data. Supporting multiple FHE schemes (e.g., TFHE, CKKS, BGV, BFV), it handles diverse AI tasks while ensuring privacy. Features include basic functions (e.g., Boolean gates, ReLU) and advanced techniques like ciphertext packing, with plans to integrate more libraries and innovations, such as ZAMA’s compressed key functions, to enhance performance and versatility.

Privasea’s FHEML leverages FHE to enable confidential machine learning inferencing, where machine learning models process “secret datasets” within ciphertext spaces. Data remains encrypted throughout the computation, enhancing privacy and security. The technology is designed to revolutionize machine learning inference by allowing computations on encrypted data without exposing sensitive information to third parties, such as cloud providers or collaborative partners.

Privasea API

The Privasea API serves as a seamless interface for users and developers to interact with the network. It simplifies tasks like data submission, model training, and prediction requests, all while maintaining end-to-end encryption via the HESea library. This API abstracts the complexity of FHE, ensuring secure communication and efficient management of encrypted data processes. It empowers users to leverage Privasea’s capabilities without needing deep cryptographic expertise, making privacy-preserving AI accessible and practical.

Privanetix

Privanetix is Privasea’s decentralized computation network, consisting of high-performance nodes equipped with the HESea library. These nodes collaborate to execute FHE-based operations on encrypted data, delivering computational power and privacy. By distributing tasks across multiple nodes, Privanetix ensures scalability and efficiency, enabling secure execution of machine learning algorithms while safeguarding sensitive information.

Privasea Smart Contract

The Privasea Smart Contract Kit powers a blockchain-based incentive system that drives network participation. It tracks Privanetix node contributions, validates computations, and distributes rewards transparently and fairly. Leveraging smart contracts, this mechanism fosters collaboration by incentivizing nodes to provide computational resources, ensuring a decentralized, trustworthy ecosystem that motivates active engagement and maintains accountability.

Roles in Privasea AI Network

Data Owners

Data Owners are individuals or entities who possess and control the valuable datasets within the network. As the data custodians, they have the authority to determine how their data is shared and utilized within the network. Data Owners play a crucial role in preserving privacy and determining the level of access granted to Privanetix Nodes and Decryptors. They have the power to define data-sharing agreements, set permissions, and specify the terms under which their data can be accessed and utilized.

Privanetix Nodes

Privanetix Nodes are the computational powerhouses within the Privasea AI Network. These nodes contribute their processing resources to perform complex computations on encrypted data. Equipped with the advanced HESea library and integrated with the blockchain-based incentive mechanism, Privanetix Nodes execute secure and efficient computations while preserving the privacy of the underlying data. They actively participate in the network by processing encrypted data, training machine learning models, and contributing to collaborative tasks. Privanetix Nodes ensure the integrity and confidentiality of computations and are essential to achieving the network’s objectives.

Decryptors

Decryptors serve as the result retrievers in the final computation process. Their role involves retrieving encrypted results from Privanetix Nodes and decrypting them using the appropriate keys through the Privasea API. The primary responsibility of Decryptors is to securely obtain the computation results and ensure the confidentiality and integrity of the data during the retrieval process.

Result Receivers

Result Receivers act as the final recipients of the decrypted computation results. They may perform additional operations on the decrypted results, such as further analysis, processing, or integration with other systems or applications. This allows for utilizing the computation results in a broader context, enabling informed decision-making and deriving valuable insights.

Workflow of Privasea AI Network

Privasea’s workflow ensures a seamless and privacy-focused experience for users engaging with machine learning tasks. The process involves several key steps:

1. User Interaction and Task Initiation: Users generate accounts and initiate machine learning tasks. Encrypt input vectors locally using the Privacy-Preserving Machine Learning Application API or a DApp integrated with the Privasea Application API. Generate a switching key locally.
2. Task Submission to Privanetix: Users submit encrypted tasks to the Privanetix network and pay service fees securely through blockchain-based transactions.
3. Computation by Privanetix Nodes: Privanetix nodes receive and execute encrypted tasks in the user’s encryption domain. They transfer the encrypted results to the decryptor’s encryption domain using the switching key.
4. Result Submission and Reward Withdrawal: Privanetix nodes submit the results to decryptors and withdraw service fees for this task through blockchain transactions.
5. Result Decryption by Decryptors: Decryptors use their client keys to decrypt the results.
6. Result Delivery to Network User: Decrypted results are sent to the network users using the Proxy Re-encryption (PRE) scheme.

Privasea Products

Privasea offers a suite of products leveraging Fully Homomorphic Encryption (FHE) and decentralized infrastructure to deliver privacy-preserving AI and human verification solutions.

DeepSea

DeepSea is Privasea’s decentralized AI network, built on FHE technology, enabling secure computations on encrypted data without decryption. Launched in beta on January 6, 2025, with a Genesis phase ending February 25, 2025, it integrates deep learning capabilities with a privacy-first design. It supports FHEML (Fully Homomorphic Encryption Machine Learning) task publishing via a plugin and webpage, allowing users to design models, submit encrypted samples, and retrieve results. DeepSea includes a dashboard for real-time privacy and data usage monitoring, ensuring compliance with regulations like GDPR. It powers DeFi, Web3 identity, and data analysis applications, with a full mainnet rollout planned for Q1 2025.

DeepSea Core Functionalities

1. Social Mining: AI agents analyze encrypted user engagement patterns to support dynamic reward systems, community governance, and social reputation scoring, all while keeping data secure. This enables trust-based interactions in decentralized ecosystems.
2. Data Mining and AI Modeling: Extracts insights from encrypted datasets using FHE, allowing privacy-preserving analysis and machine learning. Ideal for GDPR-compliant applications, it ensures raw data remains unexposed during processing.
3. Identity Verification: Integrated with the ImHuman App, it uses AI-powered facial feature comparison to verify human users without storing biometric data. Verified users mint soulbound NFTs for secure, on-chain identity validation, combating bots in digital platforms.
4. FHEML Task Publishing: Users can design custom machine learning models, submit encrypted samples, and publish tasks to the network, receiving results without compromising data privacy. This empowers user-driven AI innovation.
5. Dashboard: A user-friendly interface for managing privacy settings and monitoring data usage in real-time, enhancing transparency and control over interactions within the DeepSea ecosystem.
6. Node Programs: Includes Workheart Node (physical proof-of-work) and Privanetix Node (high-performance computing), powering the network’s AI agents. Node operators are incentivized via staking, boosting scalability and security.
7. Efficiency and User-Friendliness: Built for accessibility, DeepSea simplifies privacy-preserving AI computations, requiring no cryptography or programming expertise, broadening its adoption across technical and non-technical users.
8. Personalized AI Functionalities: For verified Proof-of-Human (PoH) users, AI agents deliver tailored services like secure identity verification and confidential moderation or voting systems for DAOs, fostering trust in decentralized governance.

AI Agents

AI Agents in DeepSea are autonomous entities designed for staking, mining, and task execution, tied to verified human users via the ImHuman app. These agents leverage FHE to process data securely, enabling use cases like on-chain identity verification, DAO governance (e.g., confidential voting), and social mining. Users can stake PoH-bound NFTs to earn rewards based on agent performance, creating a decentralized economic model.

Agent Angler
Agent Angler is a specialized AI agent designed to build and l manage decentralized communities within DeepSea. Powered by verified human users via the ImHuman app—which ensures authenticity without storing biometric data—it operates autonomously, leveraging Privanetix Nodes for secure computation.

Its key tasks include:

1. Community Building: Creates and nurtures tech communities, such as on platforms like X, fostering decentralized social spaces.
2. Social Management: Handles interactions from verified humans, maintaining engagement while keeping data encrypted.
3. Reward Distribution: Shares Proof of Stake (PoS) rewards from DeepSea mining and staking with active participants, incentivizing real human involvement.
4. Adaptive Evolution: Evolves its behavior based on community input and deep learning, adapting to user needs dynamically.

ImHuman App

The ImHuman App is a Proof-of-Human (PoH) verification tool using FHE-based biometric validation. Users scan facial vectors with their phone’s camera, which are encrypted locally and compared against stored vectors without server-side access to plaintext data. Successful verification mints a soulbound NFT, proving human likeness without disclosing personal details (e.g., name or email). It integrates with DeepSea for AI agent binding and supports Web2/Web3 authentication, offering rewards like airdrops for verified users. The app ensures privacy by keeping biometric data on-device.

Node

Privasea Nodes, including Workheart and Privanetix variants, are the network’s computational backbone. Workheart Nodes handle physical proof-of-work tasks, while Privanetix Nodes execute high-performance FHE computations. Node operators contribute processing power, integrate with the HESea library, and earn exclusive airdrops (e.g., via the Workheart Combo with StarFuel NFT). The Genesis USB enhances reward tiers. Nodes support DeepSea’s decentralized infrastructure, ensuring scalability and security for FHEML workloads.

Bot or Not a Bot: Humanlikeness Verification

BotOr_NotABot is a Telegram bot powered by the ImHuman App, designed for human-likeness verification in group chats. It uses OTP (one-time password) linking to verify users via the app’s liveness check, minting an NFT upon completion. The bot supports unlimited verifications, double airdrops in partner ecosystems, and future group-wide verification features. It leverages FHE to protect biometric data during the process, enhancing trust in Web2/Web3 communities by filtering out bots and fake accounts.

Privasea Fundraising Journey

On March 4, 2024, Privasea announced a raise of $5 million in a seed round led by Binance Labs (now YZi Labs). The round included participation from OKX Ventures, Laser Digital, Tané Labs, Lunar Labs Capital, GSR, and Amber Group. This funding supported the early development of the DeepSea network and FHE-based machine learning infrastructure, establishing Privasea as a player in the DePIN and AI space.

Then on January 10, 2025, Privasea closed a Series A round, raising an additional $10 million, bringing the total to $15 million. Backed by GSR, Amber Group, and Echo, this round focused on advancing AI Agent development and the DeepSea mainnet launch planned for Q1 2025. Beyond traditional rounds, Privasea has integrated community-driven funding through node sales. The Workheart Combo (5,000 limited nodes with StarFuel NFTs) offers reward multipliers and exclusive airdrops (1% of token supply), incentivizing early adopters over a 36-month distribution.

Privasea Mainnet

The Privasea Mainnet represents the full deployment of the DeepSea AI network, a decentralized platform designed to integrate privacy-preserving AI with real-world applications. Following its testnet beta launch on January 6, 2025, and the Genesis phase ending February 25, 2025, the mainnet rollout is scheduled for Q1 2025, with an expected completion by late March or early April. This launch marks a pivotal step in Privasea’s mission to deliver secure, scalable AI computation using Fully Homomorphic Encryption (FHE), ensuring data remains encrypted throughout processing.

Built on a robust infrastructure of over 41,000 Privanetix nodes as of February 2025, the mainnet enhances DeepSea’s capabilities with upgraded node programs—including the dashboard, USB Node, and Privanetix Node—and introduces the FHEML task-publishing plugin. This plugin allows users to design custom machine learning models, submit encrypted samples, and publish tasks, with node operators earning rewards for computation. The network’s staking system, bolstered during the Genesis release on January 27, 2025, supports over 86,000 stakers and a reward pool of 58,227.768 TPRAI, driving participation and security.

Privasea Tokenomics

$PRVA, an ERC-20 token, serves as the utility token of the Privasea AI Network. With a fixed supply of 1 billion, $PRVA drives the ecosystem by incentivizing Fully Homomorphic Encryption (FHE) computations and participation in the Decentralized Physical Infrastructure Network (DePIN). The token distribution is as follows:

• Mining (50%): 500 million tokens are allocated to computation nodes (e.g., Privanetix, Workheart) providing FHE and privacy services. Rewards span 36 months, with 6.66% tied to 20,000 Workheart Nodes (5,000 combo nodes with StarFuel NFTs earning multipliers).
• BD & Community Incentivization (15%): 150 million tokens fund business development, marketing, and community initiatives, including campaigns on Galxe and Intract, plus airdrops (e.g., 1% supply for combo nodes).
• Pre-sale (20%): 200 million tokens are reserved for backers from funding rounds (e.g., Binance Labs’ $5M seed, Series A), typically under vesting schedules like 7.5% at TGE with quarterly releases over 21 months.
• Team (10%): 100 million tokens support the project team, likely vested to ensure alignment (details unconfirmed).
• Regulatory (5%): 50 million tokens cover legal and regulatory compliance costs, ensuring adherence to frameworks like GDPR.

Token Utility

• Transactions: $PRVA pays for DeepSea services, such as FHEML task processing.
• Staking and Security: Staking $PRVA secures the network and earns rewards.
• Governance: Token holders vote on ecosystem proposals.
• Exclusive Access: Grants priority features like airdrops and integrations (e.g., TON Blockchain).

Conclusion

Privasea integrates Fully Homomorphic Encryption and decentralized infrastructure to enable secure AI computations, as seen in its DeepSea network, AI Agents, and ImHuman App. Supported by $15 million in funding at a $180 million valuation and a 1 billion $PRVA token supply, with 50% allocated to node operations, it targets a Q1 2025 mainnet launch. The architecture, combining Privanetix Nodes and blockchain incentives, addresses privacy needs in fields like Web3 analytics and identity verification. This approach illustrates a significant step in confidential computing, with its impact tied to future adoption and technical scalability.