AI and Restaking are widely recognized as the leading narratives of this bull market cycle. The former has produced various star AI projects, while the latter, centered around EigenLayer, has given rise to numerous LRT projects and various points-earning strategies. However, it is evident that these two narratives seem to have entered a mid-stage pause. Despite an increase in projects, they have become increasingly homogenous, making innovation from zero to one harder to find.

Moreover, while AI and Restaking are deemed “narratively correct,” this correctness does not imply perfection. Are many AI/Depin projects truly decentralized? Recent data also shows a decline in EigenLayer’s TVL. Can Restaking only ensure the security of Ethereum ecosystem AVSs?

Therefore, in the latter half of these hot narratives, projects that address critical common issues are the hidden gems awaiting discovery. From this perspective, Mind Network has caught our attention. It addresses the decentralization issues of numerous AI/Depin projects and enhances the utility and value of Restaking.

If EigenLayer is the restaking solution for the Ethereum ecosystem, then Mind is the restaking solution for the AI field. By utilizing more flexible restaking combined with fully homomorphic encryption for consensus security, it ensures the token economy and data security of decentralized AI networks.

Moreover, the project has already secured $2.5 million in seed funding in 2023, with participation from prominent institutions like Binance. Currently, it has deep collaborations with popular new AI/Depin projects like io.net and Myshell. The anticipation of the mainnet launch and incentive activities further raises expectations.

However, for most readers encountering this project for the first time, the combination of complex fully homomorphic encryption and profit-driven restaking may seem perplexing. How can these two elements work together to solve the key issues of AI projects?

In this issue, we delve into Mind Network to explore this promising project that integrates the trending narratives of AI, Restaking, and fully homomorphic encryption.

AI Projects Striving to Slay Dragons, Yet Gradually Becoming Dragons Themselves due to Failing to Achieve “Zero Trust”

To understand what Mind Network specifically does, it’s important to grasp the current challenges faced by AI projects. Perhaps, the narrative of the dragon slayer turning into the dragon itself serves as the best epitaph for describing today’s encrypted AI projects.

From a dragon-slaying perspective, the core narrative of encrypted AI (or DePIN) projects revolves around decentralization. This involves combating monopolies held by large corporations over AI elements such as computational power, algorithms (models), and data. The goal is to break the trust in authority held by these corporations.

While this narrative is correct and resonates deeply with the public, decentralized AI projects often face the challenge of potentially becoming dragons themselves:

They struggle to achieve “zero trust” among validators in decentralized environments.

Difficult to understand? Let’s explore specific examples.

For instance, in typical encrypted AI projects, participants often need to decentralize verification/voting on AI models to determine which performs best. However, in practice, validators (nodes) within the project typically decide which AI model performs best. How can you ensure that the model selected by validators is truly the best performer? Following a Proof of Stake (POS) mechanism doesn’t guarantee that the selection is optimal or fair.

Similarly, in AI agent businesses, ranking services based on performance raises the question of how to ensure that top-ranked services actually deliver the best results. In a DePIN scenario, when tasks are assigned to nodes within the network, how can you ensure that validators fairly distribute tasks to appropriate nodes rather than favoring familiar ones?

These examples highlight a critical common issue: in decentralized AI networks, decisions made by validators often become centralized points of trust. You end up having to trust validators or key participants within the network to make decisions correctly and without malice.

Projects advocating decentralization find themselves constrained by the need to trust internal network mechanisms. Achieving “zero trust” remains elusive, highlighting imperfections in the current narrative surrounding AI.

So, what’s needed in the face of these challenges?

Clearly, we need technological mechanisms and economic designs to minimize the trust dependencies on key participants involved in verification, voting, and decision-making within AI networks. This is precisely the area where Mind Network focuses and demonstrates its value.

The Holy Grail Of Fully Homomorphic Encryption Is Placed In The Most Suitable Position By Mind Network

Mind Network Excels in Leveraging Fully Homomorphic Encryption, Revered as the Holy Grail of Cryptography.

However, how does this relate to the issues exposed in AI and DePin projects mentioned above?

At their core, these issues all point to resource allocation, selection, and decision-making—beyond technology, they’re about governance.

Where governance allows for misconduct is where participants openly share known information (I know large holders are investing, so I’ll follow suit).

You’ve likely already sensed the potential of FHE:

What if information isn’t universally known anymore?

Fully Homomorphic Encryption (FHE), touted as the Holy Grail in cryptography, recently emphasized by Vitalik Buterin for its role in Web3. Here, we won’t delve into the mechanics of FHE; you just need to know its function—performing complex computations on encrypted data without decryption, offering a solution where data remains secure and private throughout analysis.

Yet, to embrace the Holy Grail means shouldering its weight.

While FHE’s encrypted calculations are impressive, they incur significant resource costs, making it impractical for AI model training in encrypted AI projects.

Mind Network’s approach to FHE hints at leveraging its potential effectively, positioning the Holy Grail where it belongs.

Without using FHE for AI model training and parameter adjustments, but rather employing “human governance” in areas such as cross-validation, selection, ranking, and voting after the AI model has been trained, resource expenditures can be controlled. The problems to be addressed are also very clear:

If participants in the AI network conduct business without knowing each other’s choices/voting results, there will be no “following the big players” or blindly copying behaviors from authoritative nodes. This eliminates decision biases caused by identity influence, returning decentralized decision-making to its essence, thereby identifying truly effective AI models and AI services.

Therefore, using FHE for general computing faces significant obstacles, but employing FHE for specific decentralized stages like validation is internally consistent and feasible. Ensuring zero-trust in the validation process achieves consensus security for encrypted AI projects and genuine decentralization.

On the other hand, security is synonymous with fairness.

We can also use a specific case to see how Mind Network’s fairness is reflected in the encryption execution of Validation:

• 1. An AI project integrates the Validation SDK provided by Mind, accessing homomorphic encryption verification services.
• 1. Simultaneously, the AI project registers on the Mind network to verify its identity. Mind generates a smart contract on the target project network/chain to synchronize subsequent operation changes and execution results.
• 1. The AI project publishes validation tasks on the Mind network that require fully homomorphic encryption (FHE), such as determining which AI model performs better. The FHE voting service begins, allowing validation nodes of the AI project to vote without seeing plaintext results, yet still executing the voting process.
• 1. Voting results and related data changes are transmitted to Mind’s own chain via the smart contract and promptly synchronized and recorded.
• 1. In the above steps, when the AI project utilizes Mind’s services, Mind’s project tokens will be charged as gas fees (tokens not yet issued).

Similarly, when it comes to a project like DePIN, using Mind Network can also lead to more equitable resource allocation. Let’s take IO.net, which collaborates with Mind Network, as an example:

• 1. IO.net integrates the product SDK provided by Mind to access fully homomorphic encryption verification services.
• 1. After integrating these services, nodes with GPUs gain consensus capabilities under fully homomorphic encryption. This means that when AI computing tasks arrive, both requests and data are encrypted, making it feasible to distribute tasks fairly among suitable nodes.

Wait, But What Does This Have To Do With Restaking?

The preceding discussion seems entirely focused on the technical aspects. What does this have to do with asset-level Restaking? Mind Network provides an FHE-based solution, which technically enhances the security of AI network validation. However, to participate in and benefit from this security validation, it’s closely intertwined with the economic network structure of most AI/Deepin projects.

PoS, or Proof of Stake, is the foundational consensus logic for the majority of cryptocurrency projects. Therefore, the size of staked assets under each node’s governance and their eligibility to participate in fair validation guaranteed by FHE are closely related.

Mind Network’s pivotal move at the asset level is to expand Staking and Restaking openly, safeguarding AI network validation consensus through homomorphic encryption. Different roles in the network can thus meet their respective interests.

For validation nodes of AI projects, increasing the amount of Restaking provides more opportunities and voting rights in conducting FHE validation tasks within Mind Network.

For regular users, delegating their LST/LRT assets to the aforementioned nodes in a proxy staking manner allows them to earn APR income.

This seems to bear similarity to EigenLayer’s Restaking familiar to us, fundamentally converging on different paths:

EigenLayer secures various AVS in the Ethereum ecosystem using restaking; Mind Network secures consensus for different AI networks across the entire cryptographic ecosystem using restaking.

It’s worth noting that the concept of “entire ecosystem” is inseparable from another key feature of Mind Network: Remote Restaking.

Because of Remote Staking, there’s no need to cross-chain one’s LRT tokens across different chains. Instead, users can stake their LRT tokens from different chains to a validation node of an AI network through remote staking. This significantly lowers the entry barrier for user participation and integrates fragmented liquidity in a multi-chain environment.

Broad Ecosystem Development and Solid Technical Strength

Mind Network currently has several more catalysts worth noting:

Firstly, on the product front, the test network has attracted 650,000 wallets and processed 3.2 million transactions, indicating promising prospects for the full mainnet functionality rollout.

Secondly, in ecosystem development, given its platform’s focus on empowering other AI projects, collaborations with top-tier projects are crucial. Presently, Mind Network provides AI network consensus security services to io.net, Singularity, Nimble, Myshell, AIOZ, among others. It also offers an FHE Bridge solution for Chainlink CCIP and AI data security storage services for IPFS, Arweave, Greenfield, and similar projects. These partnerships span leading AI, storage, and oracle projects, potentially positioning Mind Network as a “golden shovel.”

Additionally, from a background perspective, in 2023, the project was selected by Binance Incubator and completed a $2.5 million seed round financing involving prominent institutions like Binance. It has also received grants such as the Ethereum Foundation Fellowship Grant, inclusion in the Chainlink Build Program, and becoming a Channel Partner of Chainlink.

In terms of technical capabilities, apart from its team comprising top-notch professors and PhDs specializing in AI, security, and cryptography from leading universities and enterprises, a noteworthy focus lies in collaborations with industry-leading fully homomorphic encryption research companies.

In February this year, Mind Network officially announced a partnership with ZAMA, a leading open-source encryption company in the field of fully homomorphic encryption (FHE). ZAMA recently completed a $73 million Series A financing round led by Multicoin and Protocol Labs.

More recently, the collaboration between Mind Network and ZAMA has expanded to jointly launch a new Hybrid FHE AI network. This initiative aims to advance AI algorithms’ applications in encrypted data, adding another layer of technological advancement to the project.

According to sources close to the matter, Mind Network chose to utilize ZAMA’s foundational technology library in its own R&D efforts. This strategic decision demonstrates Mind’s expertise in optimizing FHE resources, ensuring maximal security capabilities without compromising performance.

Beyond enhancing its own capabilities with superior technology, Mind Network is also contributing to the improvement of the crypto ecosystem. In May, the project partnered with Chainlink to introduce the first FHE interface based on the Cross-Chain Interoperability Protocol (CCIP). This collaboration enhances the security of cross-chain communications and transactions while fostering a more trustworthy and user-centric Web3 ecosystem.

As of the latest update, Mind Network has established partnerships with numerous top-tier projects across different ecosystems and fields. Given its focus on empowering other projects, it may be poised to achieve a “golden shovel” effect in the future.

Conclusion

When fully homomorphic encryption meets Restaking, Mind Network could indeed become a new driving force in the latter half of this year’s crypto mainstream narrative.

Using fully homomorphic encryption as a mediator, Mind Network aims to optimize business operations for numerous encrypted AI projects, providing genuine decentralization and zero-trust support for decentralized AI initiatives. Meanwhile, Restaking paves the way for further liquidity absorption across different chains, potentially leading to a rapid increase in Total Value Locked (TVL) for the project.

It’s undeniable that the allure of fully homomorphic encryption as a holy grail is capturing market attention with new narratives. Simultaneously, Restaking attracts market liquidity. As consensus security for AI projects becomes more accessible, the concentration of attention and liquidity is likely to foster the project’s future development.

Projects like Mind Network, which refine the narrative (AI, Restaking) through their own technology, could indeed represent a gentler form of disruption in the latter half of the mainstream narrative?

Disclaimer：

1. This article is reprinted from [Techflow]. All copyrights belong to the original author [Techflow]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.