Decentralization Cloud Computing: Opening a New Era of Computing Power

In recent years, with the rapid development of technology, companies like OpenAI and Nvidia have experienced significant increases in market value. The combination of artificial intelligence and cryptocurrency has become the main focus of the current market, with investors and businesses pouring in funds, demonstrating a strong consensus. In this broader context, decentralization, as a tool for the development of artificial intelligence, shows great appeal and potential. Although there is still a gap in practical applications compared to centralized models, leveraging the advantages of Web3 to expand the core areas of artificial intelligence has become a common goal among industry participants.

The four core areas of artificial intelligence include data, models, training, and inference. Among them, data is the most critical part and can be regarded as the raw material of artificial intelligence technology, while the other three aspects are different ways of processing data. In terms of data labeling and storage, Decentralization technology can play an important role.

If we compare data to raw materials, then Computing Power is the tool for processing these raw materials, used to maximize output efficiency. This article will analyze the ecological framework and economic model of Crypto x AI x DePIN around the theme of "Computing Power", helping readers understand the value and potential of decentralized Computing Power.

1. DePIN and Decentralized Computing Power Ecological Framework

Currently, high-quality Computing Power, as a necessity for artificial intelligence research and development, has been monopolized by traditional giants. This has made it difficult for startups and individual users to access reasonably priced Computing Power resources, with high prices becoming an obstacle that most buyers find hard to accept.

To address this issue, projects in the DePIN field often adopt a peer-to-peer economic model to provide high-quality resources to resource-demanding parties. This model allows each user to become a provider of physical facility resources while earning token rewards.

With the surge in demand for decentralized artificial intelligence Computing Power, a comprehensive and balanced framework for the supply ecosystem of decentralized AI Computing Power has been established to better meet customer needs. Among them, leading projects such as Io.net, Exabit, and PingPong play different important roles in the ecosystem, and the technical barriers of these projects as well as their layouts for the future development of decentralized Computing Power are impressive.

The decentralized AI Computing Power ecosystem consists of three main parts, which serve the roles of resource agents, resource providers, and channel distributors:

Resource Agent - Io.net

Io.net is a decentralized computing network that serves as a computing power agency, providing high-quality AI computing power to clients at low prices. On the supply side, it has globally distributed GPU resources. Its clients are mainly seed to Series B startups focusing on AI inference.

This DePIN project based on the Solana chain recently completed a $30 million Series A financing, participated by several well-known investment institutions.

As a top AI Computing Power resource agency, Io.net aims to aggregate 1 million GPUs to form a large DePIN Computing Power network, providing customers with lower-priced Computing Power. Users can contribute their idle GPU and CPU Computing Power to the io.net platform and receive token incentives. Its core goal is to provide high-quality AI Computing Power under decentralized price control, helping AI startups reduce costs.

The computing service IO Cloud provided by Io.net adopts a cluster-based modular construction, ensuring that all GPUs remain interconnected, enabling large-scale coordinated work during training and inference processes. This allows GPUs to concentrate Computing Power to access larger databases and compute more complex models, enabling AI startups to deploy computing hardware at a cost far lower than that of Decentralization.

Currently, the number of GPU clusters aggregated by io.net ranks first in the industry, with over 200,000 GPUs available online, among which the GeForce RTX 4090 and GeForce RTX 3090 Ti are the most numerous models.

Resource Provider - Exabit

Exabit, as the most promising AI Computing Power provider, is able to provide ample chips for deep machine learning. Its team has significant advantages in traditional AI Computing Power resources and was a first-level agent for NVIDIA. Relying on this technological resource barrier, Exabit can directly access hundreds of data centers and gain usage rights for high-end machines such as A/H100, RTX4090, and A6000.

Exabit provides large-scale machine learning Computing Power for Web3 giants. Compared to other cloud service providers, customers can significantly reduce costs while improving efficiency by using Exabit's services.

Exabit's goal is to provide customers with the fastest, highest quality, and most reliable Computing Power through unique Computing Power supply channels. High-quality Computing Power can not only save users costs but also provide customers with a full range of service options.

Currently, the quality of AI Computing Power provided by Exabit has been recognized by multiple AI Computing Power agents, and collaborations have been established with computing giants such as Renders Network and Io.net, aiming to contribute to machine learning through Decentralization.

Resource Channel Provider - PingPong

PingPong, as a DePIN resource channel provider, offers services through demand matching. It adopts a platform-based open protocol, providing underlying aggregated resources before offering services. PingPong's goal is to become a service aggregator for DePIN, which can be understood as the 1inch or Uber of the DePIN field.

PingPong obtains information about various networks and strategies, resource status, performance, stability, and other aspects through the control layer, provides an SDK, and then delivers the SDK to users through routing algorithms.

This approach aims to address the issues of limited resources and services in various DePIN networks, as well as the poor quality of services caused by the overly centralized allocation of global resources in certain regions. Through routing algorithms, PingPong obtains basic information about data, networks, and machines, aggregates it to generate strategies, and matches services based on customer requirements. The goal is to improve the quality and service of the DePIN application layer and to find the optimal price for Computing Power networks in situations of resource scarcity.

2. Analyzing the Decentralization Computing Power Ecosystem

Io.net and Exabit have reached a strategic partnership, with Exabit, as a supplier with a rich GPU machine library, committed to enhancing the speed and stability of the io.net network. Io.net allows customers to directly purchase and lease high-quality Computing Power provided by Exabit on its network in the capacity of an agent. Both parties believe that the success of the decentralized computing industry and the integration of Web3 and AI requires close collaboration among early industry leaders to achieve.

With the continuous growth of Computing Power demand, traditional cloud computing faces some problems:

1. Limited availability: Access to hardware through mainstream cloud services usually takes weeks, and commonly used GPU models are often unavailable.
2. Limited Choice: Users have restrictions on their choices regarding GPU hardware, location, security level, latency, etc.
3. High Costs: High-quality GPU prices are expensive, and the project's monthly expenses during training and inference may reach hundreds of thousands of dollars.

The vision of decentralized computing is to provide an open, accessible, and affordable alternative to address the core issues of centralized cloud service providers. However, to challenge the dominance of the major players in cloud computing, innovators must work together and support one another.

Asset Mode

1. Heavy Asset Model

As a supplier, Exabit has an absolute barrier backed by NVIDIA. The most valuable machines in machine learning computing power, such as the A100, RTX 4090, and H100, have a single unit price of about $300,000 and have become highly scarce resources, long monopolized by traditional AI giants. The resources that Exabit connects to on the supply side are extremely precious, playing a key and irreplaceable role in the decentralized computing power ecosystem.

The heavy asset model adopted by Exabit requires a large amount of fixed asset investment, and such a scale of capital and technological investment makes it difficult for startups to replicate. If Exabit can collaborate with more decentralized computing power agents to continuously expand the supply side and meet the industry's demand for computing power resources, it is expected to achieve industry monopoly and generate economies of scale in the B2B decentralized computing power sector.

However, the biggest risk lies in the inability to continuously provide resources for Computing Power agents after investing a large amount of capital. The value on the supply side is highly dependent on whether the Computing Power agents can continuously meet customer demand.

2. Light Asset Model

Io.net, as the most outstanding computing power agent at present, relies on a large decentralized computing network formed by globally distributed GPUs. From a business perspective, Io.net adopts a light asset operation model, establishing a strong brand in the AI computing power agency field through community operation and building a high level of consensus.

The core business of Io.net includes:

• Aggregate retail GPU Computing Power and reward tokens
• Acquire high-quality Computing Power from the supply side to sell to AI startups.

From the perspective of enterprises:

• Buy low and sell high quality Computing Power from the supply side to C-end customers.
• Help users earn tokens by sharing idle GPU Computing Power.
• Provide customers with Computing Power mining and staking platform

From the customer's perspective:

• The computing power price of the Io.net network is about 80% cheaper than other centralized cloud computing services.
• Earn rewards through staking and sharing
• After the client invests a certain capital, they can achieve compound interest.

As a typical light asset model company, Io.net's biggest advantage is its lower risk, as the team does not need to invest a large amount of machine costs upfront like the supply side. The reduced capital investment makes it easier for the company and investors to achieve higher profit margins. However, due to the low entry barriers in the industry and the ease of replicating the business model, long-term value investors need to consider this carefully.

3. Future Development

If the collaboration between Exabit and Io.net can help the decentralized Computing Power ecosystem go from 1 to 10, then joining PingPong may have the opportunity to reach 100.

PingPong's goal is to become the largest DePIN service aggregator, directly competing with Web2's Uber. As a channel provider, it connects clients to the resources with the best prices and quality by aggregating the real-time status of various resources. PingPong adopts a B2B2C light asset business model, connecting the supply side, resource agents, and end customers.

As a platform, if a channel vendor can develop into a platform that can issue assets, it will make the products more valuable. PingPong provides an SDK through a routing algorithm that can calculate resources to create AI Agents and convert new financial assets, while also helping clients with dynamic mining through the SDK, focusing on mining Computing Power that is useful for computing resources. This "asset-on-asset" model can greatly enhance the liquidity of resources and funds.

For PingPong, they hope to see more suppliers and agents entering the Decentralization Computing Power ecosystem, to highlight their own advantages, expand their business lines, and acquire more customers. Just like Baidu and Dianping are able to dominate the information field because more merchants and information are uploaded to the internet, creating a high demand for channel providers among customers.

IV. Outlook

Decentralization cloud computing is still continuously evolving. Although its ecological framework and model have become clear, and the leaders of various roles are fulfilling their respective responsibilities, it is still too early to shake the position of traditional cloud computing giants. Compared to traditional centralized cloud computing, decentralization can indeed conceptually address many issues faced by customers well, but the overall resources and scale of this market are still relatively small.

In the context of insufficient computing power resources to support AI development, the market needs another model to break through the dilemma. Currently, decentralized cloud computing can meet some of the needs of startup AI companies, but how it will develop in the future remains to be seen. Let us witness the evolution of this disruptive path together and participate in this technological revolution!