Web3 constructs an economic paradigm and cultural system that is entirely different from the offline world through blockchain technology. This paradigm shift showcases its infinite potential but also brings compatibility issues with Web2. For mass adoption, Web2 users often possess completely different production materials and assets. We envision that if we could bring the assets existing with current Web2 users back on-chain, it would not only inject more dynamism into the on-chain economic ecology but also enable users to obtain greater economic value. At the same time, it also brings richer off-chain asset categories to on-chain applications, promoting diverse development of the on-chain ecology.

In this regard, WiFi万能钥匙 (WiFi Master Key) is a typical case. Established in 2012, it coincided with China's transition from 3G to 4G, when internet data fees were high. At that time, WiFi万能钥匙 addressed the pain points of lacking public networks and expensive private networks through shared internet access. Its primary business model was to open users' home WiFi, creating a shared WiFi ecosystem. At its peak, WiFi万能钥匙 had 900 million users and 370 million monthly active users. It rode the wave of the trend but declined due to changes in the era. On one hand, China's data fees rapidly decreased, with many operators starting to offer low-cost home WiFi and unlimited mobile data, while public WiFi infrastructure gradually improved. On the other hand, its product relied on cracking others' WiFi passwords for sharing, leading to higher data fees and network lag, which gradually annoyed users and ultimately cost it market advantage.

Through this example, we can see that the WiFi万能钥匙 (WiFi Master Key) has built a reusable shared network through the existing massive WiFi infrastructure, achieving nearly 1 billion users with near-free network infrastructure. Web2 users possess some large existing assets, but due to the inherent incompatibility of Web3 asset attributes, these large assets cannot realize more value.

We also realize that many existing infrastructures in Web2 can further unleash their potential through Web3's global liquidity, shared and open economy, verifiable consensus blockchain, DeFi, and other characteristics. We refer to this model as ReFi (Repurpose Fi).

ReFi is fundamentally different from current DePIN projects. Depin devices often require purchasing hardware, but the scalability of these hardware is poor. To achieve true success, it is necessary to establish sales networks and supply chains in various places. However, for products that meet essential needs, merely relying on 'making money' is insufficient; the competitiveness of the product itself must be adequate to meet user demands. Many teams lack funding and the ability to build a complete ecosystem, often relying on KOLs and large sales agents to promote mining machines, ultimately devolving into a simple 'Ponzi scheme.'

In contrast, ReFi significantly lowers the entry barriers by repurposing the existing stock market of 8 billion users globally without the need for additional hardware production and sales.

We will list some significant gains that may arise from putting Web2 assets on-chain, providing the community with entrepreneurial insights. This model directly avoids the 'show-off' business model of current DePIN that relies on hardware production and sales, widely onboarding existing Web2 assets and aiding entrepreneurs' success.

Case Study

WiFi

When tokenomics combines with WiFi, we find that existing WiFi infrastructure can be opened up for direct use by third-party users, allowing users to earn income through renting and sharing WiFi. Projects can incentivize this through token distribution.

This model primarily targets users who wish to obtain free, fast outdoor WiFi and typically need to stay in one place for a longer time. Those providing free WiFi are households with existing WiFi infrastructure. As data costs have significantly decreased, providing free networks often fails to incentivize them anymore due to slower speeds and lag as costs. However, the introduction of tokenomics can significantly encourage these users to reopen network resources, achieving value exchange and profit sharing.

The global Wi-Fi market size was valued at $14.5 billion in 2023, projected to reach $39.4 billion by 2028, with a compound annual growth rate of 22.2%. In 2024, mobile data market revenue is expected to reach $0.6 trillion. The annual growth rate (CAGR) is projected to be 4.30% from 2024 to 2029, with the market size reaching $0.8 trillion by 2029. This data indicates that the WiFi sharing model combined with tokenomics is likely to occupy a place in this rapidly growing market.

Bandwidth

Grass is a successful example of ReFi. Its goal is to directly contribute unused internet bandwidth and receive corresponding rewards. Currently, large companies utilize residential proxy networks to access data, leveraging local bandwidth from millions of users, so servers recognize this as access from real users, avoiding DDoS attacks on single nodes of large companies. Residential proxies are widely used in fields such as web scraping, data analysis, market research, social media management, and electronic ticketing.

In the past, companies with demand for residential proxies may have purchased services from residential proxy providers, who obtained residential IPs through three main channels: enticing users to install plugins, purchasing from ISPs, and IP leasing companies, without compensating users who are the important source of proxies. Grass aims to allow users to contribute unused internet bandwidth, becoming a proxy for large companies requesting internet resources, helping them access the web and making this market more open and transparent.

For the consumer group, the market is currently relatively small, with global residential agency service sales of about $620 million in 2023, projected to grow to $840 million by 2030, with a compound annual growth rate (CAGR) of 4.6%.

Data

The data contribution of Web2 users also represents a breakthrough potential track, especially in the current context where LLMs and AI Agents face challenges with computing power and quality data. Scale AI stands out in this field, with the slogan 'Empower AI with your data.'

Scale AI Business Model, source: Scale AI

Scale AI's primary business model connects user data, task platforms link cost-effective Southeast Asian data workers, who then label the data uploaded by users and input it into large models while scoring the results of these models. Data is extremely valuable for training LLMs or Specifically Models, and some have likened Scale AI to the shovel sellers of the gold rush. The average hourly wage for data labelers is $1-2, and this low-cost labor model enables Scale AI to achieve efficient large-scale operations.

For Web3 projects, we can incentivize workers through tokens, significantly increasing their hourly wages. We note that Vana, invested by Paradigm, is similar to this Data DAO model. Although it has a well-established business model, it may face certain challenges in the attention-investment environment of Web3, as its user reach is relatively small and it connects with Data DAO, leading to lower user awareness of the product, which may result in a lack of buying interest for the token.

A rather positive example is Aggregate, which directly engages retailers by uploading ChatGPT conversation content to Aggregate for further training of other models. Aggregate has already received investment support from Binance Labs. We believe that through the distribution of tokens, data providers and data labelers can be effectively incentivized, with most of the profits flowing to business participants rather than just the platform.

Currently, the data labeling market is projected to be $838.2 million in 2024, with further projections to reach $10.3462 billion by 2033, a compound annual growth rate of 32.2%.

Energy

Daylight is another successful example, with the slogan 'repurpose your surplus energy becoming a new revenue.' The company has secured two rounds of funding from A16Z and Lattice Fund Lead, totaling $13.2 million. Daylight chooses to manage and generate the grid at the network edge, incentivizing people to install various electronic devices in their homes, such as solar panels and water heaters, through tokenization, allowing for control of these devices via mobile phones while simultaneously earning token rewards. By adopting token incentives, Daylight promotes the widespread use of clean energy and balances the load on centralized grid transmission.

This model is closer to DePin, but if we can improve existing clean energy devices without purchasing electronics, achieving improvement goals through tokenization would be easier. Essentially, Daylight sells clean energy devices, while the core of ReFi is to reuse the broadly existing asset base. Collaborating with existing manufacturers to build a system or any specification interface standard, such as allowing manufacturers of those clean energy devices to directly receive token incentives on-chain, could also be a feasible business model. We look forward to teams innovating in this area.

According to reports, the global power generation equipment market was valued at $110.4 billion in 2022, projected to reach $173.1 billion by 2032, with a compound annual growth rate of 4.8% from 2023 to 2032.

Summary

In this article, we find that there is an inherent incompatibility between Web3 and Web2 users, especially on the Web2 asset side. Therefore, we propose a new direction for ReFi (Repurpose Finance), utilizing existing vast Web2 infrastructures and user assets, guiding these assets onto the blockchain through blockchain technology and token economics, thus redeveloping their potential.

The core of ReFi lies in tapping into the vast array of existing resources widely used around the world, including home GPUs, WIFI networks, repurposed bandwidth, data, energy, etc., which can be upgraded based on a large existing scale. The concept of ReFi has been around for a long time, and its biggest difference from Depin is that there is no need to repurchase hardware; it only requires leveraging the large existing stock market with token economics to find new directions for repurposing.

However, it is also necessary to consider real demand and efficiency. Some repurposing directions may not be ideal, such as using bandwidth for residential proxy networks like Grass, whose market size is small, but the demand itself does not have high efficiency requirements. In contrast, contributing bandwidth for data transmission to help AI or contributing GPUs for LLM training seems unrealistic, as these require very high bandwidth and GPU resources, and achieving efficiency in these scenarios is often difficult under current technological conditions.

We look forward to seeing more application scenarios of ReFi, which will help entrepreneurs directly utilize the existing assets of the eight billion Web2 users worldwide, integrating Web2 assets into the Web3 world through the advantages of blockchain's global liquidity, distributed state storage, and token economics, thus opening a new era of value creation and resource utilization.