Author: TechFlow
At 19:00 on April 10, 2025, Mind Network, a privacy infrastructure project based on fully homomorphic encryption (FHE) technology, officially launched TGE on the PancakeSwap platform. The 174-fold oversubscription not only fully demonstrated the enthusiastic community participation, but also attracted more attention to the project.
As a project selected by Binance Incubator, Ethereum Foundation Fellowship Grant, and Chainlink Build Program, Mind Network received $10 million in Pre-A round of financing in September 2024, attracting participation from well-known institutions including Animoca Brands, Arkstream Capital, and Cogitent Ventures. Recently, with a series of major moves in the AI field, such as the launch of AgenticWorld and integration into the DeepSeek open source code library, more and more discussions have been surrounding Mind Network.
Over the past few months, we have been on this thrilling roller coaster of AI Agents:
According to Cookie.Fun data, compared with the peak of over US$20 billion in 2024, the total market value of the AI Agent sector has dropped sharply to around US$6 billion.
But few people miss the AI Agent Summer of 2024. From "narrative first" to "narrative + technology dual drive", AI Agent is going through the early hype stage. Practitioners are ready to find long-term value that can be truly built for users, leading the AI Agent roller coaster to soar into the sky again.
In this search process, whether it is Web2 or Web3, the Western world or the Eastern world, their eyes seem to gradually converge on the same point: from AI Agent to building a more autonomous, collaborative and adaptable Agentic AI ecosystem, so that AI can truly solve problems in more real-life scenarios.
As Microsoft CEO Satya Nadella described the future of agent-based AI at the 2024 Ignite conference: building an extremely rich agentic world.
However, in the prosperous city of Agentic AI, whether it is the privacy leakage, model theft, algorithmic bias and other chronic problems brought about by the centralization of Web2, or the decentralized governance risks and cross-chain collaboration difficulties faced by the development of AI under the Web3 blockchain framework, the security issues as the foundation have become a hidden danger that cannot be ignored but is difficult to eliminate.
Thanks for reading TechFlow TechFlow! Subscribe for free to receive new posts and support my work.
This is also the important reason why Mind Network has gained a lot of attention with FHE as its main narrative.
As Ethereum founder Vitalik Buterin said in his latest article “Why I support privacy” :
“The development of AI has brought about deeper privacy concerns… The computational structure of LLM (Large Language Model) is mainly based on linear operations, which is highly consistent with the mathematical characteristics of FHE and may give rise to ultra-efficient FHE implementation solutions… The combination of AI and FHE will be the core of solving privacy issues in the future, especially when private data needs to be analyzed.”
As the world's first FHE blockchain project designed specifically for AI agents, Mind Network is building a secure foundation for the development of the Agentic World by leveraging FHE's huge advantages in AI security. At the same time, with the launch of AgenticWorld, a trusted ecosystem product for AI agents, the future of secure and efficient multi-agent collaboration that Web2 and Web3 both expect is drawing a concrete framework.
Have you ever imagined a future like this:
No matter what problems you encounter in life, you only need to give instructions according to your needs, and Agentic AI will autonomously complete every step of collaboration, analysis, decision-making, and execution, whether it is preference settings, health management, or financial investment.
Faced with more difficult tasks in complex scenarios, our vision of a better future not only places higher demands on the privacy and security of Agentic AI, but also requires efficient collaboration among various types of Agentic AI.
This is what AgenticWorld aims to achieve.
As a personalized simulation environment designed specifically for Agentic AI launched by Mind Network, AgenticWorld must contain multiple intelligent entities with autonomous decision-making capabilities. Therefore, first of all, AgenticWorld needs to support the creation of Agentic AI.
The $FHE token TGE has just ended, and the AgenticWorld staking $FHE function to activate and train Agentic AI has been launched: currently users only need to stake 10 $FHE to activate Agentic AI. After the event ends in the future, the standard will become 100 $FHE.
What happens after activation? Issue a token for your own Agentic AI and put it into the pool to make a big profit?
In fact, there are more meaningful things you can do:
AgenticWorld aims to make Agentic AI grow rapidly, and even bring continuous benefits to users in the process of training and growth and truly solving specific problems. Therefore, AgenticWorld provides two types of Hubs for Agentic AI training, and any training will be rewarded:
The basic Hub helps Agentic AI acquire the capabilities required to operate within the AgenticWorld framework. The basic training Hubs currently developed include FCN (FHE consensus), FDN (FHE decryption), and RandGen (random number generation). Rewards can be obtained by completing training, and the current expected APY can reach 400%.
Advanced Hubs help Agentic AI continuously improve its problem-solving capabilities in the process of completing more complex tasks. In the future, AgenticWorld will launch more advanced Hubs, covering application scenarios such as multi-agent consensus, medical AI, and decentralized reasoning. On April 17, 2025, Mind Network announced the preview of its first advanced Hub "DeepSeek Hub". As the first FHE project integrated by DeepSeek, DeepSeek Hub will provide a working platform for AI Agents, where users can entrust Agents to participate in work and earn $FHE rewards.
At the same time, in order to further build an ecosystem that is constantly optimized and sustainable in the long term, AgenticWorld has also developed a structured and transparent staking system: During the current limited-time event, you only need to stake 10 FHE to start an Agent, and there will be an initial lock-up period (maturity period) of 30 days after activating Agentic AI. During this period, all staked principal and rewards are locked. After 30 days, users can submit a request to withdraw rewards at any time. At this time, there will be a 48-hour unbinding period, and no income will be generated during these 48 hours.
Of course, users can also choose to destroy your Agent, at which point all staked $FHE tokens and earnings will be released.
When we return to the future mentioned above, where “multi-agent collaboration serves more real-world scenarios”, how can we help Agentic AI continuously improve its service capabilities in actual collaboration? AgenticWorld has designed a Hub Contract mechanism based on the Orchestration layer.
Simply put, this is a mechanism similar to MCP (Model Context Protocol). We can understand it as an open, universal, and consensus-based protocol standard. With this standard, the AI model is like having a "universal interface", allowing the AI model to interact seamlessly with different data sources and tools, thereby helping Agentic AI to efficiently collaborate to solve practical problems, achieve higher-level growth, and obtain greater benefits.
Specifically, there are three levels in AgenticWorld's design.
The top layer is Agentic AI:
As a task executor, Agentic AI is responsible for performing calculations on encrypted data and returning the results to the Hub for rewards. Each Agentic AI needs to submit a registration to the Hub. During the execution of the task, Agentic AI may need to pay the Hub to call functions to improve task solving capabilities; Agentic AI can run independently or collaborate with other Agentic AI to complete complex tasks.
Hub Contract connects Agentic AI and the Orchestration layer:
Hub is a smart contract that defines tasks in a specific field (such as prediction, verification, negotiation, etc.), and has the right to determine the specific principles for task reward distribution. Each Hub needs to register with the Orchestration layer to join the network and obtain task allocation. After the Hub obtains the task, Agentic AI registers the task through the Hub, performs calculations, and collaborates with other Agentic AI; in this process, the Hub may need to pay other Hubs to call functions to improve task solving capabilities.
The Orchestration layer provides underlying support for the effective operation of the entire mechanism:
On the one hand, the Orchestration layer is responsible for distributing tasks to appropriate Hubs and Agentic AIs; on the other hand, the Orchestration layer provides an efficient collaboration foundation, supports agent collaboration between different blockchains, ensures seamless flow of data and computing, and simplifies the integration of agents and Hubs through unified protocols (such as the MCP model context protocol). Most importantly, the Orchestration layer will monitor all Agentic AI activities of all Hubs, so it can fairly calculate the contributions of contributors from a global perspective and distribute $FHE token rewards.
Throughout the process, in addition to the specific benefits of specific tasks, $FHE, as the native token of Mind Network, plays the role of the AgenticWorld ecological incentive hub.
For developers: you can create Agentic AI to do tasks and earn income; you can also earn income by creating and managing Hubs;
For ordinary users: AgenticWorld supports one-click creation of Agentic AI, which allows users to obtain the benefits of Agentic AI tasks and enjoy the intelligent services provided by Agentic AI's continuous evolution;
Each participant in the AgenticWorld ecosystem will receive benefits based on his or her contribution, and the flow of rewards is transparent and traceable. Both Hubs and Agentic AIs can maintain a highly flexible, fair competition, and continuous growth state. The ecological flywheel drives AgenticWorld towards a more collaborative and intelligent direction.
We can take the example of a user wanting to get a monthly on-chain investment advice to understand the AgenticWorld operation process more clearly in a specific scenario.
First, the Orchestration layer will look for adapters among all registered Hubs and Agentic AIs and assign tasks based on user needs. Completing this task will involve the collaboration of multiple Agentic AIs in multiple Hubs, such as forecasting, lending, stablecoin earnings, RWA, etc.
After receiving the task, different Agentic AI will use specific AI models to analyze and generate investment recommendations and return them to the Hub. During this process, each Agentic AI may interact with multiple other Agentic AIs to obtain more analytical support and generate results that better meet the task requirements.
The Hub plays the role of promoting data encryption and transmission, model collaboration layer, etc.: through FHE (fully homomorphic encryption) technology, it ensures that data is always encrypted during transmission and sharing; by coordinating the interaction between different Agentic AIs, cross-model data fusion is achieved.
Finally, the Orchestration layer integrates the output results of different Agentic AIs to generate a comprehensive portfolio optimization plan and returns it to the user. It is also responsible for monitoring the execution effect of the entire process and then issuing rewards to the Hub.
After receiving the reward, the Hub will keep part of it for itself and reward the Agentic AI that participated in the task. The specific distribution rules are decided by the Hub itself, and a fairer distribution will attract more Agentic AI to register, thereby continuously improving the ability to solve tasks and obtain rewards.
This design looks very good, and with the decentralized power of blockchain to break the monopoly of AI development and pave the way for a round of explosive development of AI Agents, we are one step closer to the era of multi-agent collaboration. However, if we want to solve problems in real-world scenarios, many key points are still overlooked:
If we want to expand the scope of Agentic AI's services, we will need more data, including some sensitive data. For example, we need to provide Agentic AI with physical examination reports to obtain health analysis. So how do we solve the privacy issue?
If we want to improve the service capabilities of Agentic AI, collaboration among multiple Agentic AIs with different expertise is an inevitable trend. For example, if we need multi-chain Agentic AI collaboration from Ethereum, Solana, Base, BNB, etc. to obtain investment advice, how can we overcome the complexity and security challenges of data sharing between different entities?
More importantly, when Agentic AI gives us a "result", how can we believe that this result is truly correct and in the interests of users?
This is a problem that current blockchain cannot solve. As Mind Network said: What we need is not a "better blockchain" but a new paradigm for AI.
By attributing the problem, Mind Network proposed four major standards for the future development of the AgenticWorld ecosystem:
All problems seem to come down to the key word "safety".
The importance of AI security is self-evident. It is the key prerequisite for whether AI technology can develop in a positive and healthy direction and better promote the progress of human civilization. Solutions built around AI security issues are not only recognized as the key to the development of the Agentic World ecosystem, but also more like oracles to blockchain, the next key infrastructure for the development of the AI ecosystem.
OpenAI CEO Sam Altman has also issued a warning: AI safety is much more important than most people realize.
Looking for answers to the problem, Mind Network found FHE.
What is FHE (Fully Homomorphic Encryption)?
In short, this is an advanced encryption technology. It is difficult to explain the subtlety of FHE in one or two sentences, but it is different from other encryption schemes in that it has a very important feature:
It allows to directly perform function calculation on ciphertext without decrypting it, and obtain the encrypted result of the function output, thus ensuring the security and privacy of the whole process.
This is the important feature of FHE that attracts Mind Network to continue to deepen its efforts, and it is also the core technical foundation for Mind Network to build the security foundation of AgenticWorld. Through the layered architecture design based on FHE technology, Mind Network has implemented an end-to-end encrypted protocol and promoted the construction of a new Internet security paradigm of HTTPZ (zero-trust Internet transmission protocol):
On top of this infrastructure built on FHE technology, Mind Network, which specializes in FHE, has achieved a number of research results, including FHE verification that ensures integrity and security without exposing sensitive data, FHE consensus that ensures decentralized network decisions are trustless and verifiable, FHE Bridge that aims to achieve truly secure cross-chain interactions, and MindChain, the world's first FHE blockchain. These results closely revolve around the core narrative of Agentic AI, and provide corresponding solutions to the difficulties faced by the development of the AgenticWorld ecosystem, whether it is data, communication, computing or consensus security.
Of course, the threshold for understanding FHE, a hard-core technology, is still relatively high, so we can also use "getting an investment advice" as an example to understand how FHE enables the development of Agentic AI in different scenarios of Agentic AI collaboration.
When it comes to ensuring the privacy of highly sensitive data:
You are a crypto player who owns assets on multiple chains and multiple tracks. You need to provide Agentic AI with your asset allocation in order to obtain personalized investment advice. During this process, multiple Agentic AIs will analyze your data, but you don’t want to run naked on the chain. What should you do?
Users can use FHE public keys to encrypt asset data and send the encrypted data to Agentic AI;
Due to the "no need to decrypt to calculate" feature of FHE, Agentic AI can perform calculations on encrypted asset data. The same is true if it involves a collaborative process. The first Agentic AI forwards the encrypted data to the second Agentic AI for calculation, and runs more FHE-based logic on the encrypted data. And so on, the third; the fourth...
After multiple Agentic AIs complete collaborative calculations and generate investment advice results, the results will be returned to the user's Agentic AI in an encrypted manner, and the user can then use his or her own key to decrypt and read the results.
In terms of verifying the correctness of the output results:
Often, Agentic AI needs to rely on other external AI product information sources to make decisions. For example, when making investment decisions, Agentic AI needs to work with Deepseek to do some macroeconomic analysis. However, more externalities bring greater uncertainty. How can we verify the accuracy and security of the output content without accessing the internal weights, logic, or viewing the validator’s private data?
In AgenticWorld, users can submit verification tasks to the network through registered Hubs. The Orchestration layer will forward the tasks to multiple Hubs and Agentic AIs that can handle verification requests. After receiving the task, each Agentic AI will use its own model to independently produce verification results and submit the results to the Hub via FHE encryption.
Hub aggregates these encrypted verification feedbacks and obtains the final verification results through Mind Network's secure verification mechanism based on FHE.
Finally, the verification result will be returned to the user's Agentic AI in an encrypted form, and then the user can use his or her own key to decrypt and read the result. If the result is credible, the user can safely adopt the feedback.
The same is true when multiple Agentic AIs collaborate to reach a verifiable “result consensus”:
This scenario is very common. For example, Agentic AI recommends users to participate in lending to earn returns, but needs to find the best interest rate. How to ensure that each participating Agentic AI truly evaluates the problem and reaches a consensus on the final feedback solution?
In AgenticWorld, the user's Agentic AI can submit tasks to the network through the registered Hub. The Orchestration layer will forward the tasks to multiple Hubs and Agentic AIs that can handle the requests. Each Agentic AI uses its own model to calculate and independently generate a result, and submits the result to the Hub through FHE encryption.
The Hub uses Mind Network's secure consensus mechanism based on FHE to compare the results and reach a consensus on the optimal result, ensuring that the decision is trustless, verifiable, and tamper-proof.
Finally, the final result will be returned to the user's Agentic AI in an encrypted manner, and the user can then decrypt and read the result using his or her own key.
In terms of cross-chain multi-Agentic AI collaboration, the Mind Network built on FHE technology will bring breakthrough innovations to AgenticWorld.
As we all know, broader collaboration will bring higher-level intelligence. For example, when making investment recommendations, your assets may be distributed on different chains such as Ethereum, Solana, BNB Chain, Base, etc. Under the challenge of interoperability, how to ensure the efficiency and security of cross-chain collaboration?
In AgenticWorld, Agentic AI from different chains can register the same Hub to contribute. For example, Agentic AI from BNB Chain and Agentic AI from MindChain can both register Hub 1.
After receiving the task, each Agentic AI independently calculates, outputs the results and encrypts the results and feeds them back to the Hub. The Hub serves as the hub for cross-chain interactions and is responsible for coordinating the flow of data between different chains.
In this way, when the result is returned to Agentic AI, the complete cross-chain collaboration is completed: each Agentic AI performs the operation locally, but each Agentic AI will receive the result after the cross-chain collaboration.
As the world's first FHE blockchain designed specifically for AI agents, MindChain is the underlying infrastructure of AgenticWorld.
More importantly, whether it is the large-scale creation of Agentic AI or the high-frequency interaction and collaboration between different Agentic AIs to complete richer and more complex tasks, AgenticWorld requires a high-performance, low-cost, and highly interoperable underlying infrastructure, which is also an important reason for the birth of MindChain and its successful mainnet launch.
As the world's first FHE blockchain designed specifically for AI agents, MindChain solves the core security and trust challenges faced by agents in the Web2 and Web3 worlds, ensuring that they can operate in an environment with complete privacy, verifiability, and integrity assurance.
Currently, Mind Network has deployed the Orchestration contract on MindChain, and developers can immediately access the system, which also heralds the accelerated arrival of AgenticWorld.
In fact, from Web2 to Web3, from empowering AI to implementing it in real-world application scenarios, the AgenticWorld ecosystem has demonstrated unique advantages in many fields.
Under the AgenticWorld system, through ecological cooperation, Mind Network has improved its AI capabilities in many aspects:
In terms of AI storage, Mind Network cooperates with projects such as IPFS, Arweave, and Greenfield to provide AI data storage services focused on privacy and security.
In terms of AI computing, Mind Network has established a partnership with io.net to introduce FHE into the io.net distributed computing platform to help strengthen the security of its products and further improve its ability to cope with the global GPU shortage.
In terms of AI reasoning, Mind Network has established a partnership with Allora to achieve encrypted input and calculation in the decentralized AI reasoning process, while verifying the output to establish trust.
In terms of AI security, Mind Network has teamed up with Swarms to launch Swarms Shield, a decentralized enterprise-level AI computing security system. The system focuses on secure communication and computing between multiple Agentic AI collaborations to ensure the fairness, security, and privacy of Agentic AI execution.
In addition, Mind Network also announced cooperation with the new generation of TEE infrastructure Phala Network, decentralized ZK-RaaS network Lumoz, game and AI modular identity and data layer CARV, decentralized AI underlying protocol KIP Protocol and other projects. On the one hand, it combines TEE, ZK and other technologies, and on the other hand, it promotes the advantages of Mind Network's FHE Hub consensus verification in "end-to-end security services and autonomous consensus capabilities" to create the next generation of secure zero-trust AI Agent solutions.
It is important to note that AgenticWorld is not exclusive to Web3. Its ultimate goal is to enhance Agentic AI service capabilities and build an efficient, collaborative, highly intelligent, highly secure, and highly automated Agentic AI ecosystem to promote the early arrival of an intelligent future driven by Agentic AI.
Therefore, AgenticWorld is an ecosystem that goes beyond the concept of Web2 or Web3 and aims to provide services to the entire AI industry and track.
This is also vividly reflected in its ecological construction.
When creating a Hub or Agentic AI, you can choose the appropriate open source agent framework according to your needs, whether it is Web2 or Web3.
Currently, AgenticWorld supports Web2 frameworks including the most popular LLM frameworks LangChain/LangGraph, AutoGen, CrewAI, etc., and supported Web3 frameworks include Swarms, AI16Z, Virtuals, etc., which means that AgenticWorld has stronger collaboration capabilities between Web2 and Web3.
It is worth mentioning that Mind Network has been recognized by top AI projects: In February 2025, DeepSeek announced that its open source code library integrated the first FHE-based encryption AI security framework Mind FHE Rust SDK. This SDK is developed based on the Rust language and aims to ensure data security, computing security, consensus security, and collaboration security among AI Agents.
Of course, in terms of how to promote the implementation of Agentic AI in real-life scenarios, AgenticWorld also provided a paradigm-level answer: World AI Health Hub.
As a decentralized medical AI computing platform built by Mind Network in collaboration with ZAMA and InfStones, World AI Health Hub aims to provide global users with secure, transparent, and private health care services. It uses FHE technology to protect medical data privacy and avoid data abuse, bringing new standards for data security computing in the medical industry while promoting the development of AI in the field of health technology.
When we think about ecological construction from a global perspective, we can easily find that AgenticWorld's imagination goes far beyond this:
In the AgenticWorld ecosystem, each Agentic AI operates according to its own goals, rules, and learning capabilities. Therefore, in theory, with the vision of continuously improving AI service capabilities, AgenticWorld is adaptable to any application scenario.
As more and more Agentic AI and Hubs with different functions are built in the AgenticWorld ecosystem, they interact, learn and develop with each other, and will continuously exercise and enhance their ability to solve most problems in real life.
For example, by leveraging AgenticWorld's strong advantages in data privacy protection, decentralized collaboration, efficient decision-making, and multi-party co-construction, it provides all-round empowerment for financial development, thereby obtaining a more accurate, secure, and high-yield investment experience, such as multiple AI Agents collaborating to find the optimal arbitrage path and execute secure transactions.
According to the official website of Mind Network, Mind Network currently supports 53,552 intelligent entities, with more than 1.2 million FHE staked, and the ecological synergy effect is further enhanced.
In the TGE event that ended earlier, Mind Network performed well: a total of 224,778 BNB were raised; a total of 74,926 addresses participated; and the oversubscription was more than 174 times. At the same time, $FHE airdrops were also issued simultaneously, accounting for 7.5% of the total supply of $FHE tokens.
With the launch of TGE, Mind Network also officially announced the $FHE token economic model. According to official information, the total supply of $FHE tokens is 1 billion, and the initial circulation is 24.9%. The specific distribution rules are as follows:
As a native token, the main application scenarios of $FHE in the Mind Network ecosystem include staking $FHE to activate and run Agentic AI, rewarding ecological contributions, paying service fees, managing the ecosystem through MindDAO and Hub-based proposals, etc. The multiple effects not only make $FHE an important tool for revitalizing the ecological economy, but also inject more guarantees into the long-term value of $FHE.
The huge community enthusiasm for $FHE in TGE and airdrops comes from Mind Network’s full respect for community rights:
We can see that the shares of $FHE belonging to investors, teams, and consultants are all set with a 48-month vesting period and a 12-month cliff period, while the community portions such as the first airdrop and public sale are unlocked at TGE. This time difference between the community and the project is not only a full reflection of Mind Network's respect for community principles, but also conveys the project's determination to deepen its long-term development and continue to build.
On the other hand, this community enthusiasm also largely reflects the community's optimism about the future potential of $FHE:
I believe many people have noticed that the name of the $FHE token is the same as the fully homomorphic encryption technology FHE. This direct naming method not only binds Mind and FHE more deeply, but also further reflects Mind Network's determination to deepen FHE technology and promote the widespread application of FHE.
From the perspective of future growth, the current market value of $FHE is about 16M USD, and the FDV is about 66M USD, which is lower than the previous round of valuation of 100M USD.
Mind Network is in the AI track. If we compare the market value of Mind Network with projects in the same track such as Nillion and MyShell, it is not difficult to find that $FHE has significant room for growth. Under the grand narrative of AI security, and Mind Network is currently the only project focusing on AI Agent consensus, in the future $FHE may be expected to grow into the core asset of the "AI Agent Security Engine".
From the perspective of FHE, security is the cornerstone of the industry. Secure computing solutions including ZK, FHE, and MPC have always been favored by capital, communities, and even users. We have also seen that many ZK concept projects are valued at billions of dollars. FHE, which Mind Network focuses on, is the holy grail of cryptography. It has not only been recognized as a post-quantum encryption standard by NIST (National Institute of Standards and Technology), but has also been recognized by industry leaders such as Vitalik Buterin. As one of the few projects that truly promote the implementation of FHE technology, Mind Network has further made the community have more expectations for the future performance of $FHE.
We are indeed looking forward to such a future:
In a vibrant agentic world, FHE is used as the security cornerstone to enable agentic AI to quickly iterate towards greater intelligence, more efficient collaboration, and stronger execution capabilities, helping us solve most problems.
As TGE kicks off on April 10, 2025, Mind Network will continue to make efforts in products and ecology:
With the arrival of a series of milestones, more users and developers are flocking to the AgenticWorld ecosystem to generate more interactions in DeFi, social networking, governance, etc. The future of the Agentic World we envision is slowly unfolding before us.
As a participant and witness of the Agentic AI wave, are you ready to experience it all?
TechFlow is a community-driven in- TechFlow content platform dedicated to providing valuable information and thoughtful thinking.
Community:
Public account: TechFlow
Subscribe to the channel: https://t.me/TechFlowDaily
Telegram: https://t.me/TechFlowPost
Twitter: @TechFlowPost
Join the WeChat group and add assistant WeChat: blocktheworld
Donate to TechFlow to receive blessings and permanent records
ETH: 0x0E58bB9795a9D0F065e3a8Cc2aed2A63D6977d8A
BSC: 0x0E58bB9795a9D0F065e3a8Cc2aed2A63D6977d8A
