EMC: Driving new momentum for the integration of AI technology and Blockchain

1. Introduction

In the rapid development of the technological wave, the integration of artificial intelligence and blockchain technology is becoming a new focus, and the narrative of building Computing Power DePin based on the Graphics Processing Unit (GPU) is beginning to set off a new wave in the Web3 field.

In recent years, the widespread application of AI technology has led to an increasing demand for Computing Power resources in many industries. However, the monopoly of high-performance GPUs in the market makes it difficult for many small and medium-sized enterprises to obtain the necessary Computing Power support. The birth of the EMC (Edge Matrix Computing) project is based on this demand trend, intending to solve the problem of insufficient allocation of Computing Power friction by integrating idle graphics card resources from around the world.

The EMC team pioneered the concept of "DeAI", which is different from traditional GPU cloud services. Through its Computing Power scheduling platform, the project provides an efficient AI training model, aiming to enable developers to obtain computing resources at low cost. This innovation promotes the integration of AI and blockchain in resource utilization, data sharing, and empowers the development of the Web3 ecosystem, creating real-world application value.

2. Overview of the EMC Project

2.1 Project Overview

EMC (Edge Matrix Computing) was established in 2022 and is a high-performance Decentralization AI Computing Power application network aimed at addressing the frictional shortage between AI technology development and GPU Computing Power resources. As of October 2024, Computing Power networks and AI+Web3 communities have been established in more than 30 countries and regions worldwide, dedicated to providing a more equal and open opportunity for entrepreneurs and developers.

As the first platform in the Web3 field to seamlessly connect GPU Computing Power assets with AI applications, EMC's core products are designed for different AI and Web3 application scenarios, and build distributed high-performance Computing Power DePIN services. For example, EMC Hub is responsible for Decentralization Computing Power scheduling, providing Computing Power resources worldwide to help AI developers efficiently complete tasks; JarvisBot focuses on rich AI service applications, optimizing user experience through Depth learning, and providing intelligent support for various business scenarios; OmniMuse, as an innovative platform, aims to promote the research and development and promotion of AI technology.

In this context, EMC is committed to promoting the construction of a decentralized AI ecosystem, providing not only low-cost and efficient computing resources for developers, but also opening up new possibilities for innovative applications in various industries. By integrating distributed Computing Power, Smart Contracts, and AI services, EMC hopes to become a significant driving force for the integration of AI and Blockchain in the future, creating broader development space for global developers and entrepreneurs.

2.2 Founding Team

The main founder of the EMC project (Image source: Edge Matrix Chain (emc.network))

The EMC core team includes a number of industry veterans with extensive backgrounds in cloud computing, AI, and marketing:

• Alex Goh

MC's co-founder and Chairman of EMC Foundation, holds an MBA from Macquarie University, with over 20 years of experience in global market development. Previously served as General Manager of Greater China at Improbable.io and Global GM at AWS (Amazon), and is currently mainly responsible for the commercialization and global promotion of EMC in Singapore.

• BK Khur

Co-founder and CTO of EMC, graduated from Nanyang Technological University (NTU) College of Engineering, and was a researcher at NTU. With a rich technical background, he was responsible for digital transformation at Deloitte Consulting and is the co-founder of companies such as JuzToday and ShopperBoard, with management and technical experience in multiple innovative projects.

• Allen Foo

EMC Foundation board member and product and technical advisor, founder and CEO of UCCVR early-stage venture capital fund and VooX, responsible for business expansion in the Greater China region for Unity and Microsoft, with significant leadership experience in cloud services.

• Terrence Ngu

EMC Foundation board member, global market promotion strategy consultant, founder of Hashmeta, former chief community officer of StarNgage, and held key positions in many high-tech companies, focusing on global market strategy and community building.

2.3 Current Financing Status

As of now, the EMC project has completed multiple rounds of important financing, demonstrating its strong development potential in the global AI and Web3 fields. The first round of financing was completed in January 2024, with major investors including Swiss Bochsler Group, Future3 Campus, 1783 Labs, Frontier Research, DMC, VOFO Corp, Exabits.ai, Hashm ta, CEEX Labs, and other institutions and organizations.

In February 2024, the EMC team announced the completion of the second round of strategic financing worth millions of dollars, led by Faculty Group and Flow Capital. The funding sources cover the global Web3 community, DAOs, and AI developer communities, once again accelerating the layout, development, and construction of EMC Computing Power Nodes.

On August 30, 2024, EMC announced the successful completion of a $20 million Series A financing round, led by Amber Group and P2 Ventures, with other participants including One Comma, Kapley Judge and Associated Corporations, Cyberrock Venture Fund and other well-known investment institutions, further strengthening its industry position as a DecentralizationComputing Power scheduling platform and AI innovator.

2.4 Related Resources

3. Project Advantages

3.1 Distributed GPU

In the background of the high-performance GPU market being dominated by giants such as Nvidia, EMC effectively solves the problem of Computing Power supply and demand imbalance with its unique distributed DecentralizationComputing Power network, utilizing idle GPU resources worldwide. Especially after the ETH merge, the closure of numerous Mining Farms has led to a large number of idle GPU devices, which provides EMC with an opportunity to provide the required Computing Power support for AI developers at a low cost.

Currently, over 100 GPUNodes have been deployed on the EMC network, covering multiple countries and regions globally. The main models include A100, H100, RTX 4090, and 3090. These Computing Power resources are provided jointly by Internet Data Centers (IDC), Cloud Service Providers (CSP), Mining Farms, and EMC AI Workstations specifically designed for AI development. The EMC network adopts a mechanism that combines Proof of Work (PoW) and Proof of Stake (PoS), where participants obtain Token rewards through contributing Computing Power and stake, achieving dual benefits of Mining and stake.

In terms of user experience, the EMC AI Workstation is designed with a focus on plug-and-play. The first batch of products is equipped with high-performance hardware, such as Intel Core i7 CPU, 2TB solid-state drive, 32GB DDR5 6400Hz memory, and RTX 4090 graphics card, which ensures the computing resources and data processing capabilities required for complex AI tasks and provides users with an efficient working environment, promoting innovation and development of the entire ecosystem.

Global GPU Computing Power distribution diagram (Image source: Geomap）

3.2 Innovative DeAI Ecosystem

EMC has built a complete system from infrastructure to application development through its unique DecentralizationAI ecosystem (DeAI), with the core principles of openness, transparency, and democratization. It aims to solve the centralization problem of traditional AI through the model, data, and Computing Power of Decentralization. For example, in traditional AI models, they are often controlled by a few companies, resulting in the closed nature of data and algorithms. In EMC's DeAI system, algorithms and data are shared through a distributed network, allowing users to independently manage data, greatly enriching the data ecosystem, and enhancing users' contribution and control over AI models.

At the onset of the Bull Market, there is a particularly urgent demand for new technologies and innovative models in the market, and the combination of AI and Web3 is an important trend for the future market. By integrating these two hot fields, EMC has created a new market narrative, bringing new investment opportunities to investors, especially in the decentralized AI development and application market, which is expected to usher in a new wave of investment frenzy.

3.3 Dual-Token + Dual-Deflation

EMC adopts a 'Dual Token + Dual Deflation' model: one Token is used for governance and participating in ecosystem decisions, while the other Token serves as the primary Medium of Exchange. This design enhances the project's flexibility and allows Tokens to play their respective roles in different functions.

In addition, EMC's dual deflation mechanism reduces the Circulating Supply of Token through specific economic design to ensure its scarcity. Specifically, this includes regular repurchasing of Tokens to reduce the market's Circulating Supply, and further dropping the Circulating Supply through the destruction of some Tokens (such as fees collected in transactions). This mechanism not only maintains the scarcity of Token, but also enhances its long-term value.

In the EMC community, users can actively participate in the EMC ecosystem by depositing tokens, participating in real-world asset (RWA) transactions, and selling AI models, thereby promoting the circulation and use of tokens. In short, this "dual token + double deflation" model has laid a solid economic foundation for EMC, and with a diversified revenue model, it has incentivized more developers and users to participate in the EMC ecosystem.

EMC Economic Model Design (Image Source: Edge Matrix Chain (emc.network) )

3.4 Low Technical Threshold

By launching the EMC Hub tool, EMC has greatly lowered the technical threshold for AI DApp development. Developers can easily create and deploy AI applications using its rich SDK and toolset. This open and user-friendly development platform not only attracts more developers to join the EMC ecosystem, but also promotes the popularization of AI technology in the Web3 ecosystem, paving the way for the rapid application of intelligent AI technology.

4. Introduction to EMC Ecology

As a project that combines artificial intelligence and Web3 technology, the entire ecosystem of EMC is divided into four layers, namely Protocol Layer, Network Layer, Application Layer, and Asset Layer. Technically, it provides users with efficient AI Computing Power solutions through unique network topology, edge computing power scheduling, and multi-level Node design. EMC ecosystem composition (image source: Edge Matrix Chain (emc.network) )

4.1 EMC Protocol

EMC Protocol is a distributed AI Computing Power scheduling scheme based on the EVM ecosystem, which utilizes the high-performance infrastructure of Arbitrum Onemainchain to achieve the submission and verification of state machines, and is committed to scheduling idle Computing Power resources globally to meet the high computational demands of AI training tasks.

As shown in the figure below, the network topology of EMC can be divided into four types of Node: Computing Nodes, RelayNode Router Nodes, Validator Nodes, and Transaction Storage—these Nodes responsible for different functions are uniformly scheduled to transmit and execute transaction submission and confirmation, and collaborate to complete AI model training and inference tasks; ultimately, all transaction statuses and computation results will be stored in the transaction storage layer of Arbitrum One, ensuring the long-term availability of data.

EMC network topology (Source: https://cryptoviet.info/edge-matrix-chain-la-gi)

The core technical implementation of the EMC Protocol relies on its efficient commitment and confirmation mechanism, which gives it a unique advantage in Computing Power scheduling and validating Node management. Firstly, the commitment mechanism records the state machine as a commitment structure sent to the Arbitrum mainchain for recording, called 'commitment.' At this stage, users can immediately proceed to the next operation without waiting for actual confirmation to be completed. Once the transaction is submitted to the Smart Contract, it is considered completed, and the entire process is asynchronous, greatly reducing the perceived latency in user experience, although it takes some time.

In the PoS mechanism, the validating Node fulfills its responsibilities by staking EMC Tokens to ensure fairness and reliability. Once validation fails, the staked assets may be confiscated, further strengthening the security of the system. The incentive mechanism is linked to the amount of staked EMC, with Nodes staking the most EMC having priority to become validators. Smart routing also depends on stake to ensure the prioritized allocation and stability of tasks. There are two options for computing Nodes: staking EMC for higher rewards, or choosing to perform tasks that do not require long Computing Power, increasing operational flexibility and profit, especially suitable for smaller Nodes.

Meanwhile, EMC Protocol has significantly improved the utilization efficiency of Computing Power through the scheduling of edge Computing Power. Compared to traditional centralized data centers, the EMC network calls idle GPU resources worldwide to optimize the allocation of computing power. Through collaboration with the EMC Partner Network (EPN), EMC achieves global Decentralization Computing Power support, ensuring the stability and scalability of the system under large-scale concurrency. This design enables EMC Protocol to effectively address the challenges of today's complex computing environment and provide a solid foundation for AI and real-time applications.

4.2 EMC Hub

By integrating the AI model library with computing resources, the EMC HUB platform makes development and deployment more efficient. For developers, they can package AI models into Docker containers and upload them to the platform, along with sample code and parameter instructions, in order to receive platform rewards. This mechanism greatly reduces the burden on developers in terms of model publishing and distribution. Therefore, for users, they only need to subscribe to compute nodes to run these model Docker containers in a one-click deployment manner, quickly launching complete AI instances, and the system will automatically configure the corresponding APIs.

EMC Hub official website introduction (image source:EMCHub）

In the scheduling of Computing Power, EMC Hub relies on the collaboration of intelligent routing and Nodes: the former optimizes paths and data transmission, while the latter performs computing tasks, dynamically scheduling GPU resources in the computing pool, and intelligently allocating them based on task load and priority. Compared to traditional methods, this approach avoids the tedious process of purchasing cloud services, selecting models, and deploying environments, allowing developers to focus more on innovation.

In terms of security and efficiency, EMC Hub uses a PoS+PoW mixed Consensus Algorithm, with a total of 3F+1 validation Nodes in the entire network to jointly maintain the mechanism, and completed the verification process through the Byzantine fault tolerance (IBFT) Algorithm in a 2/3 confirmation manner. PoS ensures the security of the Nodes and prevents malicious attacks, while PoW is responsible for verifying the completion of computational tasks. This mixed mechanism not only enhances the security of the platform, but also shortens the AI training cycle. According to statistics, the cost of this method is only 30% of the traditional method, and the workload can be reduced to within a few hours.

EMC Multilevel Framework and Application Distribution Diagram (Source: EMCprotocol (EMC) · GitHub)

4.2.1 JarvisBot

The EMC AI assistant Jarvis is a revolutionary AI development platform, based on the architecture of the EMC network and Decentralization, combined with Depth learning Algorithm, making it not just an AI chatbot, but also able to improve the accuracy of Computing Power allocation through Depth learning Algorithm. While retaining powerful conversational abilities, it automates complex computation tasks and model training, optimizing the AI deployment process.

In terms of functionality, JarvisBot implements a variety of AI applications, including generating copy, images, translations, and article rewriting. At the same time, users can create customized bots for customer support, lead generation, order updates, and personalized recommendations, among other scenarios. By combining the Web3 economic model, users can not only enjoy AI services but also earn profits through contributing resources. This setup sets JarvisBot apart from traditional AI applications that can only be used through user subscription, truly achieving the sharing of AI development and creation. This model makes it highly followed in the market.

In addition, JarvisBot greatly simplifies the deployment process of AI models from a design perspective. Developers can easily call its functions and quickly launch AI models without tedious manual configuration through the Web3 tools provided by JarvisBot. This not only improves the efficiency of model training but also brings more efficient and economical solutions to the AI and DeAI development fields, making it a Decentralization version of "ChatGPT".

Jarvis Bot page (Image source: https://docs.jarvisbot.ai/)

4.2.2 OmniMuse

OmniMuse is an innovative platform designed to promote the development of AI technology through Decentralization Artificial Intelligence (DeAI). The platform offers a range of features, including customizable Smart Contract templates and frameworks designed for model minting, transactions, and data sharing, significantly accelerating the development process of AI applications. At the same time, OmniMuse integrates popular blockchain development tools, simplifying the creation of Decentralization applications.

OmniMuse uses Decentralization storage solutions such as FIL to ensure the permanence and tamper resistance of data assets, while promoting secure data sharing and transactions, all while prioritizing data privacy. The platform's advanced privacy and security features benefit from state-of-the-art encryption tools such as homomorphic encryption, secure longer computation, and verifiable computation, further enhancing the platform's security.

In addition, the DeAI Store being developed by OmniMuse will become an innovative platform for aggregating Decentralization AI applications, helping users discover and access the latest smart technology applications. DeAI Store provides Decentralization AI data storage, Smart Contract templates, and development frameworks, while integrating encryption tools to ensure user privacy and security. The platform aims to create a collaborative environment without technological "boundaries," where everyone can share, to unleash the enormous potential of AI and attract numerous AI developers, creators, and users to join together to promote the innovation and development of AI technology.

OmniMuse main page (Image source:OmniMuse）

4.3 Openverse Framework

Building on EMC Hub, Openverse further expands functionality, integrates multiple developer tools and SDKs, enhances developers' application capabilities in a Decentralization environment, and promotes seamless integration with EMC Hub, enabling developers to quickly deploy AI applications.

In terms of functionality, Openverse is a platform that integrates various SDK tools for Web3 developers, including EMC SDK, Web3 SDK, 3D Scene SDK, and DID SDK. These tools can support core Web3 functions. For example, the 3D Scene SDK can quickly build virtual 3D worlds, while the DID SDK provides blockchain-level identity verification to ensure data privacy and security.

Developers can not only upload AI models to the platform, but also easily start and manage AI instances through the one-click deployment feature of Openverse, simplifying the development process. This integrated platform significantly lowers the threshold for Web3 development, allowing developers to focus on application innovation and business rise.

Main steps for using Openverse (Image source: EMCprotocol (EMC) · GitHub)

5. Overview of Related Tokens

5.1 Token Allocation

$EMC is a Token based on the issuance of Arbitrum One public chain, with a total supply of 1 billion. The distribution of these Tokens covers multiple purposes, including community rewards, development funds, Liquidity, etc. Its design concept is to provide opportunities for developers and users to participate in the Decentralization computing ecosystem, motivate users to actively participate in the construction of the ecosystem, and achieve efficient utilization of Computing Power and economic circulation.

EMCToken initial distribution model (Image source: Token Distribution | EMCWhite Paper)

5.2 Income Deflation Model

EMC has introduced a dual-coin economic model, which includes the basic $EMC coin and the Stable CoinCredits as the medium of exchange in the EMC market. The core of this mechanism is that users need to use $EMC to purchase Credits, and in this process, $EMC will be completely destroyed, thereby increasing its scarcity and value. This design not only helps maintain the price stability of $EMC, but also attracts more users to join the EMC ecosystem.

EMC's profit deflation model includes specific profit deflation and Computing Power consumption deflation, which is used to maintain the balance of Token supply and demand.

1. In the income deflation mechanism, transactions use stable coins or other Tokens, and then the Computing Power cluster rents through the Computing Power protocol. The transaction fees of the RWA platform will be converted into $EMC and destroyed, where 80% of the rental income goes to the lessor, and 20% is destroyed. At the same time, the transaction parties will also receive $EMC as a reward, further promoting the use and destruction of Tokens.
2. Under the Computing Power consumption deflation mechanism, EMC Hub provides abundant Computing Power and application subscription services. Users need to stake other cryptocurrencies to obtain these services. In this case, the staked assets will be converted into $EMC and deposited into the consumption account. The $EMC in the consumption account cannot be withdrawn but will be destroyed after the service is used, thereby increasing the consumption of $EMC and enhancing its value.

5.3 Release Model

EMCToken's initial Token Generation Event (TGE) launched on November 9, 2023. The entire Token release plan will last for 24 to 48 months, covering investors and project teams. In Token distribution, ecosystem rewards (including governance tokens) account for 47% of the total supply. In addition, the EMC economic system has built-in deflation mechanisms and destruction plans, aiming to maintain the ecosystem through these measures and enhance the long-term value of the Token.

$EMCToken 36-month circulation change trend chart (source: Token Acquisition | EMC White Paper)

6. Future Value Analysis

The EMC project is a combination of traditional Web2 and Web3. Compared to Web2 projects, its advantage lies in the use of distributed GPUNodes, which can effectively aggregate dispersed Computing Power resources and alleviate the supply-demand imbalance caused by traditional centralized systems; compared to other Web3 projects, EMC provides a cost-effective solution for AI model training by combining AI with DePINDepth, establishing a market for knowledge, data, and computing asset sharing. In addition, the unique Credits mechanism can accelerate economic circulation and provide investors with new sources of income and opportunities.

In terms of future applications, EMC makes high-performance computing accessible and affordable, and also opens the door for AI applications in more industries.

For example, in the medical field, EMC can use its powerful Computing Power to support the processing of large-scale medical data, promote personalized medical care and precise diagnosis, and more effectively formulate treatment plans through the analysis of patients' historical data and genetic information; in the financial industry, EMC's Computing Power can be used to process complex financial transactions and risk assessments, reducing costs while ensuring the security and transparency of data.

The most promising applications are focused on smart cities and internet of things IoT. EMC's distributed architecture can support real-time data processing of a large number of devices, promoting the optimization of intelligent transportation, energy management and other systems, thereby improving the operational efficiency and sustainability of cities.

Currently, the engineering technology of large models is relatively mature, but the stability of Computing Power and the reliability of code encapsulation still need to be closely followed and continuously optimized. Considering that the EMC project is in the hot track of DEP, it preliminarily has the feasibility of CX (Customer Experience). On the other hand, from the disclosed information, the project's Chinese background is obvious, and future market expansion may require a diversified strategy to enhance global influence.

1. https://docs.emc.network/whitepaper
2. https://docs.jarvisbot.ai/
3. https://twitter.com/EMCprotocol
4. https://linktr.ee/emcnetwork
5. https://emchub.ai/#/home
6. https://coinmarketcap.com/currencies/edge-matrix-computing/