Introduction

#Phala Network was established in 2018. At the beginning of the project, Phala Network positioned itself in the privacy computing track, combining the smart contract execution environment with TEE to achieve safe and reliable smart contract execution. At the same time, Phala Network also provides a complete privacy protection mechanism to ensure that users’ data is in their own hands. In 2023, because the US stock market drove the trend of AI in the market, Phala Network closely followed the trend of AI at that time and started an artificial intelligence agent. It also followed the concept of modularization and built itself into The execution layer of artificial intelligence.

Project basic information

Basic information

Website: https://phala.network/

Twitter: https://twitter.com/PhalaNetwork, 122,000 fans

TG: https://t.me/phalanetwork

Discord: https://discord.com/invite/phala-network

white paper: https://docs.phala.network/

Online time: Token will be online in 2020

Project team

core team

Marvin Tong: Founder & CEO. Graduated from Communication University of China with a master’s degree. After graduation, he has been working as a product manager in companies such as Tencent and Didi.

Hang Yin: Co-founder & Chief Technical Development. Graduated from Fudan University in China. He was previously an expert in Google search ML-related fields, one of China’s Polkadot ambassadors, and the founder of Bitcoin Gold.

Zhe Wang: Co-Founder & COO. Graduated from Huake University in China with a master’s degree. He is an expert at Huake Computer Hardware Laboratory, CTO & CEO of Bitcoin Gold, Xiaohei Technology and Haha Retail.

Jun jiang: Co-founder & CTO. Used to be the webmaster of RubyChina. Former CTO of KnewOne, former software architect of DJI.

Advisory team

Shun Fan Zhou: One of the authors of the Phala Network white paper is a Ph.D. from the System Software and Security Laboratory of Furi University. He has published research on attack transactions and defense methods in the E-Dafang ecosystem at USENIX Security, the top international security conference, and has collaborated on multiple publications at top international security conferences. paper.

Sandro Gorduladze: Angel investor and partner of HASH CIB. Sandro established the research department at HASH, which has gained a reputation for its in-depth reporting. Before joining HASH, Sandro worked at PricewaterhouseCoopers in Russia, providing tax consulting for companies in the TMT industry.

Konstantin Shamruk: PhD in Economics from the University of Toulouse, France. Led the Boban analysis work on the economic design of Phala Network.

Jonas Gehrlein: Research Scientist, Web3 Foundation. He is responsible for studying the economic issues of the Polkadot ecosystem at the Web3 Foundation. Prior to joining W3F, Jonas received his PhD in Behavioral and Experimental Economics from the University of Bern, where he studied human behavior in markets and organizations. Prior to this, he obtained a master’s degree in quantitative economics at the University of Konstanz.

From Meckbach: Senior Polkadot Ambassador, researcher and advocate for Web3 and cybersecurity. She is currently COO of MH-IT& Service GmbH and held an application analytics position at Google before joining MH-IT.

Financing status:

Phala Network has raised approximately $10 million in a funding round.

Seed round

In July 2020, Candaq Group, IOSG Ventures, SNZ, Incuba Alpha Group, nfiChain, Exoplanet Capital and Blue Mountain Labs invested in this round, amounting to US$10 million.

Development strength

The Phala Network project was established in 2018 and was founded by founder Marvin Tong. Key events in project development are shown in the table:

Judging from Phala Network’s project development roadmap, the Phala Network team was able to complete the project schedule on time at various critical time points of the project. It can be seen that the strength of the technical development team led by Marvin Tong and Hang Yin is very Strong, although Phala Network has broader development prospects after entering the AI ​​track, the technical challenges it faces will become greater, and the requirements for the technology development team will become higher.

Technical features

Phala Network is a privacy computing infrastructure based on the Polkadot ecosystem, aiming to achieve data confidentiality and privacy protection through Trusted Execution Environment (TEE) technology, cloud computing services and cross-chain data confidentiality layers. At the same time, Phala Network uses artificial intelligence co-processors to provide secure and verifiable computing solutions for integrating encryption and artificial intelligence at the application layer.

It can be simply understood that Phala Network adds an artificial intelligence co-processor to the original privacy computing infrastructure, so that artificial intelligence agent contracts can be deployed on Phala Network. Phala Network has become the execution layer of Web3 artificial intelligence. Phala Network has protected its privacy through TEE technology. Top-level LLMs can now be accessed through Phala Network’s artificial intelligence agent contract, and control is derived from smart contracts written on the chain.

Schematic diagram of Phala Network’s AI Agent (picture from Phala Network white paper)

Before Phala Network added the artificial intelligence agent contract, the main technical feature of Phala Network was to achieve data confidentiality and privacy protection through a variety of technical means. First, Phala Network uses hardware privacy computing technology, especially the Trusted Execution Environment (TEE), to ensure the confidentiality and integrity of data during the computing process. All calculations running in TEE ensure the correctness of program execution and data security. At the same time, Phala Network also incorporates zero-knowledge proof technology, which can realize private transactions so that the data of smart contracts does not have to be disclosed, thus protecting User privacy. In this way, Phala Network is able to provide universal and easy-to-use confidential smart contracts to support zero-trust cloud services.

The focus is on the Trusted Execution Environment (TEE) and the multi-layer key rotation system.

Trusted Execution Environment (TEE):A Trusted Execution Environment (TEE) is a privacy-focused technology that isolates a node’s code execution and operations from the computer processor’s main operating system. Phala Network uses Intel’s Software Guard Extensions (SGX) for TEE processing. This system ensures that even individuals with physical access, such as Phala Network operators or malicious third parties, cannot view the status of the nodes or manipulate the data processed by the nodes. Often, trusted execution environments are used in place of zero-knowledge proofs (ZKPs). In Phala Network, TEE is not only used to protect the confidentiality and integrity of data, but also supports complex computing tasks. For example, by transferring the execution of smart contracts to off-chain security workers powered by TEE, Phala Network is able to fully utilize the computing power of a single worker without worrying about the risk of data leakage or tampering.

Multi-layer key rotation system:Phala Network has a special group of worker nodes called “gatekeepers” whose main task is to ensure the security of computing. In Phala Network, Worker nodes do not have permission to exit the network at any time, so tasks need to be deployed to redundant Worker nodes with the same key to ensure that encrypted input, output, and program status can be accessed. Gatekeeper nodes are responsible for managing encryption keys and dynamically assigning secrets to worker nodes, ensuring data security and integrity. At the same time, Phala Network adopts a comprehensive key rotation mechanism to ensure the privacy of its contracts and the overall security of the system. This process periodically updates the master key according to election rules. In the new cycle, a new gatekeeper is elected and a new master key is generated through a secure key exchange protocol. The new key is used to re-encrypt the goalkeeper state, which is then distributed and confirmed among the new goalkeepers. During this transition period, communication between miners and gatekeepers will be temporarily interrupted to maintain consistency and security. Communication will resume only after key rotation is complete, with a delay of two on-chain confirmations. In addition to rotating master keys, Phala Network also regularly updates cluster keys and contract keys to enhance security.

After Phala Network added the artificial intelligence agency contract, Phala Network’s main business has changed from the original privacy computing infrastructure to the execution layer of Web3 artificial intelligence. Its original privacy computing infrastructure has become the current execution layer of Web3 artificial intelligence. Basically, the existing artificial intelligence agent contract operates on the basis of the original privacy calculation.

Artificial Intelligence Agency Contract:The AI ​​Agent Contract is a tool for users to build smart contracts that allow users to control AI agents that can run on or off the blockchain. These AI agents are coded in TypeScript/JavaScript and executed in Phala Network’s nodes, and they have the ability to aggregate data from external APIs and plug into custom AI agents. The artificial intelligence agent leverages Phala Network’s trustless cloud computing infrastructure for off-chain confidentiality procedures. On a technical level, the AI ​​agent contract runs on TEEs within worker nodes, which are managed by Phala Network. A subset of gatekeeper worker nodes are responsible for managing keys to ensure network security. These contracts provide developers with powerful tools to leverage artificial intelligence agents to perform a variety of tasks and are guaranteed to handle sensitive data due to the security and privacy of the environment in which they operate. The development of artificial intelligence agency contracts provides new possibilities for the application of blockchain technology, and also brings more room for innovation to developers.

pRuntime：By hosting the artificial intelligence agent contract in the node’s TEE, the working node protects the code integrity and privacy of the workload and ensures that the contract is not subject to malicious attacks and tampering during execution, while providing powerful computing power for the network. . pRuntime (Phala Runtime) is a program running inside the TEE. It is the core component of the Phala Network’s working node. pRuntime is responsible for receiving and running computing tasks from the blockchain to ensure the tamper-proof and security of the calculation.

To sum up, by introducing artificial intelligence agent contracts, Phala Network makes artificial intelligence agents run as simply as smart contracts, and makes artificial intelligence agents managed by smart contracts, allowing artificial intelligence agents to accept management and access control at any time. . And various artificial intelligence agents can freely call each other to form complex applications, and all operations are run in TEE, ensuring the integrity and privacy of the code. This allows Phala Network to fully utilize the advantages of the previous privacy computing infrastructure when transforming into the execution layer of Web3 artificial intelligence.

Innovation compared with projects on the same track

After Phala Network introduced the artificial intelligence co-processor, although the underlying operating logic of the Phala Network project has not changed, the main narrative of the project has changed from the original privacy computing infrastructure track to a subdivided track in the AI ​​track. : In the algorithm track, its competitors have also become Morpheus, QnA3.AI, Fetch.AI, SingularityNet, ChainGPT and other projects.

Advantages of Phala Network (picture from Phala Network white paper)

1. Super privacy: Thanks to the fact that Phala Network was engaged in the privacy computing business before transforming into the AI ​​​​track, it has innate privacy advantages. Phala Network uses hardware privacy computing technology, especially TEE, to ensure the confidentiality and integrity of data during the computing process. All calculations in TEE ensure the correctness of program execution and data security. At the same time, it also combines zero-knowledge proof technology to realize private transactions and protect smart contract data from being disclosed, thus maintaining user privacy. In this way, Phala Network provides users with safe and reliable privacy computing solutions and provides solid technical support for privacy protection in blockchain applications.
2. Powerful features: Phala Network has introduced artificial intelligence agent contracts, giving it an absolute advantage in the artificial intelligence agent track. Artificial intelligence agents contracts make artificial intelligence agents run as simply as smart contracts, and make artificial intelligence agents managed by smart contracts. , allowing artificial intelligence agents to accept management and access control at any time, and at the same time enabling EVM smart contracts to expand their functions outside the chain, greatly improving the deployment and work efficiency of artificial intelligence agents. This achieves an artificial intelligence agent centered on smart contracts.
3. Rich integrations: Through the use of artificial intelligence agent contracts, Phala Network enables it to aggregate data from external APIs and insert custom AI Agents. It also eliminates the learning curve of accessing decentralized applications. This allows easy and large-scale access to excellent external AI tools and projects, such as: OpenAI, LangChain, GPT4, etc.
4. Wider applications: Because of the use of artificial intelligence agent contracts, the deployment of Phala Network’s artificial intelligence agents is very simple and the deployment areas are also very wide. Users can link their own artificial intelligence agents to Social, Defi or Dao according to their own preferences and needs. To improve the efficiency of your work and life.

To sum up, Phala Network has made the deployment and use of artificial intelligence agents a very simple matter with its superior performance. It not only solves the simplicity of operation and deployment for users, but also solves some users’ concerns about privacy. Concerns have given Phala Network a very large advantage in the competition on the AI ​​track.

Project model

Business model

Since the basic narrative of Phala Network has now shifted from privacy computing to the AI ​​track, the economic model of Phala Network consists of four roles: Miners, staking users, users using artificial intelligence agent contracts, users using privacy computing.

miner: The miners of Phala Network are providers of network computing power. Miners provide the computer CPU computing power they can control to Phala Network. They are the only people responsible for actual calculations in Phala Network. They are the basis for the existence of the entire network. The computing power they provide The more power you have, the richer rewards you can get.

Staking users: Because the mining design of Phala Network requires miners to provide more than just computing power, in order to prevent mining from going bad, a certain number of PHA tokens must be pledged as a deposit. However, some miners are only willing to provide computing power and do not want to provide additional token pledges. Therefore, staking users can be divided into two types: Stakepool and Vault. Among them, Stakepool only belongs to miners who provide computing power and are willing to provide pledged tokens or directly provide their PHA tokens to Stakepool. If currency holders do not want to pledge their tokens directly to a Stakepool, they can also send their tokens to a Stakepool. The coins are placed in Vault, which is pledged by Vault based on the performance of each Stakepool.

Users who use artificial intelligence agent contracts: Users who use artificial intelligence agent contracts are now the main users of the Phala Network project. When they run artificial intelligence agent contracts on Phala Network, they need to pay fees to Phala Network. The fees are paid in the form of PHA tokens, which are also Phala One of Network’s main sources of income.

Users of privacy computing: With its excellent privacy computing business, Phala Network will still attract some users with real needs for privacy computing. The fees paid by users using private computing are also one of the main sources of income for miners and Phala Network.

From the above analysis, we can see that Phala Network’s revenue is:

• Usage fees paid by users using artificial intelligence agent contracts
• User fees paid by users of privacy computing

Token model

According to the white paper: the total amount of $PHA is 1 billion, and the current circulation is 738.3 million, with a circulation rate of 73.83%. The overall token distribution is as follows:

Token empowerment:

According to the white paper, the uses of PHA in Phala Network are as follows:

1. Miner Rewards: Miners in the network receive PHA tokens for mining new blocks, requiring their computing power.
2. Usage fees: Users who use artificial intelligence proxy contracts and use privacy computing pay fees in PHA tokens.
3. Participate in governance: After staking PHA tokens, you can participate in the governance of PhalaDAO.
4. Participation in staking: The mining design of Phala Network requires miners to not only provide computing power, but also to pledge a certain number of PHA tokens as a deposit to prevent mining from going bad.

Value Determination of PHA

According to the white paper, in Phala In the Network network, the empowerment of PHA tokens is quite satisfactory, and a staking mechanism has been added, which in disguise increases the key point of locking PHA tokens to increase the value of the project. Now the circulation rate of PHA tokens has reached 73.83%. From the perspective of circulation rate, the circulation rate of PHA is slightly higher now, but it is still at the stage where nearly one-third of the tokens have not been developed, so in the future development process of the project, the value of PHA tokens will be more The increase is from Phala The value of the Network project itself is the degree to which market users use artificial intelligence agent contracts. If the usage is high, then as the token price rises, the miners’ income will become higher and higher, leading to more miners. Entering mining locks more tokens to reduce the circulation rate of tokens, thereby forming a positive flywheel. However, there is no centralized or regular destruction scenario for PHA tokens, which makes its token economics insufficient. If Phala Network’s business does not develop well, its token price will not rise accordingly.

Token price performance

PHA price trend (data source: https://www.coingecko.com/en/coins/phala-network）

According to Coingecko statistics, the price of PHA has increased more than 4 times in the past year since July 2023 (the lowest point is $0.0816, the highest point is $0.3465). The main trading venues are first-tier exchanges such as Binance, OKX, HTX, etc. This year’s The increase is mainly due to the dividends from the development of the AI ​​​​track.

Market value

The current currency price of PHA is US$0.126, the current circulation is 738,334,838 pieces, and the market value is: US$93.43 million.

FDV

The current currency price of PHA is US$0.126, the total circulation is 1 billion, and the market value is: US$126.54 million.

Average daily trading volume:

PHA’s average daily trading volume is around US$5.66 million.

PHA daily trading volume (data source: https://www.coingecko.com/en/coins/phala-network/historical_data）

PHA’s daily trading volume is US$5.66 million, and its circulating market value is approximately US$126.54 million. Although the turnover rate is only 4.47%, which is a low level, it is more due to the current sluggish market environment, which has recently been affected by the Mtgox incident. , causing most users to be in a wait-and-see state.

Top ten token holding addresses

Top ten addresses held by PHA (data source: https://etherscan.io/token/0x6c5bA91642F10282b576d91922Ae6448C9d52f4E#balances）

As can be seen from the figure, the top ten addresses held by PHA account for a total of 92.19%.

Number of PHA currency holders

Number of PHA currency holders (data source: https://phala.subscan.io/tools/charts?type=holder）

It can be seen that although the number of people holding PHA tokens has been steadily increasing in the past year.

Daily active account

Daily active account status (data source: https://phala.subscan.io/tools/charts?type=account)

It can be seen that Phala Network’s daily active accounts have continued to increase over the past year.

Pledge situation

Actual circulating amount of PHA (data source: https://phala.subscan.io/）

From the figure, we can see that the circulating amount of PHA should be 847.27 million pieces, but the actual circulating amount is 738.33 million pieces. It can be concluded that the actual pledge amount should be 108.94 million pieces.

Project risk

Although Phala Network changed the track of its project to the AI ​​track after launching the artificial intelligence agency contract, which is conducive to the development of its project and the growth of currency prices, the technical challenges it faces have also increased exponentially, especially for Phala Network uses a relatively complex artificial intelligence agent contract based on privacy computing, and will continue to access new AI Agents in the future, which also poses a lot of challenges to its technical team, so there will definitely be some system problems in the future. , this depends on the problem-solving ability of the Phala Network team and the Whether Network’s artificial intelligence agency contract is as powerful as advertised, there are currently many uncertainties.

Summary

By introducing artificial intelligence agent contracts, Phala Network makes artificial intelligence agents as easy to run as smart contracts and managed by smart contracts, achieving real-time management and access control of artificial intelligence agents. Different artificial intelligence agents can freely call each other to build complex applications, and all operations are performed in a Trusted Execution Environment (TEE) to ensure the integrity and privacy of the code. This allows Phala Network to take full advantage of the previous privacy computing infrastructure and transform it into a Web3 artificial intelligence execution layer. This innovation provides a new solution for the security and privacy protection of artificial intelligence applications, and provides strong support for the integration of smart contracts and blockchain technology.

After launching the artificial intelligence agency contract, Phala Network positioned the project on the AI ​​track, which is beneficial to both project development and currency price growth. However, the technical challenges that come with it also multiply. In particular, Phala Network uses complex artificial intelligence agent contracts based on privacy calculations, and will continue to access new AI Agents in the future, posing huge challenges to the technical team.

All in all, the Phala Network project itself has gained certain advantages in the privacy computing infrastructure track through the introduction of the Trusted Execution Environment (TEE). And now with the support of the artificial intelligence agency contract, if it can proceed as scheduled according to the project plan, Phala Network will also have certain advantages in the AI ​​track.

References

Understanding Phala Network: A Comprehensive Overview

Phala Network White Paper

DePin project in Polkadot ecosystem, can Phala Network break out of the siege?

Disclaimer:

1. This article is reproduced from [X]. The copyright belongs to the original author [@FrontierLab_ZH]. If you have any objection to the reprint, please contact Gate Learn Team, the team will handle it as soon as possible according to relevant procedures.
2. Disclaimer: The views and opinions expressed in this article represent only the author’s personal views and do not constitute any investment advice.
3. The Gate Learn team translated the article into other languages. Copying, distributing, or plagiarizing the translated articles is prohibited unless mentioned.