Original David Foresight Research 2024-04-30 11:01 UK
Author: David Zhang@Foresight Ventures
With the long-term rapid development of world science and technology, the market value of giant companies such as OpenAI and NVIDIA has increased several times in the past two years. Crypto x AI has become the core narrative of this cycle. The high market sentiment and continuous capital investment prove that a strong consensus has been formed. In the context of AI as a goal, decentralization as a powerful tool for the development of AI does have great appeal and imagination. Although there is still a huge gap between the actual business implementation and the centralized model, it has become a common goal of web3 participants to expand the four core aspects of AI with the advantages of web3 and to exert greater potential through continuous optimization.
data
Model
train
reasoning
At present, decentralization can be supported by technology in the four aspects mentioned above. First of all, data must be the core. Models, training and reasoning are all ways to process data, so it can be said that data is the raw material of AI technology, and the others are processing methods. Whether it is data labeling or data storage, decentralization has a great role and value here.
If data is the raw material, then computing power is the tool for processing the raw material to maximize the efficiency of output. Next, let’s go straight to the topic of this article. This article will analyze the ecological framework and economic model of Crypto x AI x DePIN around “computing power”.
In this article, I will mainly explain the ecological framework and market situation of "Crypto x AI x DePIN" to help readers understand the value and potential of decentralized computing power⬇️
1. DePIN & decentralized computing power ecological framework
Pain point: High-quality computing power is a must for AI research and development. This scarce resource has been monopolized by traditional giants, making it difficult for startups and individual users to buy computing power with reasonable cost performance. This high price is unacceptable to most buyers.
Decentralized solution: Currently, most projects in the DePIN track adopt a P2P economic model to provide high-quality resources to resource demanders, allowing every user to serve as a physical facility resource provider and obtain token rewards at the same time.
With the surge in demand for decentralized AI computing power, the development of the decentralized AI computing power supply ecosystem has formed a balanced and comprehensive framework to better meet customer needs. Among the leading projects, Io.net, Exabit and PingPong play different important roles in the ecosystem. The technical barriers of these three projects and the future development pattern of decentralized computing power are quite shocking.
The decentralized AI computing power ecosystem is mainly composed of three parts, which act as resource agents, resource providers and channel providers in the ecosystem respectively:
Resource Agents - Io.net
Io.net is a decentralized computing network that provides high-quality AI computing power to customers at a low price as a computing power agent. On the supply side, it has GPUs distributed around the world. The client is currently in the seed round to the B round, focusing on startups in AI reasoning.
Recently, the DePIN project based on the Solana chain completed a US$30 million Series A financing round, led by Hack VC, with participation from Multicoin Capital, Foresight Ventures, Solana Labs and others.
As the top AI computing power resource agent, Io.net is aggregating 1,000,000 GPUs to form a huge DePIN computing power network, with the goal of providing customers with computing power at a lower price. Users can manually contribute their idle GPU & CPU computing power to the io.net platform to get $IO token incentives. The core goal is to provide high-quality AI computing power while controlling prices through decentralization, so as to help AI startups reduce costs.
IO Cloud is a computing service provided by Io.net. IO Cloud uses cluster building blocks to keep all GPUs connected to each other, which enables GPUs to coordinate work on a large scale during training and reasoning. When GPUs work in coordination, they can concentrate computing power to access larger databases and calculate more complex models. AI startups can use io.net's products to complete computing hardware deployment at one-tenth of the centralized price while obtaining what they need. What is even more striking is that io.net focuses on aggregating computing power for machine learning. Io.net can help DePIN giants such as Render Network and FileCoin format GPU supply for machine learning and achieve the most fundamental and direct resource support for the underlying technology.
Currently, the number of GPU clusters assembled by io.net is the largest in the industry. There are more than 200,000 GPUs available online on io.net, of which the GeForce RTX 4090 has the largest number of available GPUs, with nearly 50,000, followed by the GeForce RTX 3090 Ti with more than 30,000.
Resource Provider - Exabit
As the most promising AI computing power provider, Exabits, as an AI computing power service node, can provide sufficient chips for deep machine learning. The Exabits team can also be called a standout in traditional AI computing power resources. The team used to be a first-tier agent of AI giant NVIDIA. Relying on such a technical resource barrier, Exabit can directly access hundreds of computer rooms on the resource supply side, and has access to A/H100, RTX4090 and A6000 machines.
Exabits provides large-scale machine learning computing power for web3 computing giants on the client side. Compared with Nebula Block, which requires customers to spend more than $140,000 per month to obtain cloud services, after migrating to Exabits, customers' monthly cloud service fees are around $40,000, which reduces expenses by more than 70% and increases efficiency by 30%.
Exabits' main purpose is to provide customers with the fastest, best quality and most reliable computing power through a unique computing power supply channel. High-quality computing power can save users' costs while providing customers with a full range of service options.
The quality of AI computing power provided by Exabits has been recognized by many AI computing power agents. It has now reached cooperation with computing power giants such as Renders Network and Io.net, aiming to contribute to machine learning through decentralization.
Resource channel provider (Uber) - PingPong
PingPong, as a DePIN resource channel provider, provides services by matching requirements. PingPong adopts a platform-based open protocol, providing underlying aggregated resources before providing services. PingPong's goal is to become a service aggregator for DePIN, which can be understood as DePIN's 1inch, or an aggregated Uber.
How to provide services: PingPong obtains SDK from various networks and strategies, resource conditions, performance, stability, etc. through the control layer, and then provides the SDK to users through the routing algorithm.
Pain point: The resources and services in each DePIN network are limited. The global search for resource allocation results in poor service quality due to excessive regional concentration.
Solution: Routing algorithm - obtains data, basic network information and machine information, etc., generates strategies after aggregation, and provides services through customer requirements matching. The purpose is to improve the quality and service of DePIN's application layer, and to find the best-priced computing network when resources are insufficient.
2. Analysis of the decentralized computing power ecosystem
Io.net and Exabits have reached a strategic cooperation. As a supplier with a rich library of GPU machines, Exabits is committed to improving the speed and stability of the io.net network. Io.net will provide the highest quality computing power provided by Exabits as an agent to allow customers to purchase and rent directly on the io.net network. Io.net and Exabits agree that the success of the decentralized computing industry and the combination of web3 and AI can only be achieved through close cooperation between early industry leaders. With the growing demand for computing power, some of the problems currently faced by traditional cloud computing:
Limited availability: Using cloud services like AWS, GCP, and Azure often takes weeks to gain access to hardware, and the most commonly used GPU models are often not available.
Limited choice: Users have limited choices in GPU hardware, location, security level, latency, etc.
High cost: Good GPUs are expensive, and the monthly training and inference costs of a project can easily reach hundreds of thousands of dollars.
The vision of decentralized computing is to provide an open, accessible and affordable alternative that can solve the core problems of centralized cloud service providers, including limited availability, limited hardware selection and high training and reasoning costs. Judging from the current situation, challenging the status of major giants in cloud computing still requires innovators to work together and support each other to take a revolutionary step.
Asset Model
Heavy asset model
As a supply side, Exabits has an absolute barrier backed by NVIDIA. The only machines with valuable machine learning computing power are A100, RTX4090 and H100, and the price of each of these three machines is about 300,000 US dollars. At the same time, these machines have become highly scarce resources and have been monopolized by traditional AI giants for a long time. In this case, the resources that Exabits can connect to on the supply side are extremely valuable. Since the quality of retail investors sharing their own personal GPU idle computing power is not enough to support the calculation and processing of large-scale AI models, the role played by Exabits in the decentralized computing power ecosystem is crucial and not easily replaced.
Exabits's heavy asset model requires a large amount of fixed asset investment. This amount of capital investment and technology investment makes it difficult for startups to copy and imitate. Therefore, if Exabits can cooperate with more decentralized computing power agents, and provide the computing power resources needed by the industry while the supply side continues to expand, it will be easy to achieve industry monopoly and scale effect in the field of B2B decentralized computing power.
However, the biggest risk is that after investing a large amount of capital, it is impossible to continuously provide resources to the hashing agents, so whether the supply side can make large-scale profits depends extremely on whether the hashing agents can have continuous customers. No matter who the hashing agents are, as long as there are customers and demand, the value of Exabits as a supply side will increase with the growth of demand.
Light asset model
As the most outstanding computing power agent at present, Io.net relies on the GPUs distributed around the world on the supply side to form a huge decentralized computing network. From a business perspective, io.net adopts a light asset operation model and builds a strong brand in the AI computing power agent through community operation and building a high degree of consensus.
Io.net's core business:
Aggregate retail GPU computing power and reward tokens
Acquire high-quality computing power from the supply side and sell it to AI startups
Enterprise perspective:
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
Providing customers with a computing power mining and staking platform, but an initial investment of about $4,000 is required to have a relatively good return. Based on this, Exabits also offers fragmented H100 machines for leasing to improve liquidity.
Customer perspective:
The computing power price of Io.net network is about 80% cheaper than other centralized cloud computing services.
Stake to earn & Share to earn。
After customers invest a certain amount of capital, they can enjoy compound interest.
As a typical light asset model company, the biggest advantage is that the risk is relatively low. The team does not need to invest a lot of machine costs to get started like the supply side. Due to the small amount of capital investment, it is easier for the company and investors to obtain a higher profit margin. At the same time, because the industry has a low entry threshold, the business model is easy to be copied and replicated, which is a point that long-term value investors need to consider carefully.
3. From 10 to 100?
If the cooperation between Exabit and Io.net can help the decentralized computing power ecosystem go from 1 to 10, then bringing PingPong along may have a chance to reach 100.
PingPong's goal is to become the largest DePIN service aggregator, directly targeting Uber in web2. As a channel provider, it aggregates the real-time status of various resources and connects customers to the resources with the best price and quality. PingPong adopts a B2B2C light-asset business model. The first B-end is the supply end, and the second B-end is the resource agent. The C-end provides customers with the best resource options through information.
As a platform, if the channel business can develop into a platform that can issue assets as much as possible, the product will be more valuable. PingPong can use the SDK provided by the routing algorithm to create its own AI Agent with computing resources. While converting new financial assets, the SDK can dynamically help customers using the application to conduct dynamic mining, focusing on mining computing power that is useful for computing resources. This model is understood as Assets on assets, which can greatly enhance the liquidity of resources and funds.
For PingPong, they hope to see more suppliers and agents enter the decentralized computing power ecosystem, so that they can better highlight their advantages, expand longer business lines and have more customers. It is very simple to understand that the reason why Baidu and Dianping can dominate the information field is that more businesses and information are uploaded to the Internet, which makes customers have a high demand for channel merchants.
4. The future is promising
Decentralized cloud computing is still developing step by step. Although the ecological framework and model of decentralized cloud computing have become very clear, and the leaders of various roles are also fulfilling their responsibilities in the ecosystem, it is still too early to shake the position of traditional cloud computing giants. When compared with traditional centralized cloud computing, decentralization can indeed solve many problems of customers in concept, but the overall resources and volume of this market are still very small in comparison. In the case that the computing power resources supporting AI are far from enough, the market needs another clear stream, or a model to solve the dilemma. The decentralized cloud computing we can see now can indeed meet some of the needs of start-up AI companies. What will happen next? Let us witness this road of subversion together and follow the evolution of the revolution together as participants!
