Interpreting io.net: DePIN Nova on Solana, reconstructing the GPU resource utilization model

By Revelo Intel

Compiler: Deep Tide TechFlow

DePIN quickly gained popularity as one of several key narratives, especially on Solana. With projects like Render and Helium making significant progress, DePIN and related projects of the genre are well worth looking into.

In the current digital era, blockchain technology is increasingly becoming a key factor in driving innovation and efficiency. In this article, we’ll take a deep dive into how io.net, a revolutionary project on the Solana blockchain, is reshaping our understanding and use of GPU computing by consolidating idle GPU resources.

Project Background

• io.net is a DePin protocol built on top of Solana to help solve coordination problems between GPUs.
• Similar to how DeFi disrupts traditional finance by eliminating middlemen and pooling liquidity, io.net aims to do the same with computing resources. It provides a coordinated solution where all participants are motivated to work together.
• io.net identified inefficiencies in the market, such as supply constraints and centralized control of GPUs. By addressing these issues, io.net aim to offer products that are cheaper than centralized services.

What is DePIN?

• DePin refers to the work of replacing existing infrastructure through hardware investment and token incentives.
• Examples include 5G networks, routing 5G through people, and replacing legacy telecom infrastructure.
• Hardware participation in the network brings value to people.

Why is DePIN growing on Solana?

• Solana’s friendly and supportive community fosters growth and collaboration, with Solana providing tools for microtransactions, data storage, and other features. All in all, the combination of community support and helpful tools makes Salana an ideal platform.
• DePin represents the purpose of the blockchain.
• The GPU space is monopolized by industry oligopolies, highlighting the need for disruption.
• The DePin project has the potential to bring significant change to various industries.

More and more crypto applications are focused on merging off-chain tokens with real-world economic output, particularly Solana-based crypto applications.

• Render is creating a peer-to-peer network that enables GPU rendering tasks to be completed using an idle GPU. Rendering is actually one of the sources of io.net aggregate GPUs.
• Hivemapper aims to develop the world’s newest AI-powered map, rewarding drivers who use dashcams with $HONEY tokens.
• Helium incentivizes users to provide local hotspot coverage in exchange for native $HNT tokens.
• Shdw Drive is a decentralized storage solution that involves native $SHDW, similar to FIL, Arweave, etc.

The team behind io.net

• Ahmad Shadid is the founder and CEO of io.net. Ahmad said that before creating the io.net, he developed a machine learning model to detect institutions that were trying to fill orders in the market and hide them. He expanded the model to cover 2,000 stocks and 150 cryptocurrencies.
• Tory Green is io.net’s Chief Operating Officer and Chief Financial Officer, having joined full-time three months ago. He began his career in corporate and has held positions in banking, private equity and corporate strategy. He co-founded a digital publishing company in the web2 era, where he transitioned from investor to operator, helping portfolio companies with sales, marketing, finance, operations, and strategy. He spent nearly two years as CEO and CFO at Hum Capital, a machine learning company. During this time, he developed an in-depth understanding of the field of AI. He has been interested in cryptocurrencies since 2016.
• Tory says he recognizes io.net is a deep solution to a real problem. He was impressed with Ahmad’s intelligence, vision and drive. Believing in its potential, he joined io.net full-time three months ago.

io.net comparison with traditional competitors

• io.net price is a significant advantage over competitors, but there are other differences.
• io.net provides an option for users to recreate a data center similar to AWS (Amazon Web Services) with comparable performance. Users can move data centers closer to them to reduce latency and provide a competitive advantage for trading scenarios.

• Tory says the user experience in io.net is better than deploying on AWS, which often involves complex processes and certifications that take months.
• Ahmad said that io.net offers GPUs that are faster than the 40,000 or 80,000 GPUs of Lambda Cloud (Amazon’s cloud computing).
• Startups are moving from Lambda Cloud to io.net due to faster internet speeds and better network performance.
• Small details like internet speed are important – a provider with fewer servers can offer higher speeds due to less network congestion.
• The goal of io.net is to leverage the profitability of AI computing, similar to AWS. By attracting more GPU vendors, io.net can provide vendors with consistent profits and expand their customer base.
• Customers don’t have to move everything away from AWS, but can benefit from lower prices, comparable performance, and an improved user experience, and the growth of the network could eventually lead to the establishment of cloud-sourced data centers and other resources, Ahmad said.

What is the trade-off between competitors and io.net?

• Traditional competitors may counter that io.net lacks specific features such as automated deployment, CI/CD, and API defense that they have spent years developing.
• Ahmad notes that io.net has built these capabilities within the ML (machine learning) industry for training and inference purposes. He added that while there may be some shortcomings in the range of features currently available, the overall experience provided by io.net is superior to that of traditional vendors.
• With their platform, users can easily access a wide range of GPU offerings in 45 countries, Ahmad said. Just click deploy, and the deployment is complete in 10 seconds.
• Whereas traditional methods of using AWS or other providers often involve begging for GPU access by submitting presentations and LinkedIn profiles. Ahmad shared how, as a developer, he was hurt by being asked to provide promotional materials and a LinkedIn profile when all he wanted was to use the GPU.
• Whereas io.net aims to return power to engineers by streamlining processes and eliminating unnecessary requirements. io.net provides a supportive community on Discord where engineers can seek help from others they deploy with.
• Ahmad added that by engaging with and providing support to the community, they aim to create a better overall user experience for engineers. This approach is in line with the needs and mindset of the current generation of developers, who value simplicity and accessibility.
• He added that there is an emerging market driven by startups that are willing to invest large sums of money to acquire computing resources. This generation is building innovative solutions that are critical to serving their specific needs and ways of working. io.net platform aims to change the dynamics of the market through on-demand supply.

What is a GPU cluster?

io.net technology allows 50,000 to 70,000 GPUs to be clustered into a single cluster. This clustering feature differentiates io.net from competitors who typically offer separate GPU instances. io.net build a network infrastructure that enables efficient communication between cluster GPUs. If users have enough resources, it is possible to build large language models through the platform.

• This product is designed to cater to those building on specific infrastructure, targeting customers who have financial resources but are struggling with costs. io.net team recognizes the importance of alignment with customers. The product has been demonstrated to well-known companies such as OpenAI, Instacart, and Uber.
• Ahmad expressed confidence in the scale and revenue potential for specific customers. He expects a lot of money to pour into io.net network as more startups join. Transactions can be made using traditional payment methods such as Visa, and users don’t need to know anything about crypto. The goal is to attract customers who pre-order GPUs for a long period of time (e.g., six months) in order to maximize revenue.
• In addition, Ahmad foresees a network where the flow of funds exceeds the token issuance. Tory sees competitors as suppliers or aspiring suppliers rather than direct competitors. He claims that they have a unique advantage in serving enterprise AI customers. All transactions within the io.net are subject to network fees ranging from 5% to 10%.

In conclusion, io.net is not only an example of technological innovation, but also an example of how blockchain can be applied to real-world problems. As their team continues to push the boundaries, we look forward to seeing more projects like io.net emerge that bring more possibilities to our Web3 world.