’s Revolutionary GPU Cloud

Greythorn Asset Management
11 min readMar 12, 2024



Opening Remarks

In the digital age, computing power is fundamental to technological progress, defining a computer’s ability to perform operations through its memory, processor speed, and number of processors. This power determines a device’s efficiency in executing tasks and its ability to handle multiple processes simultaneously, influencing both performance and cost. As artificial intelligence and deep learning models have surged in popularity, so has the demand for increased computing power, particularly in GPUs, leading to a global shortage.

The central processing unit (CPU), often regarded as the computer’s brain, alongside the graphics processing unit (GPU), plays a crucial role in determining a device’s computing power. A more potent CPU can process more operations per second, while a GPU enhances the CPU’s ability by handling parallel tasks more efficiently.


In response to this growing need, has emerged with an innovative solution: building the world’s largest AI compute cloud by leveraging underutilised resources from data centres, crypto farms, and consumer GPUs. This initiative promises to deliver almost unlimited computing power at a fraction of traditional costs, addressing the critical gap between the burgeoning demand for computational resources and the limited supply.’s approach not only aims to alleviate the current shortage but also supports the ongoing evolution of AI technologies, ensuring the continuous advancement of digital capabilities.

Project Overview aims to revolutionise cloud computing with its decentralised GPU network, specifically designed for machine learning. By building the largest AI compute cloud, offers scalable, accessible, and affordable computing power, challenging traditional cloud services’ limitations. Through partnerships with entities like Filecoin and Render, and the upcoming IO token, enables AI/ML engineers to easily scale their projects across a vast network of GPUs.

The platform supports essential AI development tasks like batch inference, parallel training, and reinforcement learning, optimised for Python workloads. With the introduction of the IO token, is advancing towards a fully operational, cost-effective decentralised network. Leveraging DePIN technology, it caters to a wide range of GPU-dependent applications, offering a flexible, user-controlled computing environment. seems to be setting new standards in decentralised cloud computing, leading innovation in the field. Features

IO Cloud

IO Cloud is designed to simplify the deployment and management of decentralised GPU clusters, offering ML engineers and developers scalable and flexible access to GPU resources without the need for heavy hardware investments. This platform delivers an experience akin to traditional cloud services but with the benefits of a decentralised network.


  • Scalability and Affordability: Aims to be the most cost-effective GPU cloud, reducing AI/ML project costs by up to 90%.
  • Integration with IO SDK: Enhances AI project performance through seamless integration, creating a unified, high-performance environment.
  • Global Reach: Distributes GPU resources worldwide to optimise ML serving and inference, similar to a CDN.
  • RAY Framework Support: Uses the RAY distributed computing framework for scalable Python application development.
  • Exclusive Features: Provides private access to OpenAI ChatGPT Plugins for easy deployment of training clusters.
  • Crypto Mining Innovation: Seeks to revolutionise crypto mining by supporting the ML and AI ecosystem.

IO Worker

IO Worker is engineered to streamline and optimise supply operations for users within a WebApp. This includes user account management, real-time activity monitoring, temperature and power consumption tracking, installation support, wallet management, security, and profitability analysis.

It bridges the gap between the demand for AI processing power and the supply of unused computing resources, facilitating a more cost-effective and streamlined AI learning process.


  • Worker Homepage: Offers a dashboard for real-time monitoring of connected devices, with functionalities like deleting and renaming devices.
  • Device Details Page: Displays comprehensive device analytics, including traffic, connectivity status, and job history.
  • Earnings & Rewards Page: Tracks earnings and job history, with transaction details accessible on SOLSCAN.
  • Add New Device Page: Simplifies the device connection process, supporting quick and easy integration.

IO Explorer

Designed as a comprehensive platform to provide users with deep insights into the operations of the network, similar to how blockchain explorers offer transparency for blockchain transactions. Its main goal is to enable users to monitor, analyse, and understand the details of the GPU Cloud, ensuring complete visibility into network activities, statistics, and transactions while maintaining privacy for sensitive information.


  • Explorer Homepage: Offers insights into the supply, verified suppliers, active hardware quantity, and live market pricing.
  • Clusters Page: Displays public information on clusters deployed across the network, along with real-time metrics and booking details.
  • Devices Page: Shows public details of devices connected to the network, providing real-time data and transaction tracking.
  • Real-time Cluster Monitoring: Offers instant insights into cluster status, health, and performance, ensuring users have up-to-date information. Architecture and Technology

IO-SDK: The Core Technology

IO-SDK, a fork of Ray, underpins’s functionalities by allowing tasks to run in parallel and handle different languages. It works well with key ML frameworks, making flexible and efficient for various computing needs. This setup, along with a well-defined set of technologies, makes sure Portal can handle today’s requirements and adapt to future changes.

That being said, uses a multi-Layered Architecture, including:

  • User Interface: Acts as the visual front-end for users, including the public website, customer area, and GPU provider area. It’s designed to be intuitive and user-friendly.
  • Security Layer: Ensures system integrity and safety, incorporating network protection, user authentication, and activity logging.
  • API Layer: The communication hub for the website, providers, and internal management, facilitating data exchange and operations.
  • Backend Layer: The core of the system, handling operations like cluster/GPU management, customer interactions, and autoscaling.
  • Database Layer: Stores and manages data, with main storage for structured data and caching for temporary data.
  • Task Layer: Manages asynchronous communications and tasks, ensuring efficient execution and data flow.
  • Infrastructure Layer: The foundation, housing the GPU pool, orchestration tools, and execution/ML tasks, with robust monitoring solutions.

IO Tunnels

Reverse tunnels technology creates secure connections from a client to a remote server, allowing engineers to bypass firewalls and NATs for remote access without complex configurations.

Workflow: An IO Worker connects to an intermediate server ( server). Then, the server listens for connections from the IO Worker and the engineer’s machine, facilitating data exchange through the reverse tunnel.


Application in

  • Engineers connect to IO Workers via the server, simplifying remote access and management without network configuration challenges.


  • Ease of Access: Direct access to IO Workers, eliminating network barriers.
  • Security: Ensures protected communication, maintaining data privacy.
  • Scalability and Flexibility: Efficiently manage multiple IO Workers across different environments.

IO Network

The IO Network employs a mesh VPN architecture to provide ultra-low latency communication between antMiner nodes.

Mesh VPN Networks:

  • Decentralised Connectivity: Unlike traditional hub-and-spoke models, mesh VPNs connect nodes directly to each other, facilitating multiple data paths for enhanced redundancy, fault tolerance, and load distribution.
  • Advantages: Robust against node failures, scalable, reduced latency, and better traffic distribution.

Benefits for

  • Direct connections lower latency, optimising application performance.
  • No single point of failure, the network remains operational despite individual node failures.
  • Enhances user privacy by making data tracking and analysis more challenging.
  • Easy addition of new nodes without impacting performance.
  • Efficient resource sharing and processing across nodes.


$IO Coin is set to launch in Q2 2024 as the native cryptocurrency and protocol token of, serving AI Startups, Developers, and GPU Suppliers in the IO Ecosystem with various functionalities aimed at streamlining payments, rewarding contributions, and facilitating efficient compute job transactions.

For AI Startups and Developers:

  • Cluster Payments: IO Coin simplifies payment processes for cluster deployments and compute jobs.
  • Model Inference Gateway: Requires a microtransaction in IO Coin for every model inference on
  • IOSD Credits: A $USD-pegged utility token derived from burning IO, used for transaction fees on the Network.

For Suppliers:

  • GPU Rewards: Rewards for GPU contributions, including payments for hired time and incentives for idle time utilised for inferencing hosted models on the network. also employs a dual token system with IO and the USD-pegged IOSD, ensuring stable transaction costs. IOSD is generated by burning IO coins, a process overseen by net emissions to maintain supply equilibrium.

Tokenomics and Governance

  • Burn and Mint Economics: $IOSD production through $IO burning maintains a static supply of $IO month-on-month, adjusting based on network usage.
  • Net Emissions: Allows the protocol to sustain $IO rewards perpetually, adjusting the minting based on IO burned for $IOSD.
  • Governance: $IO Coin acts as a governance token, with stakers influencing network decisions, including fee structures and feature additions.

Objective & Design Principles:

  • connects global GPUs for AI and ML workloads, with native assets incentivising network growth and balanced development.
  • Distribute rewards to GPU providers and offer loyalty points to AI Engineers for continuous network use.
  • Allows stakers to vote on key network decisions, promoting decentralised governance.
  • Allows IO Coin holders to stake towards node operators or users, earning a share of their rewards.
  • The Network initiates with a 5% fee, used for IO Coin burns or dollar incentives, supporting network growth and user engagement.


In the Web3 sector, there are already many compute marketplaces, with prominent ones including Akash, Render, Nosana,, Inferx, Aioz, Fluence Labs and GPU.NET. Despite this, has quickly gained significant attention, already valued at $1 billion after it raised $30 million in a Series A funding round from investors including Hack VC, Solana Labs, and cryptocurrency exchange OKX. What sets apart in the DePIN space is its superior clustering technology. This innovative feature allows to combine multiple computers into a single, powerful supercomputer. Currently, stands out as the only service capable of forming decentralised clusters on a large scale, providing a crucial edge in the development of complex AI solutions.

Source: Suyo on X

Also, in 2023, received prestigious recognition by winning an award at the Inventions Asia Awards, where it stood out among over 110 inventions from across Asia. Since the start of 2024, has shown impressive growth and engagement, evidenced by a waitlist exceeding 100,000 GPUs, delivering more than 14,000 compute hours, registering over 20,000 GPUs on its network, and successfully setting up over 650 clusters.

It’s also worth mentioning that while both and platforms like Akash aim to revolutionise traditional cloud computing with decentralised approaches, they cater to different niches within the sector. Akash Network provides a broad marketplace for cloud resources, using container technology, while is all about high-performance GPU computing for AI and machine learning, featuring unique clustering capabilities.

Furthermore, The Render Network Foundation has committed to allocate 300K RNDR, out of a total of 1.14M, to new GPU node operators, naming as its initial partner.

Bullish Fundamental Factors

  • has formed strong partnerships, notably with the Render Network and Solana, the latter chosen for its high throughput and low fees, making it ideal for DePIN projects as per Messari’s 2024 report. It has also partnered with Filecoin to offer new earning avenues for decentralised storage providers and enhance AI applications. Before this partnership, received hardware contributions from 1,500 Filecoin members.
  • The team includes Angela Yi, a Harvard alum who co-founded a consultancy aiding Filecoin projects, and Tory Green, COO with over 20 years of experience in venture capital and executive roles.
  • is set to become a significant player in the DePIN sector, highlighted as one of the most anticipated launches in Q1.
  • It’s the only project that aggregates GPUs from various locations into a unified distributed cluster, providing instant access to users.
  • With computing resources becoming as valuable as oil, aims to reduce costs for renting GPU/CPU power, crucial for AI and ML scalability. The demand for computing power has surged, yet the supply hasn’t kept pace, indicating a growing disparity.
  • The need for computing resources, especially for model training, is expected to become immense, potentially accounting for 1% of the US GDP by 2030.

Bearish Fundamental Factors

  • The market for decentralised computing is becoming increasingly crowded, with established players like Akash Network and emerging projects continuously entering the space. Stiff competition could impact’s ability to retain users.
  • The complexity of creating and managing decentralised clusters at scale involves significant technical challenges.
  • While there is growing interest in decentralised computing solutions, widespread adoption is still in its early stages. Slower-than-expected adoption by AI startups, developers, and GPU suppliers could limit growth.
  • Changes in laws for cryptocurrencies and decentralised networks can be risky. Also, using the $IOSD stablecoin might get regulators’ attention.
  • The dual-token system involving $IO and $IOSD might be complex for users to understand.

Closing Remarks is poised to capitalise on the lucrative AI computing market, boasting a highly skilled team and strong backing from reputable entities such as Multicoin Capital, Solana Ventures, OKX Ventures, Aptos Labs, Delphi Digital, among others. Considering the circumstances and its upcoming token launch, we regard it as a project deserving of close attention, but always remember to navigate this industry carefully.

You can connect with us on LinkedIn, where you can also check out our other research projects. Feel free to visit our website for more information about us, or follow us on our X profile.


This presentation has been prepared by Greythorn Asset Management Pty Ltd (ABN 96 621 995 659) (Greythorn). The information in this presentation should be regarded as general information only rather than investment advice and financial advice. It is not an advertisement nor is it a solicitation or an offer to buy or sell any financial instruments or to participate in any particular trading strategy. In preparing this document Greythorn did not take into account the investment objectives, financial circumstance or particular needs of any recipient who receives or reads it. Before making any investment decisions, recipients of this presentation should consider their own personal circumstances and seek professional advice from their accountant, lawyer or other professional adviser. This presentation contains statements, opinions, projections, forecasts and other material (forward looking statements), based on various assumptions. Greythorn is not obliged to update the information. Those assumptions may or may not prove to be correct. None of Greythorn, its officers, employees, agents, advisers or any other person named in this presentation makes any representation as to the accuracy or likelihood of fulfilment of any forward looking statements or any of the assumptions upon which they are based. Greythorn and its officers, employees, agents and advisers give no warranty, representation or guarantee as to the accuracy, completeness or reliability of the information contained in this presentation. None of Greythorn and its officers, employees, agents and advisers accept, to the extent permitted by law, responsibility for any loss, claim, damages, costs or expenses arising out of, or in connection with, the information contained in this presentation. This presentation is the property of Greythorn. By receiving this presentation, the recipient agrees to keep its content confidential and agrees not to copy, supply, disseminate or disclose any information in relation to its content without written consent.



Greythorn Asset Management

Melbourne-based asset management firm specialising in Technology, Web3, Cryptocurrency, and Blockchain. Subscribe for our latest industry insights.