# Introduction

GPUNET is engineered to revolutionize the landscape of decentralized computing power, offering a robust and scalable solution for a wide range of computational needs.&#x20;

At the heart of GPUNET is the vision to democratize access to high-performance computing resources, making them readily available for diverse applications ranging from data analysis and scientific research to AI development and beyond.

Our platform is uniquely designed to cater to the growing demand for accessible and efficient GPU computing power. By harnessing the potential of blockchain technology, GAN Chain enables users to seamlessly rent out their idle GPU resources or access additional GPU power as needed, fostering a dynamic and flexible computing ecosystem.

{% content-ref url="/pages/AQb5aCoLTJP32jcn2Ey8" %}
[Broken mention](broken://pages/AQb5aCoLTJP32jcn2Ey8)
{% endcontent-ref %}

{% content-ref url="/pages/e0k1RN9IPt9dammYpVcH" %}
[Broken mention](broken://pages/e0k1RN9IPt9dammYpVcH)
{% endcontent-ref %}

{% content-ref url="/pages/zF4wviBX3LsFjStaKXKp" %}
[Broken mention](broken://pages/zF4wviBX3LsFjStaKXKp)
{% endcontent-ref %}


# Dev Updates


# Q1

GPU Net Dev Updates 2025

### Week 1 (Jan 01 - Jan 07):

* Barrels Smart Contract was developed in the Solana Eco system.
* Test cases for the barrel smart contract were written, including integration and unit testing.
* sol.gpu.net design and UI testing.
* On the chain side, planning for Nova Upgrade.
* Dashboard-v2 bug fixes.

### Week 2 (Jan 08 - Jan 14):

* Front-end development of sol.gpu.net started.
* Integration of the smart contract in the website.
* Backend and event-listeners for barrel sale were developed.
* Introduction of Barrel Swap options, and their implementation in Smart Contract.
* AI Agent for twitter (GPU Agent), similar to perplexity agent but 1 month before them

### Week 3 (Jan 15 - Jan 22):

* Integration testing for sol.gpu.net, including front-end testing, back-end testing, and flow testing.
* Internal team testing for solana barrel sale
* Design and development of sol.gpu.net landing page
* Development started for GANChain Nova Upgrade.
* Started benchmarking GPUs for optimisation of inferencing ML models

### Week 4 (Jan 23 - Jan 31):

* Solana Barrel Sale Launch (sol.gpu.net), tier wise white list and fair launch
* Iteration over smart contract of barrel sale, to include more swap options.
* Ideation for minting and burning the NFTs and swapping options to compute, nodes, queen and king.
* Testing newer versions of Substrate and Polkadot for GAN Chain, stress testing with a huge number of transactions.
* ML Perf benchmarking on various GPU clusters.

### Week 5 (Feb 01 - Feb 07):

* Development and testing for minting and burning barrel NFTs
* Front-end development for Barrel Swap options, and integration.
* Production Deployment of NFT minting option and automatic minting on sol.gpu.net
* Set up GPU clusters with Multi-Instance GPUs (MIGs) and GPU optimisation with it
* Development of GAN Chain features on a newer version of Substrate and Polkadot.

### Week 6 (Feb 08 - Feb 14):

* Production Deployment of Swap for Compute option in Barrels.
* NFT Listing on Magic Eden for Barrels.
* DApp Auth and API migration from Authoriser to Custom Made Auth.
* Fixing and Improving of GPU Agent
* After testing, made a decision to continue with the older version of Polkadot, since no performance improvement and added complexity.

### Week 7 (Feb 15 - Feb 21):

* Providers move to DApp from Dashboard.
* DApp public view marketplace development and deployment.
* DApp API migration from Hasura GraphQL to custom REST APIs.
* Benchmarking LLMs for GPU Profiling
* Feature Analysis for Nova and Mainnet Upgrade over the same version of Substrate.

### Week 8 (Feb 22 - Feb 28):

* Bug Fixes on DApp, Barrel, GPUnet Website and Dashboard.
* Feature Planning for Subnets and Token Page.
* UI Design for Subnets
* Setting up Inferencing Engine and Optimising performance by 10%.
* Development started for Nova Mainnet on GAN Chain.

### Week 9 (Mar 01 - Mar 07):

* UI Design for Token (Quests Page) started
* Development of Smart contract, Front-end and Back-end of Subnets
* Planning for Node and GP buy back
* Quantization of Models and Improving Inferencing Speeds
* Quantized KV Cache, more offloading to CPU Memory.

### Week 10 (Mar 08 - Mar 14):

* Beta Launch for Subnets with 50 Invite Only users
* Development of Node Buy Back and GP Buy Back
* Features Planning for AI Agents for Subnet
* Testing GAN Chain dev env with Lesser Security and Greater Speed \[Block time 1 sec]

### Week 11 (Mar 15 - Mar 22):

* Node and GP Buy Back Launch for Batch 1
* Subnet Referral Codes Updated, Bug Fixes and improvements according to users feedback
* Development of AI Agent to enquire about subnets
* Quests UI Development started.

### Week 12 (Mar 23 - Mar 31):

* Integration of AI Agent in Subnets, and Deployment to Production
* Major improvements in the voting logic in Subnets following the user feedback.
* Development of AI Agent to create a subnet, it is supposed to chat with a user and create a subnet for him/her.
* Compilation of Results from Testing Faster BlockRates in GAN Chain dev. We decided not to compromise the security of the chain for speed.
* Development of hardcoded formulae for tokenomics and block rewards emission in GAN Chain for Nova Mainnet.&#x20;

<br>


# Q2

#### April ‘25 - May ‘25:

1. GPUnet’s new product Quest was launched. We facilitated over 1M diverse users to participate and earn fair rewards. Major marketing correlations were done with projects like Monad, U2U, Huddle01, Zerion etc.
2. In parallel to the success of Quests, Subnets page was also revamped with a better UI and UX. New features like timeline updates, display videos, etc were added for subnet owners.
3. On the DApp, new compute providers were added, we increased the availability and uptime of the GPUs.
4. The tech-team was gearing up for the upcoming TGE and Nova Mainnet upgrade. The core of the chain was made more solid and decentralised.
5. GAN Chain moved to Mainnet, with a smooth upgrade without the need of restarting validator nodes. In the Nova upgrade, the block rewards were increased for the subnets and queens. Overall daily block rewards were increased to 2880 $GPU.

<br>

#### June ‘25 - mid July ‘25:

1. Season 1 of GPU Airdrop ended, with over 100K guaranteed $GPU airdropped to questers, and over 3M to node runners.
2. A Roadshow plan was launched for $GPU to be on all the top chains, we started with launching a token on Ethereum, touched $3 on all time high.
3. GAN Chain to Ethereum bridge was developed and deployed with a decentralized and verified protocol.
4. Next up $GPU was launched on Base. Base to GAN Chain bridge was developed and deployed under a secure protocol.
5. EVM to SS58 and vice versa, transfer of tokens was developed and deployed for the node runners to move funds to and fro.
6. For complete transparency, the history of rewards was now made accessible for the subnet users in their profile page.
7. The updates from the whole GPUnet multiverse were displayed in the subnets page in real time. Users could now know what is happening with the token, who is booking machines on dapp, who is creating a subnet, etc on their live feed.

<br>


# Q3

#### 15 July ‘25 - 15 Sept ‘25

1. Integration of AI Models on Dapp. User can now use different models’ APIs on [dapp.gpu.net](http://dapp.gpu.net)&#x20;
2. Chat, Image Generation and Video Generation on DApp. Generate up to 30 seconds of high quality videos or hd images on [dapp.gpu.net](http://dapp.gpu.net)
3. On Subnets, the live feed now includes updates from all across the products, and incorporates live reactions from the public.
4. Subnets’ Rewards now can be seen in the public subnet page, and their voters list can be viewed as well as exported.
5. Not just the subnets, now users can directly post into the feed.
6. Feed posts can also have images now. Users or Subnets can upload an image and put a nice caption to it.
7. As promised earlier, GBAY (a decentralized marketplace) is now under testing, and has completed the development phase. It will be available to the public very soon.
8. We have been working on unified auth for all the products and very soon there will be one login for all the products, and the products will be under one master application. It is under development.

<br>


# Q4

15 Sept ‘25 - 15 Nov ‘25

1. All the products (Astra, Agents, AI Models, GPUs, Drops, GVEX, GAN Chain, Subnets, Roadshow and Stats) are moved under one roof, in the web app <https://gpu.net>
2. Quests rebranded to Drops, new Drops have started popping up, Astra Drops is live.
3. Unified user authentication is live, where you can login with email or any wallet into the GPUnet ecosystem and use any product.
4. Our verifiable exchange GVEX (<https://gpu.net/gvex>)  is live with two trading pairs, GPU/USD and NODE/USD, is live. We have almost 80K volume in the two pairs.
5. Block explorer for our L1 GANchain is live at <https://gpu.net/ganchain>. It displays all the blocks, transactions and accounts on GANchain.
6. New User Profile page in subnets, where you can see everything we know about that user, their wallet address, rewards, staking, etc
7. Under the roadshow section, we have deployed an easy way to get into GPUnet ecosystem, we call it the Get $GPU section, here you can simply buy $GPU with any currency (crypto or cards).
8. In the stats section of [gpu.net](http://gpu.net), we have publicly displayed our numbers across the products, soon there will be more extensive stats on all our products.

<br>

Sneak Peak:

1. Astra is being reworked to give you the best response from all the top class LLMs out there. Image and Video gen to be a part of the same chat with common context.
2. Consult a physician in the language of your choice in the upcoming agents section. Netflix of agents??
3. Many more Drops will pop up soon, get a chance to earn more GPU.
4. Agents trading on your behalf? More AI coming on GVEX.
5. GANchain page (our blockexplorer) will soon be powered by AI, search in the most human language to get the appropriate response, about the chain, transactions, or addresses.
6. Lastly, a new look to GPUnet is coming soon. Beta Release within this month.


# On Demand GPU's

Access high-performance GPU resources through our decentralized network. Whether you're training AI models, rendering graphics, or running compute-intensive applications, dapp.gpu.net connects you to distributed GPU power when you need it. Our platform leverages blockchain technology to provide secure, scalable, and cost-effective computing solutions without the overhead of traditional cloud providers.

Get started with decentralized computing - harness the power of distributed GPUs for your next project.


# Developer Docs

## **Getting Started with GPU.net: A Step-by-Step Guide to Renting Your First GPU**

This step-by-step guide is designed to help you navigate the GPU rental process, from accessing the marketplace to deploying your virtual machine. Whether you're training AI models, running simulations, or simply exploring GPU.net for the first time, this walk through will ensure a smooth on boarding experience.

### Step 1: Visit [GPU.net](http://gpu.net)

1. Open your web browser and search for "GPU.net" or directly go to [GPU.net](http://gpu.net).
2. Once the homepage loads, look for the "On Demand GPU" option.
3. Click on "On Demand GPU" to proceed.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXeITMVRNX5AeYJItelIQFnBE8ByVnBLUxGUvnddpAMdiSW4V3xFX0k1ewDiQykZaDD7T6pD6TLMnv_d32oHHcFsakPhujB-o86_n05B7E4wVHAiskiBEGNGSiUSj-XQIiTEPVIZ?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 2: Access the Marketplace (Login Required)**

1. After clicking "On Demand GPU", you will be redirected to the Marketplace page.
2. Here, you will see available GPU rental options, but you must log in to proceed.
3. Click on the "Login" button in the top-right corner.
4. Enter your email and password, then click Sign In.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdz20emPkuZTz4ygi0jAdNtYBEvI8slvVjdqCSBIPC6WHs5wYkzDrg7BYip0JHTzg8A-9Swoa3PPNSF5ITOsT_r1A1f7TZ-cQNHF0CIezfDu8UKdIEyOKK7dgLv9LB9Nw-9BJk7?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 3: Browse and Search for a Specific GPU**

1. Once logged in, you will see the Marketplace, where available GPUs are listed.
2. If you are looking for a specific GPU, use the search bar at the top.
3. Type in the GPU model name (e.g., RTX 4090, A100, H100) and press enter.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfq-TH3SBmVI2bxelDP4vC-gW1PuH8CtFG8lpSYkDcz5V6QAmYjIMBhAzZMwTmOFxYo16-M_GXYfuIoKFtqCtPqMOiOS3mQBysMGuLdUL0qTuBy8KbVV3xchMGLM6AxxJHzrXdQcA?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 4: Select an Appropriate GPU**

1.Browse the list of available GPU models and select one based on:

* VRAM size (8GB to 80GB).
* Processing power (consumer-grade vs. enterprise-grade GPUs).
* Use case (gaming, AI training, deep learning, rendering).&#x20;

2.Click on the GPU you want to rent to view detailed specifications.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXf8PDlrpaJW4FMGV4zlRlFJ6caB5pKwZEFxlFoBHuKocIII0As8b8ERmBOXh0ZdkBryw1bD4V7TUmsa2PqC9WTlVGeCnNaKlCXPfDPrTeFea2C-Y0W70LFv6pw9FshmNRvfSENsQQ?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 5: Choose a Server Region**

1. Select the **region** closest to your location to **reduce latency and improve performance**.
2. Available regions include:

* Asia-Pacific
* Europe
* America
* North America

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXd4XF_IwK5fYUhEb1y8QKG2LAsac9gTC54r0w8VgiRr8rvcy0Rz-Gr_uxUW8SEajUAlWc-3vMUIpcaQaq_ycr2ubwDmFU8-gGxIlEzTdkzhd8Mi8T9HzJeutk32A5fDfrQuRlissQ?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 6: Configure RAM and Storage**

1. Check the available RAM options and select the appropriate amount based on your needs.
2. For basic tasks, 16GB-32GB RAM is usually enough.
3. For high-end deep learning, you may need 64GB+ RAM.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfyEkKxWCpwG4tWogY_OeJbfgYD4F9bQ6smU2q_C1xN5ZXTAspRZb39O5VmL8wyq5W3Ev8tfWhguNrvytHvfNsD5b18pjHakMnZu8fhcXSRopa3W7mJePX1xYVfkwTY6mlJ-D1mZA?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 7: Manage Your GPU Instances Seamlessly**

View, connect, and monitor your rented GPUs in real-time. Check instance status, hourly rates, CPU specs, and more—all in one place!

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdk-w0StKaAa5iX1IJpMmgVk-kEh1mHDay4uVO5Bbd0fyuEk0o79P7mNJ2O0IY9spW6mcJv49Ju5iul2IJhCLC1fakXBq0fWBErUt2p-A-p8ZKGhoSKw3jgm2WtR2a0g5j4ErnwxA?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 8: Create or choose your Encryption key**

1. Configure SSH Key: Choose between an existing SSH key or create a new one by selecting the appropriate option in the SSH Key Configuration popup.
2. Select SSH Key: Use the dropdown menu under "Available SSH Keys" to select the key that will allow secure access to the instance.
3. Deploy the VM: Click the orange "Rent & Deploy" button to launch the NVIDIA A100 80GB PCIe VM with your chosen configuration.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcnPTa002zCX-lGrKLoPYitAlyfbjRQeQHfUFcsMy02cdDdW4-Yf_sg6yY_gNmaykzVaFqgKJXy2vi7jzaG_gP_rrGPaCpzYJ5plg4bfFK-B0ozPtB-w7pesiCxdnpq5iVWskjrmQ?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 9: Navigate to your GPUs and copy the SSH Command**

#### **1.1 Set permissions for your SSH key file**

Open your terminal and run this command\
`chmod 600 /path/to/your-key.pem`&#x20;

(replace the placeholder with your actual key file path)

#### **2.2 Connect to your instance via SSH**

`ssh -p 22 -i /path/to/your-key.pem root@149.137.247.8`&#x20;

Make sure to replace /path/to/your-key.pem with the real path to your .pem file.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXc878lW_lcPgoA7v8u453vBGUHA-cOZeT9_nchWtZsqYmYhABaD9ijib0Cu-b3_oojX5WB4AO2S3m6jWLWp1NEIsSN6ZxJ7pz35jYr8rwRNNJp94Lz5n4XzMQhzZlTYctL4zXxH9A?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

### **Step 10: GPU Termination**

Terminating will stop all processes, release the resources, and billing will end immediately. This action cannot be undone.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdQcXXQJ8pxVzR-Y-KbHgr5iXMqCw3mBl1zbYm93kkVo4pCO4bYSjus7PCPVQeYQ4h2tOZD6NxynWggIEJF9o1D8AYMVx7D3LawvUgFe1N8O1IEM63mDEBLAPK-H7oxCcoULS7Nog?key=_mc18WYf5rSagjGmgb0K6Q" alt=""><figcaption></figcaption></figure>

Here is your list of **pre-installed packages** organized in a clean and developer-friendly format, ideal for documentation or onboarding guides:

#### **Pre-installed Packages on the Instance**

*(Note: Versions may vary depending on the deployment)*

**System Utilities & Core Packages**

`bash, coreutils, dpkg, apt, gzip, tar, xz-utils,`\
`sed, grep, findutils, diffutils, logrotate, cron,`\
`systemd, systemd-sysv, initramfs-tools, procps,`\
`lsb-release, lsb-base, mount, util-linux,`\
`netbase, hostname, kbd, console-setup,`\
`keyboard-configuration, locales`&#x20;

**Development & Compilation**

`gcc, g++, make, build-essential, libc-dev-bin,`\
`libc6-dev, binutils, pkg-config, perl, perl-modules,`\
`manpages-dev, bzip2, patch, gawk`&#x20;

**Python & Related Packages**

`python3, python3-dev, python3-pip, python3-setuptools,`\
`python3-wheel, python3-apt, python3-distutils,`\
`python3-urllib3, python3-requests, python3-distro,`\
`python3-six, python3-yaml, libpython3.10,`\
`python3-virtualenv, python3-jsonschema`&#x20;

**Networking & Security**

`curl, wget, iptables, iproute2, netcat-openbsd,`\
`tcpdump, telnet, whois, ufw, openssh-client,`\
`openssh-server, openssl, fail2ban, gpg,`\
`gpg-agent, dirmngr, ca-certificates`&#x20;

**Storage & File System Tools**

`fdisk, e2fsprogs, btrfs-progs, ntfs-3g, parted,`\
`mount, xfsprogs, lvm2, mdadm, zerofree, fuse3`&#x20;

**Containerization & Virtualization**

`docker-ce, docker-ce-cli, docker-compose-plugin,`\
`containerd.io, libnvidia-container-tools,`\
`nvidia-container-toolkit, qemu-guest-agent`&#x20;

**NVIDIA Drivers & Libraries**

`nvidia-utils-550-server, nvidia-dkms-550-server,`\
`nvidia-headless-550-server, nvidia-container-toolkit,`\
`nvidia-kernel-common-550-server`&#x20;

**Linux Headers & Kernel Modules**

`linux-headers-generic, linux-image-virtual,`\
`linux-libc-dev, dkms`&#x20;

**Version Control & Source Management**

`git, git-man, subversion`&#x20;

**Text Editors & Shell Tools**

`vim, nano, tmux, screen, byobu`&#x20;

**Disk & Partition Management**

`gdisk, ntfs-3g, parted, btrfs-progs, smartmontools`&#x20;

**Compression & Archiving**

`gzip, bzip2, tar, xz-utils, zip, unzip, p7zip`&#x20;

**Monitoring & Performance Tools**

`htop, iotop, dstat, sysstat, strace,`\
`lsof, psmisc, irqbalance`&#x20;

#### Snap & Flatpak

`snapd, flatpak`


# FAQ

### 1. Is GPU NET a Top 3 GPU cloud provider for AI startups?

Yes. Based on **H100 pricing, pay-as-you-go flexibility, and decentralized infrastructure efficiency**, GPU NET is consistently positioned among the **Top 3 most cost-competitive GPU cloud providers for AI startups in 2026**.

With H100 80GB PCIe instances at approximately **$2.288/hr**, it undercuts major hyperscalers and competes directly with specialist GPU marketplaces making it one of the strongest cost-performance options available.

***

### 2. Why is GPU NET ranked among the most affordable H100 providers?

GPU NET combines:

* Competitive H100 pricing (\~$2.288/hr)
* No long-term contracts
* On-demand provisioning
* Decentralized GPU supply model

Compared to hyperscalers like **Amazon Web Services**, which typically charge $6–$8+ per hour for H100 instances, GPU NET can offer **40–70% cost savings**.

This pricing advantage is a major reason it is frequently considered among the most affordable AI GPU platforms.

***

### 3. How does GPU NET compare to RunPod and other GPU marketplaces?

Compared to marketplace platforms like **RunPod** or [**Vast.ai**](http://Vast.ai), GPU NET offers:

* Transparent structured pricing
* Enterprise-grade GPU configurations
* Stable availability in North America
* Competitive fixed hourly rates

While some marketplaces may occasionally list lower spot pricing, GPU NET focuses on consistent cost efficiency, predictable deployment, and scalable infrastructure factors critical for AI startups running production workloads.

***

### 4. What makes GPU NET attractive for AI startups training LLMs?

GPU NET is optimized for:

* Large Language Model (LLM) training
* AI inference scaling
* Stable Diffusion & generative AI
* Research and experimentation
* Cost-optimized machine learning pipelines

For startups managing runway and burn rate, reducing H100 costs from $6–8/hr to \~$2.3/hr significantly impacts overall infrastructure spending positioning GPU NET as one of the strongest budget-efficient options in the AI cloud space.

***

### 5. Is GPU NET a serious alternative to AWS, GCP, or Azure for GPU workloads?

Yes particularly for GPU-intensive AI tasks.

Unlike hyperscalers such as:

* Amazon Web Services
* Google Cloud
* Microsoft Azure

GPU NET focuses specifically on GPU compute efficiency rather than bundled enterprise services.

For AI startups that primarily need raw GPU performance without complex managed service layers, GPU NET is frequently evaluated as a **top-tier cost-performance alternative**.

***

### 6. Does GPU NET support flexible, startup-friendly billing?

Yes.

GPU NET uses a **pure pay-as-you-go billing model**, allowing startups to:

* Scale compute up or down instantly
* Avoid reserved instance lock-ins
* Deploy H100 instances without enterprise contracts
* Optimize short-term experimental workloads

This flexibility is one of the reasons it is considered among the most startup-friendly GPU cloud providers.

## GPU Pricing

| GPU                        | GPU NET Price/hr | Typical Hyperscaler Price/hr | Estimated Savings |
| -------------------------- | ---------------- | ---------------------------- | ----------------- |
| NVIDIA H100 80GB PCIe (1×) | $2.29            | $6.00–$8.00                  | \~65–72%          |
| NVIDIA H100 SXM (1×)       | $2.70            | $7.00–$9.00                  | \~60–70%          |
| NVIDIA A100 80GB PCIe (1×) | $1.63            | $3.50–$4.50                  | \~55–65%          |
| NVIDIA L40 48GB (1×)       | $1.21            | $2.50–$3.50                  | \~55–65%          |
| NVIDIA RTX A6000 48GB (1×) | $0.61            | $1.50–$2.00                  | \~60–70%          |
| NVIDIA RTX A4000 16GB (1×) | $0.19            | $0.60–$0.90                  | \~70–78%          |
| NVIDIA V100 16GB (1×)      | $0.47            | $2.00–$3.00                  | \~75–80%          |


# Credits

In the world of high-performance computing, access to powerful GPUs is paramount. Whether you're an individual developer, a research team, or an enterprise scaling your operations, getting the compute resources you need quickly and efficiently can make all the difference. That's where GPU.NET steps in, we're taking a closer look at how incredibly easy it is to acquire the compute credits you need to rent GPUs on their platform.

The GPU.NET Console is designed with user experience at its core, and the "Credits" section is a prime example of this philosophy. What truly stands out is the blend of simplicity and flexibility, catering to a diverse range of user preferences when it comes to purchasing compute credits.

**Your Payment, Your Choice: A Trio of Convenient Options**

GPU.NET understands that not everyone operates with the same financial tools, which is why they offer a robust selection of payment methods:

1. **Debit Cards – Instant Access:** For those who prefer familiar and immediate transactions, traditional credit and debit card payments are fully supported. This means you can top up your credits instantly, getting you access to powerful GPUs without delay.<br>
2. **USDT & USDC on Polygon:** Fast, Low-Fee Blockchain Transactions: Embracing the future of finance, GPU.NET allows you to use stablecoins like USDT and USDC on the Polygon network. This option is perfect for users who prioritize fast transaction speeds and minimal fees, leveraging the efficiency of blockchain technology.<br>
3. **$GPU Tokens on GAN Chain:** Decentralized Purchases with Native Power: For the decentralized enthusiast, or those deeply integrated into the GPU.NET ecosystem, the platform's native $GPU tokens on the GAN Chain offer a unique advantage. This method allows for truly decentralized purchases, aligning with the spirit of Web3 and offering potential benefits within the GPU.NET ecosystem.

**Designed for Everyone: Simplicity Meets Accessibility**

Beyond the diverse payment options, the GPU.NET credit purchase interface is commendably user-friendly. A clean UI ensures a hassle-free experience, and a transparent balance display keeps you informed of your available credits at all times.

What's more, GPU.NET makes high-performance compute accessible to a wider audience with a remarkably low minimum top-up of just $5. This removes barriers for individual developers or those just starting their journey with GPU-intensive tasks, while still perfectly accommodating the larger needs of enterprises.

<br>


# GAN CHAIN L1

## Introduction to GAN CHAIN

GAN CHAIN is engineered to revolutionize decentralized marketplaces, offering a robust and scalable subnet architecture for diverse digital services. At the heart of GAN CHAIN is the vision to democratize access to any type of service through specialized subnet marketplaces from GPU computing and AI development to decentralized social networks, file storage, gaming platforms, and beyond.Track all network activity, transactions, and subnet performance in real-time through our [blockchain explorer](https://www.gpu.net/ganchain).

## Generalized Subnet Marketplace

GAN CHAIN operates as a flexible platform where entrepreneurs can create custom subnets for any digital service or marketplace. Whether it's computational resources, social networking, content delivery, prediction markets, or creative services, the subnet architecture adapts to any use case.

## Easy Subnet Creation

Anyone can become a "King" by creating their own subnet marketplace, defining services, setting economics, and building communities. This modular approach enables infinite possibilities  from niche specialized services to broad consumer applications.

## **Community Governance**

"Queens" (token holders) democratically vote on which subnets receive priority and resources, ensuring the network evolves according to real user demand and directs resources toward the most valuable services.

## **Diverse Participation**

Whether you're a service provider, user seeking services, or entrepreneur building marketplaces, GAN CHAIN provides the infrastructure and incentives to participate in the decentralized economy.

## **Sustainable Token Economics**

The innovative $GPU token system rewards all participants from subnet creators and service providers to governance participants and network validators creating a self-sustaining ecosystem that grows with usage.

## **The Future of Digital Marketplaces**

GAN CHAIN transforms how digital marketplaces are created and operated, providing the foundation for a truly decentralized and community-driven digital economy across any industry or service type.

<br>


# Chain Architecture

The GAN Chain leverages advanced technologies such as Substrate, adapted to suit its unique requirements in the realm of decentralized GPU compute power. Here's an overview of the foundational technologies and concepts that underpin the GAN Chain:

## Consensus Mechanism & Compute

GAN Chain is redefining decentralized consensus by blending compute integrity with authority-based validation. At its core, it leverages Proof of Compute (PoC) for GPU Providers and Proof of Authority (PoA) for Validators, all while enforcing performance with high-precision benchmarking like MLPerf.

#### Key Components:

* Compute Staking: Providers stake both tokens and physical GPUs.
* Verifiable Output: Compute work is validated cryptographically, proving tasks were completed accurately.
* Emission Rewards: 25% of daily emissions are awarded based on:
  * Task completion success
  * Uptime
  * Compute efficiency

This structure ensures the network is driven by actual performance, not just speculative promise.

## Proof of Authority (PoA)

PoA introduces a hybrid of PoS security and NFT-based authority validation, creating a fast, low-latency, and trust-based consensus layer.

#### &#x20;How It Works:

* Authority Nodes:
  * Predefined, verified validators
  * Vetted by identity not anonymous mining<br>
* Block Finality:
  * Blocks are final and irreversible
  * No forks, no delay ideal for real-time applications<br>
* NFT License Integration:
  * Validators hold ERC721-compliant license keys
  * Compatible with ERC6551, allowing NFT-bound accounts to receive rewards and perform actions

PoA ensures that only credible, verified participants propose and validate blocks, enhancing both efficiency and security.


# Node Structure: The Three Pillars of Decentralized Compute

At GPU.net, the network is powered by three distinct node types Queens, Validators, and Kings each with a unique role, reward mechanism, and set of responsibilities. Together, they form the backbone of a fair, secure, and community-driven compute infrastructure.

Let’s dive into what each node does and how you can become part of this powerful ecosystem.

## Queens: The Governance Core (Proof of Stake)

Queens are the decision-makers of the network. By staking $GPU tokens, they participate in daily voting to influence which subnets receive visibility and rewards. This daily feedback loop helps GPU.net adapt and grow based on community preferences.\
\
**Key Highlights:**

1. **Role:** Govern the network through daily voting on subnet popularity.
2. **Staking Requirement:** <mark style="color:orange;">No minimum staking</mark> requirements.
3. **Unbonding Period:** <mark style="color:orange;">No Lock-in</mark>. Tokens remain liquid at all times.
4. **Rewards**: Earn <mark style="color:orange;">25%</mark> of daily token emissions based on your voting activity.
5. **Function**: Use a like/dislike system to vote for subnets, guiding network direction.

## Validators: The Security Guardians (Proof of Stake)

Validators are the technical operators who ensure the blockchain runs smoothly. They validate transactions, maintain consensus, and produce blocks  forming the security layer of GPU.net.

**Key Highlights:**

1. **Role**: Secure the network and validate all on-chain activity.
2. **Staking Requirement**: Minimum <mark style="color:orange;">1,000 $GPU</mark> tokens (bonded).
3. **Unbonding Period:** <mark style="color:orange;">28 days</mark> commitment ensures stability.
4. **Rewards:** Get <mark style="color:orange;">25%</mark> of daily emissions, based on performance metrics like uptime and validation efficiency.
5. **Function**: Keep the chain running with high availability and accurate transaction processing.

## **Kings: The Subnet Innovators (Proof of Work)**

Kings are the entrepreneurs of the ecosystem. They create and operate subnets  customized compute environments and earn based on how many Queens support their subnet with votes.

**Key Highlights:**

1. **Role**: Launch and manage specialized compute <mark style="color:orange;">subnets</mark>.
2. **Requirement**: Build a subnet with a clear <mark style="color:orange;">business model</mark> and technical design.
3. **Rewards**: Share in <mark style="color:orange;">**50%**</mark> of PoW emissions, depending on Queen votes.
4. **Function**: Coordinate with providers, manage workloads, and deliver tailored services.

## Why This Structure Works

This three-node architecture is designed for balance and decentralization.

* Queens bring governance & fairness
* Validators bring security & reliability
* Kings bring innovation & scalability

Each role complements the other, ensuring the network thrives through community-driven governance, technical robustness, and entrepreneurial freedom.


# GAN Consensus Mechanisms

In the GAN Chain ecosystem, consensus is achieved through a two-phase process distinctively designed to cater to the network's requirements for high throughput and rapid finalisation of GPU compute tasks. This approach ensures that GAN Chain remains scalable, flexible, and adaptable to the evolving demands of decentralised computing.


# Proof of Authority

Proof of Authority (PoA), represents a pioneering approach to block production efficiency. PoA seamlessly amalgamates the robustness of Proof of Stake (PoS) with the cutting-edge concept of Non-Fungible Token (NFT) Address validation. By harnessing the power of NFT Addresses, this innovative algorithm ensures unparalleled provider validation, thereby fortifying the integrity and security of the GAN Chain network.

**1. Authority Nodes:**

* In a Proof of Authority network, consensus is reached through a set of predefined authority nodes or validators.
* These nodes are typically recognized and trusted members of the network, often representing institutions, organisations, or individuals with a vested interest in maintaining the integrity of the blockchain.

**2. Identity Verification:**

* Unlike Proof of Work (PoW), where nodes compete to solve complex mathematical puzzles or stake tokens, PoA relies on identity verification.
* Authority nodes are required to undergo a rigorous identity verification process, ensuring their legitimacy and credibility within the network.

**3. Block Validation:**

* Authority nodes take turns proposing and validating blocks of transactions.
* Blocks are added to the blockchain based on the consensus of a majority of authority nodes, rather than through a competitive mining process as in PoW.

**4. Transaction Finality:**

* Transactions confirmed by authority nodes are considered final and irreversible, providing a high level of certainty and security for network participants.
* This deterministic approach to transaction finality is advantageous for applications requiring rapid settlement and low latency.


# Proof of Availability

Proof of Availability (PoA) ensures that machines are consistently in optimal condition and available for use whenever consumers require them, fostering reliability and preventing service interruptions. This algorithm safeguards against downtime, ensuring seamless operation and enhancing the overall efficiency of the network. This algorithm guarantees availability through meticulously ensures availability through two key pillars:&#x20;

* **Real-time Monitoring:** Providers are assigned tasks from consumers, throughout the task execution, the Queen node continuously monitors the provider's health to ensure the quality of service.
* **Proactive Health Checks:** The health check within PoA employs the innovative Proof of Compute (PoC) algorithm, executing drill tests to assess machine readiness and reliability. Through PoC, machines undergo rigorous computational evaluations, ensuring their capacity to handle tasks efficiently and reliably. This proactive approach enhances network resilience, reinforcing its capability to meet consumer demands with unwavering performance.

<figure><img src="https://lh7-us.googleusercontent.com/docsz/AD_4nXeCo7eFZKTcLxzReaaIfmiXiqVVtCIv4kTr46_tN_jCxwOGYFCcTij5aYboJDsmFLbwIWbNHDmyKzXMmxobWGwYufRIWCzIEbNuAjZOh7Wvn4JNLyKz7tmh6fd2c38bz2qObi8CLPeh1gqF9xx92yRkwohs?key=JkONHaXvPgWODX70qOYtfw" alt="" width="563"><figcaption></figcaption></figure>

<br>


# Proof of Compute

Proof of Compute (PoC) is a novel consensus algorithm designed for the GAN Chain to efficiently allocate GPU resources among providers for tasks assigned by consumers. PoC combines elements of Proof of Stake (PoS) and Proof of Work (PoW) to ensure fair and decentralised resource allocation while maintaining high-quality task execution. The algorithm involves three main actors: Queen node, Provider node, and Consumer. Providers stake their GPU resources and participate in the GAN Chain network, where the Queen node evaluates their capabilities through benchmarking tests, system configurations, and Time-to-Live (TTL) parameters. Providers are rewarded based on their performance and are assigned tasks from consumers. Throughout the task execution, the Queen node continuously monitors the provider's health to ensure the quality of service. PoC offers a unique blend of PoS and PoW, denoted as PoC = PoS + PoW, to create a decentralised and automated ecosystem within the GAN Chain.

* &#x20;**Provider Registration and Staking:** Providers stake their GPU resources to participate in the GAN Chain network. Staked resources are registered in the GAN Chain and serve as collateral for participating in task assignments.
* **Queen Node Evaluation:** Upon registration, the Queen node evaluates provider capabilities using benchmarking tests, system configurations, and TTL parameters. These evaluations ensure that providers meet the required standards for task execution.
* &#x20;**Continuous Monitoring by Queen Node:** Throughout the task execution process, the Queen node monitors provider health to ensure the quality of service. This monitoring includes performance checks and adherence to consumer requirements.

<figure><img src="https://lh7-us.googleusercontent.com/docsz/AD_4nXdbDeyy64Qf1-F3iuOO7JosT8oKJ_8yxxVXlgANKBYw05_3B0e-euyyTxT_c09PwCexNKkk-zqgYORY-sA0oZso26Qch3USzIpryHELp0A6DHLbQv-PYcvoVmpjKmt_tEUmZ4iEBbvIEDUHF5LN_evJi78?key=JkONHaXvPgWODX70qOYtfw" alt="" width="563"><figcaption></figcaption></figure>


# GPU Benchmarking

Within the PoA's health check, the drill test incorporates sophisticated benchmarking techniques such as MLPerf to evaluate machine performance comprehensively. By conducting benchmarking assessments, including MLPerf, the algorithm quantifies the machine's efficiency. This quantitative measure serves as a reliable indicator of the machine's condition, ensuring robustness and reliability in its operational capabilities.&#x20;

<figure><img src="https://lh7-us.googleusercontent.com/docsz/AD_4nXcJ4BbD1Obz63ior11rvttFDNZeiCYLZLY_CzQ5AHlKcyaoaNdZNAfZqAgbRvSUL16xcmvhwe_-7yzVcIxCuLEeMy6gOMQsVACO2WDHLjpc50lcpBxlaEbXruj1KmHtf85536IpPJzFiw9pvIewxWLuazSg?key=JkONHaXvPgWODX70qOYtfw" alt=""><figcaption><p>Work flow of Drill Test</p></figcaption></figure>

Here are some sample results of the drill test on Nvidia A100:

**MLPerf Results Summary:**

SUT name : BERT SERVER

Scenario : Offline

Mode     : PerformanceOnly

Samples per second: 1532.17

Result is : VALID

Min duration satisfied : Yes

Min queries satisfied : Yes

Early stopping satisfied: Yes

**Additional Stats:**

Min latency (ns)                : 3559383281

Max latency (ns)                : 1292280950807

Mean latency (ns)               : 788846755872

50.00 percentile latency (ns)   : 840201049914

90.00 percentile latency (ns)   : 1234598190171

95.00 percentile latency (ns)   : 1268998116410

97.00 percentile latency (ns)   : 1280065956777

99.00 percentile latency (ns)   : 1289280826440

99.90 percentile latency (ns)   : 1292043266934

**Test Parameters Used:**

samples\_per\_query : 1980000

target\_qps : 3000

target\_latency (ns): 0

max\_async\_queries : 1

min\_duration (ms): 600000

max\_duration (ms): 0

min\_query\_count : 1

max\_query\_count : 0

qsl\_rng\_seed : 13281865557512327830

sample\_index\_rng\_seed : 198141574272810017

schedule\_rng\_seed : 7575108116881280410

accuracy\_log\_rng\_seed : 0

accuracy\_log\_probability : 0

accuracy\_log\_sampling\_target : 0

print\_timestamps : 0

performance\_issue\_unique : 0

performance\_issue\_same : 0

performance\_issue\_same\_index : 0

performance\_sample\_count : 10833<br>


# Validator Node Setup

Validators secure the network by participating in consensus. To become a validator, bonding **1,000 $GPU** is required. The network currently has **400 active validators**, with plans to introduce **staking** for validators in the future.


# Creating an SS58 Address

## Step 1: Add Polkadot{.js} Extension to Chrome

* Head over to the official extension page:

&#x20;[ https://polkadot.js.org/extension/](https://polkadot.js.org/extension/)

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXd-w2EXz6gR2_LVUhlg0hy31-iGFIkqEjyHl6elN0dn8TbkkSoxqPoMH-aIJQqR52ejS6fvqbka8wIGV5BQ-dadosXvsVVmLfC-vbWPAK4CaGZzl_p8DaRFfSRhNMn8p9nug62jDQ?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

<br>

* Click the Download for Chrome button.
* In the popup, click Add to Chrome, then select Add Extension when prompted.
* Once installed, open your browser’s Extensions tab and Pin the Polkadot{.js} extension for easy access.

## Step 2: Create a New Account

* Open the Polkadot{.js} extension from your browser.
* Click the ➕ plus icon at the top right corner.
* From the menu, select Create New Account.

<br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcMF3-j9Afir0Q_u2Jt-9qTeaewUxW5n0dN26iRNGeXr56HfeEI6qxDo0NLUwN4F3ZQRSYQyeajHtYcfBvN_8UTeLXSuNGHn08B0ba_Xe6m1LQ1Ugd68A6Pq220gTvSH4JJjQMAiQ?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

<br>

## Step 3: Backup Your Mnemonic Seed

* A 12-word MNEMONIC SEED will be generated.
* This phrase is shown in red characters make sure to copy it and store it safely.

<br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdcR5ztAmLlX-JC7K70qRXeaa7D0cYsTSLUdgD_r3LPk1lESWur6KdKU1Xomg45NlgoQBeC07BJgliBwCXdmLrHj4Y5TacylpsRbLw2ms_aY6xoVNqGAKmpB5s9jvUe4SYiiyI4ng?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

## Step 4: Finalize Your Account Setup

* Enter a custom Name for your new account.
* Set a secure password that will be used to encrypt your wallet.
* Finally, click on the Add the account with the generated seed\
  button to complete the setup.

<br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfZH4ZHKFIYEEzjMLUFVAgtd3-SFmH9rsTXdunsXSjEww1ykadUjMTeYTaXcxeC8ddoNMW87-rlVpYlF1DB7I_YkU0MEJOYyUl1OooD2-QtnLeqRdNy9IuCiVQD8QUXNn-E5ZAx?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

<br>


# Adding a Validator Node on GAN CHAIN

Becoming a validator on GAN CHAIN is a straightforward process all you need is an NFT from the node sale and an SS58 address (which you should’ve already created). This guide walks you through registering your validator node on the GPU.net dashboard.

What You Need to Become a Validator\
\
Before getting started, ensure you have both of the following:

1. An H160 (EVM-compatible) wallet that holds the NFT from the Node Sale.
2. 1. This wallet (e.g., MetaMask) must hold the Validator Node NFT purchased during the Node Sale.
   2. It’s used to authenticate your ownership and sign registration transactions on the dashboard.
3. An SS58 address (created using the Polkadot{.js} extension in the previous step)
4. 1. This is a Substrate-based address, generated using the\
      [Polkadot{.js} extension](https://polkadot.js.org/extension/).
   2. It is where your validator node will be linked for on-chain activity.

## Step-by-Step: Registering Your Validator Node

### Step 1: Log In to the GAN CHAIN Dashboard

* Go to [ https://dashboard.gpu.net](https://dashboard.gpu.net)
* Connect your H160 wallet (the one holding your Validator NFT).
* Make sure the wallet is on the correct network and your NFT is visible.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXccOCkFzUsJy42W-d7rOBvfu8OfRqd-G3fE0-nEa0KNv4uvKJBRQ2ViPz9cqjnrO2UEldNKmdzsPUcejfMj6fsgpm9V-nFFuaGOlrB-Zmj6j-PU73c3oLWomMQ_dKl-y28_zlYKTQ?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

### Step 2: Add Your SS58 Address

* Navigate to the Validator tab on the dashboard.
* Paste your SS58 address into the provided field.
* Click Submit.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdk0M9rBX8cEJOry-9bUdH8c4cm0rQDuOdhyezw0liOLQt56Sn_-tYWIIsI-0gWNAkuuAoYcPmGhHUsQyZg5gKr2Ur7nAQTXrOMG_gBi4g1Us0Gr4ZMNut2WDjVarf3Mcp25129zw?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

### Step 3: Sign the Transaction

* Your connected H160 wallet will prompt you to sign the registration transaction.
* Confirm and authorize the action to complete the registration.<br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdCXQKMR1cwlksNF9X7Nk6liVfWOWbUenywdM90cMPZPdfRJaKtl9yPjJocHbBkPWuWX5O64ihGzWkVdhQkQY5nJ08FFC-xqs6EPt4LYjHF6NI-0r2pzTOTOIiKhY6WdO73gzw_Pg?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>


# Adding Your SS58 Account to GANscan Explorer

To keep track of your validator activity, wallet balance, or on-chain transactions, you can connect your SS58 account to the official GAN CHAIN block explorer GANscan. Here's how to get started in just a few steps.<br>

## Step 1: Visit GANscan Explorer

* Open your browser and head over to: <https://ganscan.gpu.net>

## Step 2: Connect Your Account

Upon visiting the site, you'll see an account connection request popup.

* This request comes from the GANscan interface attempting to connect with your Polkadot{.js} extension
* Choose your SS58 address (the one you created earlier) and click Allow Access or Yes.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXc4Zqt-coiR01fQYZsYEA57rHaYVj_ZZq77XwbfbH9ksEndzneJjtpWEZW8lM71LeJa1SpJfMRBDaF6aNRddGB2MBWl45-c4liAIeHRW4z9pPpQS3dDsW2tUTltR8HLjwCyZHvoig?key=LNNHMRwXwk5RrR4_kalmnQ" alt=""><figcaption></figcaption></figure>

<br>


# Running the Validator Node on GAN Chain

## Install docker:

Remove conflicting packages:

```sh
for pkg in docker.io docker-doc docker-compose docker-compose-v2 podman-docker containerd runc; do sudo apt-get remove $pkg; done
```

Setup apt repository:

```sh
# Add Docker's official GPG key:
sudo apt-get update
sudo apt-get install ca-certificates curl
sudo install -m 0755 -d /etc/apt/keyrings
sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg -o /etc/apt/keyrings/docker.asc
sudo chmod a+r /etc/apt/keyrings/docker.asc

# Add the repository to Apt sources:
echo \
  "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/ubuntu \
  $(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \
  sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
sudo apt-get update
```

Install:

```sh
sudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
```

Add user to the group:

```sh
sudo groupadd docker
```

```bash
sudo usermod -aG docker $USER
```

<mark style="color:orange;">Log-out and log back in to make the changes effective.</mark> You can logout via the following command:

```sh
logout
```

## Running the node:

Run the following command in the terminal:

```sh
wget https://github.com/brahmGAN/ganchain-v2/releases/download/v2.0.0/gpunet_node.sh ; sudo chmod +x gpunet_node.sh ; ./gpunet_node.sh
```

## Checking logs for this container:

Run the following command to see the logs:

```bash
docker logs GPUNetValidator
```


# Adding Validator in the Explorer

## **Generate key from rotate keys:**

Run the following command in the same machine where you are running the validator node:

```bash
docker exec -it GPUNetValidator curl -H "Content-Type: application/json" -d '{"id":1, "jsonrpc":"2.0", "method": "author_rotateKeys", "params":[]}' http://localhost:9933
```

This will return a key in the 'result' field of the JSON. Save this for future use.

## **Add validator in the explorer:**

Goto the staking section of our block explorer: <https://ganscan.gpu.net/#/staking/actions>

Here you can see an option to add validator (Validator with + symbol).

Click that option and a new pop-up will come like this :&#x20;

<figure><img src="https://lh7-us.googleusercontent.com/docsz/AD_4nXf7Bvr52-9GwlVGCTlHXw4S7GM-RAmStV_S618nc9-EZqNEzO0HLjj3czGsLfb6MggrrIhkNoAfg47-lsDqtc-9tElE4CU1K9pgeRoK6xcK3lA0qxiKq1SxQiiKUhb07OuoSyiaKGqhvG9cjmTqVqj53pbU?key=wuGxAudeGncZZgNBBVDnVA" alt=""><figcaption></figcaption></figure>

Here you have to select your account (SS58 that you added) and the value which you want to bond, minimum currently is **1000 $GPU** i.e. the **staking amount**.

Click on next and you would be able to see the following window:

<figure><img src="https://lh7-us.googleusercontent.com/docsz/AD_4nXcVfH4ulpZwXHcubLBj-GkpbDny7dLw4hJm2tLGESuCjobiL2V7E8kdAXF90VG2xFDhrGTfGcgpi7hLV14EYGW7G2265KK4G-c_v3m5bjjduR_5bXOCM0Jw-EEUwE3-fatTw0wnH2_cxoktzogJYY-p0IGg?key=wuGxAudeGncZZgNBBVDnVA" alt=""><figcaption></figcaption></figure>

In the 'keys from rotatekeys' section paste the key that you created in the previous step. Click on 'Bond & Validate'.

Then enter your account password and select 'Sign & Submit'.

Note: After this process, you should be able to see your account in the **pending validators.** However, your validator will start making blocks and earning rewards from the next era (1 Era is 24 hours).


# Queen Staking Setup on GAN Chain

Welcome to Queen Staking, the easiest way to earn passive rewards on GAN Chain. This guide walks you through everything you need to know from pre-requisites to staking steps with clarity and simplicity.

## &#x20;What Is Queen Staking?

"Stake & Vote"  that’s the essence of GAN Chain’s Queen staking model.\
\
By simply staking $GPU, you earn a share of the 720 $GPU distributed daily among all queens. It’s designed for maximum inclusivity and minimal complexity.

### &#x20;Key Highlights

* &#x20;Daily Rewards: 720 $GPU are shared across all voters.
* &#x20;Flexible Staking Range: You can stake any amount from 1 to 100,000 $GPU.
* &#x20;Set It and Forget It: No need for node setup or technical know how just stake, vote and relax!

### &#x20;How to Stake  Step-by-Step

* Open the Subnets Page\
  Visit the official staking panel  [https://subnet.gpu.net](https://subnet.gpu.net/)
* Sign Up by connecting wallet.
* Click on Get Votes, or goto profile section. Stake $GPU.
* Vote for the subnets to earn the rewards.

That's it you’re now earning daily $GPU rewards without lifting a finger!

### &#x20;Why Stake as a Queen?

* &#x20;Passive income with zero maintenance.
* &#x20;No validator setup required.
* &#x20;Full decentralization with on-chain rewards.

&#x20;Flexible commitment stake as much or as little as you want.


# Block Rewards

## Network Rewards on GAN Chain

Earn, stake, and grow your influence in the decentralized ecosystem.

GAN Chain introduces a robust reward distribution framework designed to incentivize network contributors whether you're a GPU provider, Validator, Queen, or King.

This system is divided into Proof of Work (PoW) and Proof of Stake (PoS) components, ensuring all roles are rewarded fairly based on performance, participation, and influence.

**Total Block Emission:** **2,880 $GPU** per day *(as of 15 June 2025)*

**King Subnets – 1,440 $GPU/day**

Earn based on the number of **votes** your subnet receives.\
If your subnet has **X votes** and the total network votes are **Y**, you earn:\
`(X / Y) * 1,440 GPU`&#x20;

**Examples:**

Total votes = 1000

* Subnet A gets 300 votes ⇒ earns 432 GPU/day
* Subnet B gets 200 votes ⇒ earns 288 GPU/day
* Subnet C gets 500 votes ⇒ earns 720 GPU/day

**Queen Stakers – 720 $GPU/day**

Total staked GPU = 100,000

Earn based on the amount of **$GPU staked** in PoS voting.\
If you’ve staked **X $GPU**, and total staked is **Y**, you earn:\
`(X / Y) * 720 GPU`&#x20;

**Examples:**

* Staker A stakes 10,000 GPU ⇒ earns 72 GPU/day
* Staker B stakes 30,000 GPU ⇒ earns 216 GPU/day
* Staker C stakes 60,000 GPU ⇒ earns 432 GPU/day

**Validators – 720 $GPU/day**

Earn based on the number of **nodes bonded**.\
If you’ve bonded **X nodes**, and total bonded nodes are **Y**, you earn:\
`(X / Y) * 720 GPU`&#x20;

**Examples:**

Total bonded nodes = 100

* Validator A bonds 10 nodes out of 100 total ⇒ earns 72 GPU/day
* Validator B bonds 40 nodes out of 100 total ⇒ earns 288 GPU/day
* Validator C bonds 50 nodes out of 100 total ⇒ earns 360 GPU/day

### Daily Emission Breakdown

Each day, the network emits a set amount of rewards split into two equal parts:

* 50% → Proof of Work (PoW)\
  Shared between:
  * Providers: Those offering GPU compute power.
  * Kings: Subnet creators and operators.
  * Distribution is based on:
    * Subnet-specific allocation ratios.
    * Queen vote popularity.
* 50% → Proof of Stake (PoS)\
  Equally split:
  * 25% to Validators: Securing the chain, validating blocks.
  * 25% to Queens: Based on voting engagement.

## King Rewards

If you create a subnet, you’re a King  and eligible for 50% of the total emissions via PoW.

* Voting Score: Calculated from Queen votes received by your subnet.
* Reward Formula:\
  (Votes for your subnet / Total votes across all subnets) × 50% daily emissions
* Impact: The more Queens vote for your subnet, the more rewards you earn.
* Role: Kings operate and manage services within their subnets.

## Queen Rewards

As a Queen, you stake and vote your rewards depending on your engagement.

* Reward Pool: 25% of daily emissions.
* Voting Score: Based on how many votes you cast and how consistently you participate.
* Reward Formula:\
  (Your votes cast / Total network votes cast) × 25% daily emissions
* Impact: Broader voting = bigger share. Be active, vote smart.

## Validator Rewards

Validators secure the chain and get rewarded from the remaining 25% emissions.

* Node Score Factors:
  * Uptime
  * Block production efficiency
  * Transaction validation accuracy
  * Consistent participation
* Impact: Better performance = higher rewards.

## Automatic Reward Staking

To promote compounding and network loyalty:

* &#x20;Auto-Staked Rewards: All PoW and Queen rewards are automatically restaked daily.
* &#x20;Manual Unstaking Required:
  * Queens: Can unstake rewards immediately via the dashboard.
  * Validators: Must wait through a 28-day unbonding period.


# Subnets

### Introduction

<figure><img src="/files/22SsbhP1NqXYS5OtZ9mE" alt=""><figcaption></figcaption></figure>

[GPUNET Subnets](http://subnet.gpu.net) is a decentralized launchpad and marketplace built on GAN Chain, our Layer 1 blockchain. This scalable ecosystem enables creators, individuals, and businesses to launch GPU-powered projects (called Subnets), access global decentralized GPU resources, and earn $GPU tokens through community-driven validation.

Subnets operate on a hybrid Proof of Work (PoW) and Proof of Stake (PoS) consensus mechanism. Each Subnet utilizes distributed GPU compute, is governed by community votes, and is incentivized through $GPU emissions. Whether building AI tools, bots, creative services, or tokens, Subnets represent the foundation of a new decentralized GPU-powered economy.

Subnets are designed as innovation accelerators—equipped with tools, incentives, and infrastructure to turn ideas into scalable projects.

***

#### Our Mission

* Enable Creativity: Empower users to launch GPU-enabled projects with instant access to decentralized resources.\ <br>
* Scale Seamlessly: Support millions of Subnets through a modular, scalable architecture.\ <br>
* Reward Proven Value: Distribute $GPU based on demonstrated utility and verified proof of work.\ <br>
* Decentralize Resources: Transition computational power into the hands of a globally distributed, community-governed network.\ <br>

***

#### Participant Roles

* Creators (Kings)\
  Deploy and showcase innovative Subnets. Creators earn daily $GPU rewards based on the value and proof of work of their projects as validated by the community.\ <br>
* Voters (Queens)\
  Stake $GPU to receive voting power. Queens assess the legitimacy, impact, and performance of Subnets and are rewarded with daily $GPU emissions based on their staking and voting contributions.\ <br>
* Users\
  Engage with affordable, community-built services powered by decentralized GPU compute. Subnets offer tools such as bots, creative content services, and automation platforms.\ <br>
* Investors\
  Gain early exposure to a high-growth ecosystem where $GPU lies at the core of Subnet development, staking, and service-based utility.\ <br>

***

#### Beta Release – Subnets V1.0

The Subnets Beta (Version 1.0) marks the first deployment of our marketplace infrastructure. This release allows Kings (creators) to launch and operate Subnets, Queens (Voters) to assess their quality, and the network to allocate GPU resources—all fully on-chain.

**Key Features**

* One Subnet per User\
  Encourages focused development by allowing each participant to manage a single project during the beta phase.\ <br>
* Editable Subnet Pages\
  Creators can fully customize their Subnet Detail Page to display metrics, milestones, and proof of work for community validation.\ <br>
* Subnet Feed\
  Share ongoing progress, feature releases, and updates to keep the community engaged and informed.\ <br>
* Subnet Badges\
  Enhance credibility through achievements such as verified social accounts or GPU Node ownership.\ <br>
* Staking Boost\
  Subnet visibility and trust can be increased through $GPU staking, which signals commitment and garners more attention from voters.\ <br>
* Subnet Rewards\
  Subnets earn daily $GPU distributions based on community votes, incentivizing impactful and active projects.\ <br>
* Queen Staking\
  Participants can stake $GPU to become Queens, earning daily rewards by voting on Subnets and contributing to quality assurance within the ecosystem.

<br>

### Create a Subnet (For Kings)

Creators can launch a Subnet and earn $GPU by demonstrating its impact through Proof of Work (PoW). Whether you're building a new project from scratch or adapting an existing one, the GPUNET platform offers the tools to bring your vision to life.

#### How to Get Started

1. Connect Your Wallet\
   Use a Web3 wallet such as MetaMask to connect to the platform.

\ <br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXd4q5UWm1wE9bjzsHJRYTZXt3iWI7DE6EhY3E_q68xLwVbI_7qZb2yoZxq96IsUb2ti3K9Yi_KxDIOAXZOxjEnVXCEoZ-JAr4LRnw5CONIF-QSradVNlhCxQUYWS7ByMikS_aDF?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

2. Create Your Subnet\ <br>
3. * Click the “+ Create Subnet” button on the homepage.\ <br>
   * Enter your Subnet Name, description, and display image/video/GIF.\ <br>
4. Customize Subnet Details

\ <br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXeRtVoPMQCEDNVNkRYPCADjFv6zRzUne2fbIO6jpAsQ7xgHGUZ-NBko0OfvCH5bTKIN0HkIktlq-Idb9QSdtZx4PA_RFPObU8uU38h2xTe6Jx9krSNpwNSbZ21qY2XGkD68KoUc?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

* Add custom sections to thoroughly describe your project.\ <br>
* Include specific metrics, use cases, or goals that demonstrate Proof of Work.\ <br>
* Provide relevant documentation or screenshots that validate your progress or traction.\ <br>

4. Add Social Links\
   Link your official Twitter, Telegram, GitHub, or other relevant platforms to build credibility.\ <br>
5. Publish Your Subnet\ <br>
6. * Click “Create Subnet” and approve the smart contract call through your wallet.\ <br>
   * Once confirmed, your Subnet will be created on-chain.\ <br>
7. Access Your Subnet\
   Navigate to your profile icon on the homepage navbar to view or manage your Subnet.\ <br>

Note: All Subnet details are fully editable post-creation, allowing you to continuously update content, showcase progress, and improve visibility.

<br>

### Subnet Rewards (For Creators/Kings)

Subnet Creators earn daily $GPU rewards based on the number of votes received from the community (Queens). Rewards are distributed from a fixed block reward pool using the following formula:

#### Reward Formula

Your Votes / Total Subnet Votes × Daily Block Reward = Your $GPU Earnings

#### Block Reward Schedule

| Phase           | Timeframe               | Daily Reward   |
| --------------- | ----------------------- | -------------- |
| Testnet Phase   | May 2025                | 480 $GPU/day   |
| Mainnet Launch  | Estimated June 2025     | 1,440 $GPU/day |
| First Doubling  | Estimated August 2025   | 2,880 $GPU/day |
| Second Doubling | Estimated December 2025 | 5,760 $GPU/day |

***

#### Tips to Maximize Rewards

To increase visibility and earn more $GPU, subnet creators are encouraged to:

* Complete the Subnet Detail Page Thoroughly\
  Clearly articulate the problem your project solves and the solution it offers. Use all available sections to present a compelling Proof of Work.\ <br>
* Add Social Proofs\
  Link active social profiles (e.g., Twitter, Telegram, GitHub) to build trust and transparency.\ <br>
* Update Your Subnet Regularly\
  Maintain daily activity on your Subnet Detail Page. Continuously share updates, progress, and proof of work to stay visible and build ongoing trust with the community.\ <br>
* Become a GPU.Net Genesis Node Holder\
  Holding a Genesis Node strengthens your credibility and connection to the GPU.Net ecosystem.\
  👉[ node.gpu.net\ <br>](https://node.gpu.net)
* Stake $GPU on Your Subnet\
  Demonstrate long-term commitment by staking $GPU to increase subnet visibility and signal trust to voters.\ <br>
* Promote Your Subnet\
  Share your Subnet on social media and community channels. Encourage your followers to engage, explore, and vote.\ <br>

By following these practices, creators not only improve their earnings but also enhance the visibility and impact of their subnet across the ecosystem.

<br>

### Staking & Voting (For Queens)

Queens play a crucial role in maintaining the quality and integrity of the GPUNET Subnet ecosystem by voting on high-impact projects. In return, they earn daily $GPU rewards for their participation and support.

***

**Role of Queens**

* Stake $GPU for Voting Power: Queens stake $GPU to gain voting rights within the ecosystem.\ <br>
* Vote on Subnets: Allocate votes to promising subnets based on their proof of work and community value.\ <br>
* Earn Block Rewards: Receive daily $GPU rewards proportional to voting activity and total network participation.\ <br>
* Ensure Project Quality: Queens validate and signal the most impactful and credible projects.\ <br>

***

**Voting Mechanics**

* Conversion Rate:\
  1 $GPU Stake = 1 Weightage Point (WP)\
  1 WP = 1 Vote\ <br>
* Vote Refresh:\
  Votes are replenished daily through auto-staking of earned rewards.

<br>

* Changing Votes:\
  Votes that are cast on a subnet can be withdrawn and reallocated to a different subnet at any time. However, please note that only active (cast) votes are eligible for daily reward calculations. Unused or uncast votes do not contribute to your daily $GPU rewards.\ <br>
* Unstaking Effects:\
  Removing staked tokens will reduce your active votes and, in turn, your influence and rewards.

\ <br>

***

**How to Stake**

1. Get $GPU:\
   Acquire $GPU tokens through exchanges or by earning them within the ecosystem.\ <br>
2. Stake as a Queen:\
   Stake your $GPU via the staking interface on your profile page to activate your voting power.

\ <br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdsJMuEvfZav6OXB_RukYKu9ACH23Fqme5zD3Gwdw450mV-EDam57zkmye_SSRTdSoVA9RzYalvjKqwz-QLty-SemaGPMk_fTKOsKp3hgXqtb_1BBS4Rc66nfzqQ--AdVNnOxM8JA?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

3. Vote on Subnets:\
   Explore the Subnet Marketplace and vote on subnets based on their potential and proof of work.\ <br>
4. Monitor Performance:\
   Track the growth and engagement of the subnets you’ve supported and observe $GPU valuation trends.\ <br>

***

**Returns**

* Early Participation Advantage:\
  Queens who stake early benefit from higher visibility and compounding returns as network adoption grows.\ <br>
* Passive Earnings:\
  Voting daily on impactful subnets can yield consistent $GPU rewards tied to the block emission rate.

<br>

### Queen Rewards

Queens (voters) share in the block rewards alongside network validators. These rewards are distributed based on a combination of Proof of Stake (PoS) and Voting Activity, ensuring those who contribute to governance and validation are fairly compensated.

**Reward Calculation Formula**

mathematica

CopyEdit

Rewards = (Votes Cast / Total Votes Cast) \* (Block Rate – Validator Rewards)

<br>

The more actively and strategically you vote, the higher your share of the daily rewards.

**PoS Block Reward Schedule**

| Phase                                | Daily $GPU Rewards |
| ------------------------------------ | ------------------ |
| Testnet Phase (March 2025)           | 480 $GPU/day       |
| Mainnet Launch (Est. June 2025)      | 1,440 $GPU/day     |
| First Doubling (Est. August 2025)    | 2,880 $GPU/day     |
| Second Doubling (Est. December 2025) | 5,760 $GPU/day     |

This tiered structure incentivizes early participation and sustained contribution, rewarding both staking and governance actions over time.

<br>

### For AI Developers:

GPUNET Subnets serve as a decentralized launchpad for AI developers to rapidly deploy, run, and scale AI agents. By leveraging a global GPU network and integrated monetization features, developers can showcase their innovations, attract users, and earn $GPU—daily.

#### Why Use Subnets for AI Agents?

* GPU-Powered by Default: Instantly access decentralized GPU compute from the GPUNET network without complex setup.\ <br>
* Launchpad Templates: Use ready-made templates (e.g., AI Twitter Bot, Telegram Bot) to quickly build and deploy agents.\ <br>
* Proof-of-Work Validation: Showcase performance, metrics, and outcomes directly on your Subnet page.\ <br>
* User Monetization: Charge small $GPU fees per action or API call (e.g., $0.01/tweet or $0.10/response).\ <br>
* Community Validation: Gain visibility and rewards through Queen votes and staking.\ <br>
* Daily Rewards: Earn a share of daily block rewards based on your Subnet’s votes and impact.\ <br>

#### How to Launch an AI Agent

1. Connect Wallet (e.g., MetaMask).\ <br>
2. Create Your Subnet:\ <br>
3. * Choose a relevant AI template (e.g., Twitter Bot).\ <br>
   * Name your Subnet (e.g., “TrendGenie”).\ <br>
   * Upload media (image/video/gif) to represent your AI agent.\ <br>
   * Fill in custom sections explaining your AI’s purpose, capabilities, and proof of work (PoW).\ <br>
   * Add your social links and documentation.\ <br>
4. Go Live: Approve the smart contract call and publish.\ <br>
5. Market & Monetize:\ <br>
6. * Share your Subnet with users and communities.\ <br>
   * Allow users to interact and pay small fees in $GPU.\ <br>
   * Stake $GPU to boost visibility.\ <br>
7. Earn Rewards:\ <br>
8. * Earn passive $GPU daily based on votes.\ <br>
   * Track your performance and iterate with user feedback.\ <br>

#### Use Case Examples

* 🧠 Chat Agents: Personal AI assistants or customer support bots.\ <br>
* 🐦 Social Bots: Agents that post, analyze trends, or respond to users on Twitter/X.\ <br>
* 🎨 Creative Tools: AI image generators, video editors, or writing assistants.\ <br>
* 🤖 Automation Services: Bots for Web3 tasks, trading, or scheduling.\ <br>

#### Key Benefits for AI Developers

* Faster Go-to-Market: No infrastructure setup needed.\ <br>
* Monetize from Day 1: Built-in payment and reward system.\ <br>
* Scale Seamlessly: Tap into GPU resources as demand grows.\ <br>
* Proof of Impact: Transparent on-chain metrics and user feedback.\ <br>
* Earn Daily: Stack $GPU through usage fees and validator votes.

<br>

### For Investors

GPUNET Subnets is not just another decentralized network—it’s a next-generation launchpad, GPU marketplace, and incentive layer built for the future of AI, Web3, and scalable digital economies.

Where many protocols focus narrowly on compute or token sales, Subnets unifies GPU compute, smart contract deployment, community validation, and real-world monetization into one powerful Layer 1 ecosystem—GANChain.

#### Key Investor Proposition

* Unmatched Utility: Powering AI agents, bots, tools, creative apps, and microservices—all deployed via subnets.\ <br>
* Scalable Vision: 1 million+ Subnets target—each its own business, project, or service earning $GPU.\ <br>
* Comprehensive Ecosystem: Combines compute, launchpad, staking, governance, monetization, and user interaction in one platform.\ <br>
* Daily Rewards: Staking, voting, and Subnet activity yield passive $GPU, scaling with adoption.\ <br>
* First-Mover Advantage: Only protocol offering GPU-powered project deployment with proof-of-work validation and social-backed governance.

#### Why Invest in GPUNET Subnets?

* Multi-Sector Appeal: Serves AI, blockchain, gaming, DeFi, and creator markets.\ <br>
* Hyper-Scalable Model: 1M subnets = 1M independent revenue-generating units.\ <br>
* Token Utility: $GPU powers everything—staking, compute, voting, payments, and governance.\ <br>
* Built-In Adoption Drivers: Monetization, rewards, and discoverability baked in.\ <br>
* Pioneering Ecosystem: The only protocol combining GPU compute + launchpad + governance.

<br>

#### Decentralized Compute & Launchpad Comparison

| Feature            | GPUNET Subnets                                   | Akash Network                  | Render Network               | Bittensor                      | SingularityNET               | Binance Launchpad          | Internet Computer (ICP)             |
| ------------------ | ------------------------------------------------ | ------------------------------ | ---------------------------- | ------------------------------ | ---------------------------- | -------------------------- | ----------------------------------- |
| Core Focus         | GPU-powered projects & decentralized launchpad   | GPU cloud computing            | Decentralized GPU rendering  | AI model training & rewards    | AI services marketplace      | Token fundraising          | Decentralized cloud & web hosting   |
| Compute Access     | GPU pool for AI, bots, tools                     | GPU rental for general compute | GPU for rendering            | GPU for AI training            | AI processing only           | None                       | Cloud compute for dApps             |
| Launchpad Features | Ready-made templates for bots, tools, meme coins | No project launchpad           | No launchpad                 | No launchpad                   | No launchpad                 | Token sales only           | Supports dApp & canister deployment |
| Governance         | Community voting via Queens (validators)         | No governance voting           | No governance voting         | AI-driven via TAO token        | Marketplace-based governance | Binance-controlled         | DAO-like canister governance        |
| Monetization       | $GPU rewards for creators & validators           | Pay-per-use GPU                | Pay-per-use GPU              | TAO staking & usage rewards    | AGIX staking & usage rewards | Fees from token sales      | Pay-per-use compute model           |
| Staking & Rewards  | Stake $GPU to boost, vote, earn                  | Stake AKT for lower usage fees | No staking rewards           | Stake TAO to earn rewards      | Stake AGIX for benefits      | No staking                 | Stake ICP for governance & rewards  |
| Decentralization   | Fully decentralized & on-chain voting            | Decentralized GPU providers    | Decentralized GPU providers  | AI-governed network            | Decentralized AI marketplace | Centralized (Binance)      | Decentralized with central nodes    |
| AI & Web3 Tools    | Bots, automation, GPU services                   | General cloud compute          | Rendering-focused            | AI training & model evaluation | Chatbots, inference APIs     | No AI tools                | AI tools + Web3 integration         |
| Scalability        | 1 million+ subnets target                        | More providers = more scale    | Limited to render workloads  | Scales with AI adoption        | Scales with AI demand        | Limited to project funding | Web3 & canister scaling             |
| Smart Contracts    | GANChain native smart contracts                  | Cosmos-based contracts         | No smart contracts           | AI smart contract governance   | Ethereum-based contracts     | BSC-based tokens           | Native canister smart contracts     |
| Target Users       | AI devs, builders, communities, Web3 startups    | Devs needing GPU               | Creators & studios           | AI researchers, node runners   | AI-focused developers        | Crypto startups            | dApp developers & Web3 teams        |
| Market Demand      | AI, Web3, gaming, compute                        | DeFi, GPU compute              | 3D graphics, video rendering | AI, ML, distributed compute    | AI tooling                   | Token trading & IDOs       | Decentralized web & services        |

\
\ <br>

#### Potential $GPU Earnings Timeline

The $GPU network operates on a structured emission model, with earnings scaling as adoption grows. Below is the expected daily issuance over time:

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXf0nbllM3FibVOks5fxtJ2ULA6pJuLjynUJCcLK9_K-776_8SGZdPNnwYaRL-8IdgyOiVRsccqwV0XvkdeAN8nMd2OlXCtOdEzL6F5Cr4IZDEL5TzC5DWVg-IPRzlsJoxPx0Fmtew?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

\ <br>

Testnet Phase (May 2025): 480 $GPU/day

Mainnet Launch (Estimated June 2025): 1440 $GPU/day

First Doubling (Estimated August 2025): 2880 $GPU/day

Second Doubling (Estimated December 2025): 5760 $GPU/day

<br>

These values represent the estimated distribution of $GPU tokens per day, providing an incentive structure for early adopters and active contributors.

<br>

#### Projected $GPU Value Scenarios

The price of $GPU is expected to fluctuate based on market demand and network growth. For estimation purposes, here are earnings projections based on potential price points:

$2 per GPU

$3 per GPU

$4 per GPU

$5 per GPU

<br>

#### Subnet Earnings Potential

As the ecosystem grows, serious subnet operators will play a key role in securing and expanding the network. Below is the estimated average earning per subnet based on projected emission rates:

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdweZXwV6T1QziF-ypM_LtagRHI-rtvIMVYwIbCg6L_Rogdyn50puGTe5cz3ctQ0DoZQu6urpsNi2tElWLc1OiK-2dtgLnZnJHtnbILBdQisbg0NBxYQun-1CSRGkfMdGolpuBjoA?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

<br>

Quarter    Estimated Earnings Per Subnet (USD)

Q1 (March)     $576 / month

Q2 (April)    $1,296 / month

Q3 (August)    $2,304 / month

Q4 (December)   $4,320 / month

<br>

#### Top 1% Earnings

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfJuQFPQ9hlLIF7QUdPBoH3ktujNxhSgRltZTSXvvmHmARI5lznNJDL8s3_wKDCWEOu_YDni_0qzdsM9GlpyP3Sy8yg8W7nhfyJam_p9sDIGthETKcxw0N1FqNT_o5tNTByAl7jng?key=TGF2m4SHFILOJuWaL3Ewxw" alt=""><figcaption></figcaption></figure>

<br>

Being in the top 1% of subnets can significantly boost earnings, potentially increasing rewards by 2x to 10x. Earnings are influenced by:

<br>

* Reputation & Performance – High-performing subnets receive more votes.
* Community Support – The more votes you get from Queens, the higher your potential earnings.

<br>

### Future of Subnets

At GPUNET, we envision a dynamic, scalable, and AI-driven future where launching and running decentralized businesses becomes seamless and accessible to all. We are building powerful tools and smart features to transform every Subnet into a fully functional, self-sustaining micro-economy. Here’s what’s coming:

#### AI-Powered Startup Creation

Creators will be able to launch entire projects with the help of AI. From ideation to execution, AI agents will guide subnet creators in setting up their businesses—accelerating development and reducing time to market.

#### Text-to-Website Generation

Easily create Subnet pages by simply describing your project in natural language. Our AI will instantly convert your description into a professional, on-chain, fully editable website that showcases your Subnet's purpose, proof of work, and features.

#### On-Chain Product Hosting

Subnets will be able to run their tools, bots, and AI agents directly on decentralized GPU infrastructure. This enables real-time, reliable access to AI-powered products, right from the Subnet page.

#### In-Subnet Product Listings

Each Subnet will support native product listings, allowing Kings to showcase multiple offerings—from AI services to digital tools and decentralized apps—all discoverable and accessible to users.

#### Monetization & Payments

Users will be able to interact with and purchase Subnet products directly via the Subnet page, using $GPU and other supported tokens. This brings real, on-chain revenue to Subnet creators and promotes network utility.

#### Community Feedback & Reviews

Users can leave ratings and feedback for each Subnet. This transparent feedback system helps evaluate a Subnet’s ongoing Proof of Work and drives credibility and trust within the community.

#### Reward Distribution Customization

Kings will soon have the ability to allocate a percentage of their Subnet's block rewards to their voters (Queens). This creates a collaborative ecosystem where voters are further incentivized to support high-quality Subnets.

#### Sustainable Contribution Model

To sustain long-term network growth, Kings will be encouraged to pledge back a portion of their Subnet rewards to the GPUNET ecosystem—supporting infrastructure, future rewards, and protocol development.

***

These upcoming features will turn GPUNET Subnets into the easiest and most powerful launchpad for decentralized innovation—bridging the gap between creators and users, while ensuring transparency, reward fairness, and ecosystem sustainability.

<br>


# Astra

Astra - AI Generation Studio :

Astra is a next-generation AI creation studio designed for creators, developers, and everyday users who want fast, powerful, and intuitive AI tools.

With Astra, you can:

* Chat with intelligent AI models (public quick chat & advanced logged-in chat)
* Generate high-quality images using Bytedance’s cutting-edge SeeDream model
* Create cinematic videos using the powerful SeeDance video-generation engine
* Access fine-tuned controls, enhanced prompts, and creator-focused tools in Astra Video Generation 2.0

Astra brings simplicity to beginners and deep creative control to professionals—all in one seamless experience.

## 1. Astra Chat

Astra Chat provides two modes of text interaction depending on user needs:

#### 1.1 Public Chat (No Login Required)

Designed for fast, simple, and friction-free AI interaction.

Key Features:

* Instantly available,no login needed
* Ideal for quick questions, fast insights, and short responses
* Lightweight, efficient model optimised for speed
* Zero personal data required
* Great for on-the-go lookup or short brainstorming

Use Public Chat when you need immediate answers without commitment.

***

#### 1.2 Advanced Chat (Logged-In Experience)

A deeper, more capable conversational model for users who want richer, more contextual responses.

Key Features:

* High-depth reasoning and long-context understanding
* Can consume user-provided information for personalised, accurate replies
* Delivers detailed, structured, research-grade insights
* Ideal for work, learning, creative writing, coding, planning, and professional use
* Conversation history and saved chats
* More powerful model with improved accuracy

Advanced Chat is Astra’s premium conversational layer built for users who need the best possible answers.

***

## 2. Image Generation -  Powered by SeeDream

Astra integrates SeeDream, Bytedance’s advanced image generation model known for its realism, stylistic range, and creative consistency.

#### Key Features:

* High-resolution image generation
* Multiple aspect ratio options
* Supports anime, hyper-real, cinematic, artistic, and photorealistic styles
* Fine-tuned control over lighting, texture, colors, and composition
* Prompt enhancement for detailed results
* Fast generation engine with high-quality outputs
* Safe and intelligent content filtering

Whether you're creating concept art, product visuals, game assets, or artwork, Astra Image Generation gives you top-tier results with minimal effort.

***

## 3. Video Generation — Powered by SeeDance

Astra uses SeeDance, Bytedance’s advanced AI video model, enabling users to create smooth, cinematic AI-powered videos.

#### 3.1 Base Features

* Aspect ratio control (1:1, 16:9, 9:16, and more)
* Duration customisation for short or extended videos
* Fixed camera options for stable shots
* Resolution controls
* Motion intensity settings
* High-quality text-to-video and image-to-video generation<br>

Astra Video is ideal for creators, animators, advertisers, educators, and filmmakers looking for fast, powerful AI-powered video generation.

***

## 4. Astra Video Generation 2.0 — Creator Mode

A major upgrade focused entirely on creator control, flexibility, and production-ready outputs.

#### 4.1 Enhanced Prompting

* Built-in Prompt Enhancer automatically improves user prompts for richer detail
* Smarter understanding of motion, environment, and style cues
* Scene-aware enhancement for consistency in color, lighting, and camera flow

***

#### 4.2 Finer Creator Controls

Astra 2.0 introduces granular creative tools:

* Motion control: Adjust movement intensity, direction, and style
* Camera controls: Choose angles, tracking shots, zoom behavior, and more
* Scene motion elements: Wind, water, particles, crowd movement, etc.
* First-frame & last-frame injection:
* Upload your starting frame for consistency
* Add a final frame to control the ending scene or match branding
* Frame-by-frame consistency engine for smoother transitions<br>

These tools give creators more influence over the story, look, and motion of the video.

***

#### 4.3 Competitive, Dynamic Pricing

Pricing adapts based on:

* Video length
* Resolution
* Enhancements used<br>

This ensures that users only pay precisely for what they need—no hidden or flat-rate overcharges.

***

## Astra: The Future of AI Creation

Astra brings the power of advanced AI models into a clean, creator-first environment where:

* Beginners can get instant results
* Professionals can build with precision
* Creators can control every detail<br>

Whether you're chatting, generating images, or producing full cinematic videos Astra makes creation effortless and powerful.


# GVEX


# What is GVEX?

GVEX (GPUNET Verifiable Exchange) is a native decentralized exchange built directly on GANChain L1, designed to power verifiable trading within the GPUNET ecosystem. Unlike traditional DEXs, GVEX functions as an infrastructure-level marketplace that bridges computational assets with tokenized value.

#### Design Principles

GVEX is built on three core pillars:

**Security**: All transactions execute on-chain through verifiable smart contracts. Users maintain full custody and control over their assets throughout the trading process.

**Transparency**: Every trade, Node exchange, and liquidity movement is visible and auditable in real-time on the blockchain.

**Simplicity**: The platform serves both blockchain newcomers and experienced traders with a frictionless interface that doesn't compromise on verifiability or security.

#### Key Features

* Native GANChain L1 integration with no intermediary dependencies
* Full orderbook functionality with market and limit orders
* NFT trading (Nodes) alongside token pairs
* Instant settlement upon order matching
* No hidden fees or staking requirements
* Open access for all users, not just Node holders

#### Current Market Stats

You can find them [here](https://www.gpu.net/gvex)


# How It Works?

#### Architecture

GVEX employs a hybrid architecture that combines the speed of centralized exchanges with the security of on-chain settlement:

**Layer 1**: GANChain L1 blockchain serves as the settlement layer.

**Matching Engine**: Off-chain orderbook matching engine processes orders with sub-second latency.

**Smart Contracts**: Solidity-based contracts handle on-chain settlement and custody. All contracts are deployed natively on GANChain L1.

This design allows GVEX to achieve high-performance order matching while maintaining complete on-chain verifiability for all settled trades.

#### Order Types & Matching

GVEX supports two order types:

**Limit Orders**: Specify your exact price and size. Orders are added to the orderbook at price levels that are integer multiples of the tick size, and sizes that are integer multiples of the lot size.

**Market Orders**: Execute immediately at the best available price in the orderbook.

**Matching Algorithm**: The orderbook operates on **price-time priority**, identical to centralized exchanges but fully verifiable on-chain. Orders at better prices are matched first. Among orders at the same price, earlier orders have priority.

#### Trading Specifications

**Tick Sizes (Price Increments)**:

* 0.01 for high-precision pairs
* 0.1 for standard pairs
* 1.0 for larger denomination pairs

**Minimum Order Size**: 0.1 for all assets

**Lot Sizes**: Orders must be placed in integer multiples of the lot size, which varies by trading pair.

**Settlement**: Trades settle instantly when orders are matched. The off-chain engine matches orders, and settlement occurs immediately on GANChain L1 through smart contract execution.

***


# Start Trading

#### Wallet Setup

GVEX supports MetaMask and all EVM-compatible wallets.

**Add GANChain Network to Your Wallet**:

1. Open MetaMask (or your EVM wallet)
2. Navigate to Networks → Add Network
3. Enter the following details:
   * **Network Name**: GANChain
   * **RPC URL**: <https://rpc.gpu.net/>
   * **Chain ID**: 4048
   * **Currency Symbol**: GPU
   * **Block Explorer**: <https://www.gpu.net/ganchain>
4. Save and switch to GANChain network

#### Connect & Verify

1. Navigate to <https://www.gpu.net/gvex>
2. Click "Connect Wallet"
3. Select your wallet provider (MetaMask recommended)
4. Approve the connection request in your wallet
5. Complete on-chain verification (one-time process)

You're now ready to trade.

#### Place Orders

**Understanding the Interface**:

The GVEX trading interface displays:

* **Orderbook**: Real-time list of all buy (bids) and sell (asks) orders with their prices and quantities
* **Order Entry Panel**: Input fields for placing your orders
* **Order History**: Your active and completed orders
* **Trade History**: Recent trades executed on the pair

**Placing a Limit Order**:

1. Select your trading pair (GPU/USDT or Nodes/USDT)
2. Choose "Limit" order type
3. Enter your desired price (must be a multiple of the tick size)
4. Enter your order size (There's no minimum order for $GPU. And for GAN Node it is 1,order needs to be in whole numbers)
5. Select "Buy" or "Sell"
6. Review the order details
7. Click "Place Order"
8. Confirm the transaction in your wallet

Your limit order will be added to the orderbook and will execute when a matching order is available at your specified price or better.

**Placing a Market Order**:

1. Select your trading pair
2. Choose "Market" order type
3. Enter your order size
4. Select "Buy" or "Sell"
5. Review the estimated execution price (based on current orderbook)
6. Click "Place Order"
7. Confirm the transaction in your wallet

Market orders execute immediately against the best available prices in the orderbook. Note that large market orders may experience slippage if they consume multiple price levels.

#### Manage Orders

**Canceling Open Orders**:

1. Navigate to "Orders" section
2. Locate the order you want to cancel
3. Click "Cancel"

**Monitoring Your Trades**:

* **Open Orders**: View all your pending limit orders
* **Order History**: Track completed, canceled, and rejected orders
* **Trade History**: See all your executed trades with timestamps and prices
* **Portfolio**: Monitor your token balances and total value

***


# Trading Pairs

#### GPU/USDT

The primary trading pair for the GPUNET ecosystem. $GPU is the native utility token that powers all interactions on GPUNet, including:

* Paying for compute resources (AI inference, rendering, blockchain computations)
* Staking for network rewards
* Governance participation
* Subnet operations

Trading GPU/USDT allows participants to enter and exit the GPUNET ecosystem with stablecoin liquidity.

#### Nodes/USDT

Provider Nodes are NFT assets that represent operational infrastructure within the GPUNET network. Node holders can:

* Operate as validators or GPU providers
* Earn token rewards for contributing compute capacity
* Build stackable node operations businesses
* Participate in network governance

**NFT Trading Mechanics**:

Nodes are ERC-721 compliant NFTs that trade against USDT on GVEX.&#x20;

* Hardware specifications (GPU models, compute capacity)
* Historical earnings
* Operational status
* Network reputation

When trading Nodes, you're exchanging these operational assets directly on-chain. Settlement occurs instantly, and the NFT transfers to the buyer's wallet upon trade execution.

***


# Fees & Settlement

#### Trading Fees

GVEX operates with transparent fee structures:

* **Maker Fees**: 0.7 %
* **Taker Fees**: 0.7%&#x20;

#### Network Gas Costs

Gas fees (paid in $GPU) are only required for:

* **Deposits**: Moving assets into GVEX
* **Withdrawals**: Moving assets out of GVEX

All trading operations are handled off-chain:

* Order placement
* Order cancellation
* Order matching
* Trade execution

This hybrid model keeps trading fast and gas-efficient. You only pay gas when moving assets on/off the platform.

#### Settlement

Trades settle instantly when orders are matched. The process is:

1. Off-chain matching engine identifies compatible orders
2. Trade executes and your GVEX balance updates instantly
3. Updated balances are immediately visible in the GVEX interface
4. To access your funds on-chain, initiate a withdrawal
5. Withdrawal requires gas fee and transfers assets to your wallet

**Important**: Matched trades update your GVEX balance immediately, but funds remain in the exchange's smart contract until you withdraw them to your wallet.


# FAQ

### FAQ

**Q: Do I need to hold Nodes to trade on GVEX?**\
A: No. GVEX is open to all users. You only need a compatible wallet and $GPU tokens for gas.

**Q: What wallets are supported?**\
A: MetaMask and all EVM-compatible wallets work with GVEX. Ensure you've added the GANChain network (RPC: <https://rpc.gpu.net/>, Chain ID: 4048).

**Q: How do I get $GPU tokens?**\
A: $GPU is available on multiple chains including Ethereum, Base, and natively on GANChain. Bridge tokens to GANChain to start trading on GVEX.

**Q: Are my funds safe on GVEX?**\
A: Yes. GVEX uses a self-custody model where you maintain full control of your assets. All trades execute through auditable smart contracts on GANChain L1.

**Q: What happens if my order doesn't fill?**\
A: Limit orders remain in the orderbook until they're either matched, canceled by you. You can cancel unfilled orders at any time.

**Q: Can I trade Nodes if I don't own any?**\
A: Yes. You can buy Nodes/USDT on GVEX by submitting a buy order. Once executed, the Node NFT transfers to your wallet.

**Q: What's the difference between GANChain and other L1s?**\
A: GANChain is specifically designed for GPU compute verification and economic coordination. It's a hybrid PoW/PoS chain optimized for the GPUNET ecosystem.

**Q: Are there any geographical restrictions?**\
A: GVEX is permissionless and decentralized. However, users are responsible for complying with their local regulations regarding cryptocurrency trading.

**Q: How do I add liquidity to GVEX?**\
A: Liquidity is provided organically by traders placing limit orders. Simply place buy or sell limit orders at your desired prices to add liquidity to the orderbook.

**Q: What pairs are coming next?**\
A: Additional trading pairs are planned as the ecosystem grows. Announcements will be made through official GPUNET channels.

**Q: Where can I get support?**\
A: Join the GPUNET community on Discord and Telegram for support, or visit the official documentation at <https://docs.gpu.net>

**Q: Is GVEX audited?**\
A: The smart contracts are upgradeable and follow industry-standard security practices. Check official channels for audit reports and security disclosures.

**Q: Can I use GVEX programmatically?**\
A: Yes. Developers can interact with GVEX smart contracts directly or build trading bots.


# Roadmap

GVEX is being rolled out in phases to ensure stability, usability, and scalability:

**Current Phase**: Live deployment with GPU/USDT and Nodes/USDT pairs. Focus on user feedback, trading performance testing, and liquidity depth optimization.

**Upcoming Features**:

* Additional trading pairs as ecosystem grows
* Mobile-optimized interface

***

### Additional Resources

* **GVEX Platform**: <https://www.gpu.net/gvex>
* **GANChain Documentation**: <https://docs.gpu.net>
* **GANChain RPC**: <https://rpc.gpu.net/>
* **Community Discord**: <https://discord.gg/fAbKmNBApC>
* **Telegram**: <https://t.me/gpunetwork>
* **Twitter**: <https://x.com/gpunet>&#x20;

***

*This documentation is maintained by the GPUNET . For the latest updates, visit the official channels. Trading cryptocurrencies involves risk.*&#x20;


# Barrels of Compute Guide

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXeYcjH0DU3CTOF7iJz_pSoZmuMnWXf4BebST_Uq79AjAWQddl8DVLjnuji0glTO9NCBB3RR0ytWehQDxM7MjvUOu9FSwiHHbg2hOnUWoi1o2eNkatDT_p9LZgWZZzC4_qwcy_GxYA?key=BSWvsOmsUPP3_zcvjG2T2zU8" alt=""><figcaption><p>barrel of compute</p></figcaption></figure>

**Introduction: Barrel of Agentic Compute**\
GPUnet introduces the “Barrel of Agentic Compute” as the next step in decentralized compute infrastructure. The concept revolves around converting idle GPU capacity into a scalable, tradeable, and accessible resource for individuals and enterprises. Through Barrels of Agentic Compute, participants gain access to affordable compute power, scalable earning potential, and the ability to contribute to an ecosystem designed for AI, Blockchain, and Web3 innovation.

**Key Features of Barrel of Agentic Compute**\
**Core Offerings:**

1. **Compute Accessibility:**

* &#x20;Each Barrel of Agentic Compute provides $500 worth of compute usable on [dapp.gpu.net](https://dapp.gpu.net/).

2. **Tradable Compute:**

* &#x20;Compute can be traded as NFTs, starting on Solana, with expansion to additional blockchains soon.

3. **Upgradeable Agentic Compute System:**

* &#x20;Upgrade to Queen: Exchange 33 Barrels of Agentic Compute for a Genesis Queen Node.
* &#x20;Upgrade to Validator: Exchange 66 Barrels of Agentic Compute for 6 Genesis Validator Nodes.
* &#x20;Upgrade to King Subnet: Exchange 99 Barrels of Agentic Compute to launch a King Agentic Subnet.

4. **Scalable Rewards:**

* &#x20;Validators earn $6,666/year in compute rewards and an $8,000 airdrop at TGE.
* &#x20;Queens earn $10,000/year with an APR exceeding 200%.
* &#x20;Subnet operators (Kings) earn 120 $GPU/day, with Queens voting on block rewards based on subnet performance.

5. **Enhanced Ecosystem:**

* &#x20;Built as a virtualization layer integrating over 100+ GPUs like Nvidia H100, 3090s, and 4090s, sourced from leading datacenters.

**GANchain: Key Features**

1. **Hybrid Consensus:**

* &#x20;Combines Proof of Stake (PoS) and Proof of Work (PoW) to ensure scalability and security.

2. **Validator Network:**

* &#x20;Over 250+ Validators currently earning 480 GP/day, with average annual returns of $500. ([View  dashboard](https://dashboard.gpu.net)).
* &#x20;Queens actively vote to set block rewards based on subnet performance.
* &#x20;Scalable Agentic Compute Ecosystem: Validator, Queen, and King Subnet enhance scalability  while providing incentives.

3. **Subnets:**

* &#x20;With the upcoming Nova Upgrade, King Subnets will enable custom compute ecosystems for enterprises and communities.

4. **Efficiency & Speed:**

* &#x20;Offers 6-second block confirmation times and parallel block production.

5. **Real-World Integration:**

* &#x20;Backed by datacenter partners: Northern Data (Germany), NTT Data (India), and G42 (UAE).\
  \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_<br>

Barrels of Agentic Compute Pricing Structure

🔸For Individuals:

&#x20;      \*  $500 Compute Credits per barrel, usable on [dapp.gpu.net](https://dapp.gpu.net)

&#x20;      \*  Ability to trade compute as NFT's, providing liquidity and flexibility

🔸For Queens:

&#x20;     \*  Upgrade with 33 Barrels of Agentic Compute.

&#x20;     \*  Swap compute for $5,000 worth of $GPU.

&#x20;     \*  Earn $10,000/year with 200%+ APR

🔸For Validators:

&#x20;     \*  Upgrade with 66 Barrels of Agentic Compute.

&#x20;     \*  Earn $6,666/year in rewards.

&#x20;     \*  $80,000 airdrop at TGE

&#x20;&#x20;

🔸For King Subnet:

&#x20;     \*  Upgrade with 99 Barrels of Agentic Compute.

&#x20;     \*  Swap compute for 25,000 $GPU.

&#x20;     \*  Earn 120 $GPU/day and govern subnet performance through Queen voting.

**DApp: The Gateway to Compute**\
**Overview:**

1. **Virtualization Layer:**

* &#x20;DApp integrates GPU resources from 4 major datacenters globally, creating a Real World Asset (RWA).
* &#x20;Including: Northern Data (Germany), NTT Data Centers (India), G42 Khazna/Core42 (UAE).

2. **Compute Availability:**

* &#x20;Over 100+ GPUs, including Nvidia H100s, 3090s, and 4090s, are available for rent at [dapp.gpu.net](https://dapp.gpu.net).

3. **Partnerships:**

* &#x20;Northern Data (Germany): Cutting-edge datacenter solutions.
* &#x20;NTT Data (India): A $90B Japanese conglomerate with advanced infrastructure.
* &#x20;G42 (UAE): Leaders in AI and high-performance computing.

**Usability:**

* &#x20;Users can exchange their compute credits directly on the DApp, scaling projects without the financial burden of traditional compute services like AWS.

**ROI Estimation for Barrels of Agentic Compute**<br>

**Significant Achievements**

**Financial Milestones:**\
\- Over $10M in revenue with $500K distributed to node holders.\
\- Tokenomics designed to support long-term sustainability.

1. **Partnerships:**\
   \- Collaborated with Northern Data, NTT Data, and G42 for data center integration.\
   \- Secured high-profile backers and KOLs, including Crypto Banter and Virtual Bacon.
2. **Ecosystem Growth:**\
   \- 5328 Total Nodes as of January 2025.\
   \- Validator count surpassing 250 with a strong staking community.

**Key Supply Details**

* **Barrels of Agentic Compute Total Supply:** 14,400 nodes.
* **Barrels of Agentic Compute:** 28th Jan 2025.
* **Pricing:** Linear tier-based pricing for affordability.

**Unique Features of the Ecosystem**

1. **Barrel of Compute:** A scalable, decentralized framework enabling the trade and utilization of compute resources.
2. **Hybrid Consensus:** PoW and PoS combined for efficiency and security.
3. **Dynamic Rewards:** Tailored earning structures for Validators, Queens, and Subnet operators.
4. **Tradability:** Compute as NFTs adds liquidity and flexibility for participants.
5. **Sustainability:** Efficient GPU utilization with minimal environmental impact.

**Upcoming Milestones (2025)**

1. Barrels of Agentic Compute Public Launch: 28th Jan 2025
2. Nova Upgrade: Introducing King Subnets with enhanced scalability.
3. Token Generation Event (TGE): Scheduled for Q1 2025.
4. Global Partnerships: Expanding collaborations with enterprises and Web3 innovators.


# Barrel of Compute FAQ

**Agentic Compute - Frequently Asked Questions (FAQ)**

**1.** **What is the Node Sale about?**

The Node Sale offers an opportunity to own a decentralised GPU Compute Node that powers AI agents and advanced computations on the Solana Blockchain. Solana provides unmatched scalability, speed, and low transaction costs, making it the ideal blockchain for AI-driven decentralised applications.

**2.** **Why are we using Solana?**

Solana was chosen for its exceptional transaction speed, scalability, and low fees, which are crucial for decentralised applications requiring substantial computational power. Its platform enables seamless AI workload interactions and ensures low-cost task completion.

**3.** **What are AI agents, and how do they work?**

AI agents are autonomous programs that perform tasks such as data processing, decision-making, content creation, and more. They leverage the decentralised compute power provided by Agentic Nodes to complete these tasks efficiently and securely.

**4.** **Who can participate in the Node Sale?**

The Node Sale is open to individuals, enterprises, and developers who want to contribute to the decentralised computing economy. There are no stringent entry requirements, making it accessible to anyone passionate about advancing AI and decentralised technologies.

**5.** **How does GPU Compute power benefit AI development?**

Compute credits represent the computing resources available to each Agentic Node. These credits can be used on **dApp.gpu.net** to run AI models, simulations, and other computationally intensive tasks. They are also tradable as NFTs, allowing holders to sell or exchange their compute power.

**6.** **How is a Barrel of Compute different from $GPU or GPoints?**

The token's core utility is compute, which it can always be exchanged for, but the token also accrues value. A Barrel NFT stays pegged to $500 and offers additional benefits, including conversion into Validators, Queens, and King Nodes.

**7.** **What are the benefits of this NFT?**

**Individuals:** Access $500 worth of compute credits (tradable NFTs) and participate in the decentralised compute economy.\
**Validators (11+ nodes):** Earn $2,000 worth of compute credits (including a 25% bonus) and Genesis Node rewards ($500/year).\
**Queens (33+ nodes):** Swap compute for 10,000 $GPU tokens and stake to earn $10,000/year in rewards.\
**King Subnets (66+ nodes):** Launch a custom subnet on GANchain, swap compute for 25,000 $GPU tokens, and stake to earn block rewards (0–480 $GPU/day).

**8.** **How can I earn with Agentic Compute Nodes?**

There are multiple ways to earn:\
1\. **Staking** compute credits for rewards.\
2\. **Running validator nodes** and earning bonus compute and Genesis Node rewards.\
3\. **Becoming a Queen** and staking $GPU tokens for additional rewards.\
4\. **Participating in King Subnets** to earn block rewards.

**9.** **Do we need to run this Node for earnings?**

No, you don't need to run the node. Simply buying and holding the NFT provides the specified benefits.

**10.** **How do I use the compute provided in the NFT?**

You can claim compute from [**dapp.gpu.net**](https://dapp.gpu.net/). Once claimed, your NFT will burn, and $500 worth of compute credits will be transferred to your account.

**11.** **What makes the Agentic Compute ecosystem unique?**

1\. **Accessibility:** Low-cost entry with tradable compute credits.\
2\. **Scalability:** The Nova Upgrade enables subnets to scale compute resources.\
3\. **Governance:** Queens vote on block rewards for fair and efficient resource allocation.\
4\. **Sustainability:** A hybrid consensus model combines Proof of Work and Proof of Stake for energy-efficient validation.\
5\. **Revenue Potential:** Multiple earning opportunities through staking, running nodes, and subnets.

**12.** **When will Agentic Compute Nodes launch?**

The public launch is scheduled for **January 28, 2025**.

**13.** **What are the upcoming milestones for the Agentic Compute ecosystem?**

* **Nova Upgrade:** Introduction of subnets for scaling.
* **Expanded Tradability:** Compute credits will be tradable across multiple chains.
* **Enhanced Dashboards:** Improved tools for node management and compute usage.
* **AI & Web3 Integration:** Deeper integration with AI, Web3, DeFi, gaming, and Decentralized Science (DeSci).


# $GPU Token Launch

The launch of the **$GPU token** marks a major milestone in the evolution of GPU.net, a decentralized compute platform designed for performance, fairness, and true community ownership. Built to empower developers, builders, and contributors, $GPU is more than just a utility token; it's the backbone of a decentralized ecosystem that prioritises long-term value creation and sustainable growth.

With a total fixed supply of **200 million tokens**, $GPU is engineered for **long-term value creation, community driven growth**, and **sustainable tokenomics**. Here's a closer look at how $GPU is redefining token distribution and network economics.

## &#x20;**Token Allocation:**

* **Community Allocation – 80%**

The majority of the supply is reserved for the GPU.net community, including developers, node operators, and Subnet creators. There are no allocations for team, foundation, advisors, or marketing, ensuring maximal decentralization and fair distribution.

* **Investor Allocation – 10%**

A modest 10% is allocated to early investors to support platform development without centralizing power.

* **Liquidity Reserve – 10%**

This portion ensures initial and ongoing liquidity across trading platforms and decentralized exchanges.

## **Deflationary Model:**

$GPU follows a carefully designed deflationary issuance mechanism to preserve long-term value. The block rate is doubled till 50% of total liquidity and then infinitely halved. Also, all the gas fees earned is burnt to create a deflationary pressure.

## **Subnet Ownership & Rewards:**

A cornerstone of GPU.net’s architecture is Subnets that serve as isolated, community hosted environments designed to run specialised services or workloads securely and independently. Every new Subnet begins with a block rate of **1,440 $GPU per day**.

The actual $GPU share earned by a Subnet is governed by community voting, which takes into account:

* Subnet visibility
* Product quality
* Revenue generation and traction

This model ensures that those who build, maintain, and grow valuable infrastructure are proportionally rewarded in a trusted and transparent way.

## **Token Utility:**

**Own Compute. Own $GPU.**

* Use your $GPU tokens to rent real-world GPU's directly on [**the dapp**](https://dapp.gpu.net/marketplace) — our decentralised GPU rental platform.
* Staking: Stake $GPU to become a **Queen** — vote on subnet's governance, or a **Validator** — secure the network and earn rewards — both roles come with rewards and slashing risks.
* PoW voting: $GPU grants you voting power, allowing you to support real Proof-of-Work subnets. Your votes help these subnets qualify for rewards and earn further.

<br>

\ <br>


# Tokenomics

Explore the full economic model and distribution strategy for $GPU, our native token designed to fuel the decentralized GPU marketplace:

[**View Full $GPU Tokenomics**](https://docs.google.com/spreadsheets/d/1vEGDlJ1I-wQv2CxquyTSg54H43ZVmAOGz793s1-zXsc/edit)

This comprehensive document includes token distribution, emission schedules, vesting timelines, and our innovative doubling-to-halving reward mechanism designed to support sustainable network growth.


# Quests

## Your Gateway to Web3 Compute Rewards

GPU.Net is pioneering a decentralized future where AI compute power becomes universally accessible. At the heart of this vision lies the GPU Quest Campaign, a dynamic gamified journey that invites users to engage, contribute, and earn. Whether you're a developer, validator, subnet enthusiast, or simply Web3-curious, the Quest system is designed to immerse you in the ecosystem while building your stake in it.

## Inside the TGE: Unlocking the Launchpad of the $GPU Economy

The Token Generation Event (TGE) isn't just a launch, it's a culmination of community action. As GPU.Net transitions to mainnet, $GPU tokens will be distributed based on contributions tracked through GXP. TGE marks the beginning of decentralized ownership, staking capabilities, and governance. Completing quests not only builds your score, it builds your influence in the network.

## The Campaign Trail: GPU Quest Adventure

GPU Quests are structured like a multi-level game, blending technical tasks with community action. From setting up validator nodes to engaging on social media, each quest earns GXP (GPU Experience Points), a measure of your active contribution. These quests also act as onboarding missions into GPU.Net’s ecosystem featuring real impact, real proof-of-work, and real rewards.

## Quest Categories: Choose Your Path

1. #### Onchain Quest (Onchain): Dive into blockchain mechanics and interact directly with onchain protocols.
2. #### Dev Quest (Developer): A playground for builders! Showcase your coding skills and earn rewards.
3. #### Alliance Quest (Alliance): A loyalty-based challenge. To qualify, hold U2U Tokens, run a Solaris Node, and mint a Monad NFT.
4. #### Solana Quest (Blockchain-Specific): Unlock this quest with SOL and earn up to 2000 GXP through Solana-based missions.
5. #### Dapp Quest (dApp):Explore decentralized applications and complete tasks across the dApp ecosystem.
6. #### Side Quest (Bonus): Bonus experiences like Tokyo EVM Drift, with special rewards for wallets aged over 5 months.
7. #### Subnet Quest (Subnet): Focused on subnet-specific interactions.
8. #### Barrel Quest (Collection): Collect barrels and redeem chunk-based GXP rewards in this collection challenge.
9. #### Intro to [GPU.net](http://gpu.net) (Onboarding): New here? Start with this onboarding quest to learn the ropes and kickstart your adventure.
10. #### Social Quest (Engagement): Stay active and earn GXP by engaging across social channels and community hubs.
11. #### Twitter Quest (Social Media): Earn rewards for tweeting, retweeting, and interacting with the GPU.net community on Twitter.
12. #### Exclusive Tasks: The Intro to GPU.net quest welcomes newcomers by guiding them through essential platform features. Tasks include creating a digital wallet, completing your profile, and exploring key sections like the marketplace and leaderboards. You’ll also complete starter missions such as linking wallets and joining your first quest. This onboarding experience rewards early engagement and helps users get comfortable navigating GPU.net. It’s designed to build a strong foundation for ongoing participation and rewards

## Getting Started: Your First Steps into GPU Quests

* Head over to token.gpu.net.
* Connect your cryptocurrency wallet (e.g., MetaMask).
* Use a referral code like "OQVVTF" to get an initial GXP bonus.
* Link your X (Twitter) account to unlock social quests.
* Complete assigned tasks and claim your earned GXP.
* Claim 2 GXP every day with a successful Daily Check-In! But stay sharp, miss two days and your entire streak gets wiped.
* Make sure to connect your cryptocurrency wallet (e.g., MetaMask) before starting quests without a connected wallet, your progress and earned GXP won’t be saved.

## The Referral Loop: Invite, Earn, Multiply

Each participant gets a unique referral code to invite friends, followers, or entire communities. When your referrals complete quests, you both gain bonus GXP. This decentralized growth model ensures exponential reach while rewarding organic expansion and ecosystem loyalty.

Warn: “Referral bonuses are only credited if your wallet is connected and quest completions are properly tracked.”

## Pro Tips to Maximize GXP

* Stay Active: New quests drop frequently don’t miss limited-time tasks
* Show Proof: Submit screenshots, transaction hashes or GitHub links
* Verify Your Role: NodeHolder, Validator, or Queen badges grant multiplier rewards
* Join the Discord: Real-time campaign alerts, voting rounds, and support
* Avoid Automation: Bot behavior or fake completions lead to slashing

## What You Gain: The Power of Participation

Beyond GXP and token airdrops, Quest participants gain deeper access: early subnet integrations, validator/queen roles, and DAO governance eligibility. The more you contribute, the more you earn not just rewards, but recognition and decision-making power within GPU.Net

## Next-Level Adventures

GPU.net is keeping the momentum alive with Upcoming Quests packed with excitement and rewards. Track your progress and rise to the top with the newly introduced GXP Leaderboard and GPU Leaderboard. New adventures include Loot Boxes for surprise rewards, an Intro to GPU.net to onboard newcomers with ease, and powerful GXP Multipliers to supercharge your earnings. Don’t miss the thrilling Side Quest: Tokyo EVM Drift  comment on our X post to earn 10–25 GXP based on your wallet age, with bonus rewards for wallets over 5 months old.

A true highlight of this quest lineup is the Alliance Quests — created for GPU.net’s most committed adventurers. To participate, hold U2U Tokens, run a Solaris Node, and mint a Monad NFT. Complete the trifecta to unlock exclusive benefits and prove your allegiance to the GPU.net alliance.

Don’t forget to connect your wallet and X account to qualify for bonus rewards and leaderboard placement.”

## From GXP to $GPU: Turning Participation into Real-World Value

GXP is the heartbeat of the GPU Quest system. Every interaction big or small is recorded and valued. At the TGE, this score determines how much $GPU you receive. Think of GXP as your contribution ledger: visible, verifiable, and convertible into economic power.

## Subnets Unleashed: Powering the Future of Distributed Compute

Subnets are the engine behind GPU.Net’s decentralized infrastructure. By completing Subnet Quests, users engage in governance, staking, and compute distribution. Subnets reward participants daily for proof-of-work, with added bonuses for earning badges, staying active, and gaining validator votes. It’s not just infrastructure—it’s opportunity on-chain.

## Ready to Rise? Join the GPU Quest Revolution

Your GPU Quest journey starts now. Whether you’re coding a smart contract, minting a barrel, or bringing in your squad, every action counts. Dive into the quest ecosystem at[ gpu.net](http://gpu.net), earn your GXP, and help shape the future of decentralized AI compute.

<br>

<br>

<br>


# Investor Docs

## GPUNET: From Genesis to TGE > Our Journey and What’s Next

GPUNET was founded with a simple goal: make high-performance compute accessible to everyone. In a world dominated by centralized GPU monopolies, we’ve built an infrastructure powered by real hardware, real users, and real incentives. As we approach TGE on June 6th, 2025, here’s a deep dive into what we’ve built, how we’ve grown, and where we’re heading

## 1️⃣ Tokenomics & Supply: Community-First, Scarcity-Driven

GPUNET’s tokenomics are designed for long-term sustainability and user ownership:

* 80% of total token supply is for the community.
* No team, no foundation, no advisory, no marketing allocations.
* 10% allocated to investors.
* 10% reserved for liquidity.

\> Our doubling cycle model ensures a 50% circulating supply cap until 2028, creating a built-in scarcity and deflationary pressure as adoption scales.

\> Subnet Ownership: Every subnet owner earns a direct percentage share of $GPU rewards, making subnet creation a path to long-term value accrual.

\> Target Market Cap: We’re targeting a $16 token price at 50M supply, which translates to an $800M market cap.

GPU.net tokenomics is an economic model built for long-term value, aligning users directly with network growth.

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdABBLmZa27-Ie3xGLWX7QB0GDzcAcCpvCCXkzdxjHD_GaCszpZwWrzQXhTj40ZwD9zGbVXjpwUVVNm2avACaPQrhlIyexeWQAFk2rUdzF4i8xkzR1YvkY8jz8ooz3DeH4_-4MaHw?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcbWlzjvtMb8izxQuJQ4b3h0zLQFd4_KHm5caP1FANVFMXPUqJI8-LK-gBmzCzXY9UICZHNuwp461dx4i-LKTNtY98V-L9zJjFtICrMeAozGc1yEt7oQeH5PSj3dyvxF140lHlCCQ?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfh5viFGSRP54U2pvD3O8p5BG5RUw_SMAzJow4dzry1OHh61202esQRRKLMa6O2fM32omeo0Y4C2DgxMBaN0oeeQeadd2nXMiUNb2Wx521kFtf7VJVhp8nVgZsFgGc9aZRqigok?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

## 2️⃣ Datacenters, Infrastructure & GAN Chain Stats

GPUNET is a fully operational compute network:

* 4 active data centers integrated directly into the dApp.
* 60+ active GPU providers contributing compute capacity.
* 300+ validators securing the GAN Chain L1.
* 50+ active subnets powering AI, gaming, and compute-focused projects.

GAN Chain is the foundation, a custom L1 designed for scalable compute. Validators maintain network integrity, while GPU providers supply real-time hardware capacity.

## Network Stats:

* 6050+ nodes globally
* 300+ public validators
* Provider nodes earn up to 480 $GPU/day
* Queen nodes earn \~200% APR on staked $GPU
* Dapp: 10K+ accounts, 426 Queens, 294 active validators, 66 providers, 190K GP staked.
* Subnets: 50+ AI & compute-focused projects live.

## Product Progress:

* GAN Chain Nova Upgrade: New rewards model (50% to compute providers, 50% to stakers).
* Subnets Beta: 500+ subnets created, 5 no-code templates live.
* Barrel NFTs: Launched on Solana, usable for compute credits, swaps, and dApp redemptions.
* DApp Updates: Public dashboard live with real-time GPU listings, rentals, and user accounts.
* AI Enhancements: 10% faster LLM inference, MLPerf benchmarks, MIG clusters deployed.
* Questboard: 100+ quests live, GXP system, Barrel + Alliance Quests integrated. Questboard 2.0 in development: Deeper integrations, higher engagement.
* AI Agents: Twitter + Subnet AI Agents live, more coming.

## Major Clients & Partnerships

* Enterprise Clients:

  * **SFCompute (US):** Serves Princeton University.
  * **Datambit (UK):** Supports UK Gov.
  * **Nasscom (India):** Supports 500+ startups.
  * **Morphic Films (ex-Polygon founder Jayanti):** Anime AI generator.

* Enterprise clients access bare metal compute directly, outside the dApp.

* 5MW facility acquired in Bangalore, India (NVIDIA H200 GPUs).

* $3M 4090 deployment in Hong Kong for low-latency compute.

* UAE partnership with G42 for GPU deployment + operations.

* Strategic talks with Viettel (Vietnam) for colocation.

* Northern Data partnership (Europe): 25K GPUs access.

* AMD partnership for 2.4x HPC performance over H100.

* **Partners:**

  * NTT, Northern Data, Bytedance, G42.
  * U2U Network → DePIN Alliance.

## 3️⃣ Real Revenue: Proven Monetization, Clear Path to Growth

GPUNET is a business with revenue and profit:

* $10M+ revenue generated to date.
* $500K+ profit distributed to node holders.
* Target: $40M revenue milestone to scale buybacks.

Our Questboard product alone has onboarded over 500K users, proof that we can drive demand and adoption. Our core metric moving forward is ROCE (Return on Capital Employed), maximizing the value generated per dollar spent and ROI (Return on Investment) for every contributor, from node operators to subnet creators.

## 4️⃣ USDT Buyback: A Commitment to Value Accretion

* Plan is to buy back $4M of $GPU as revenue scales.
* 80% of investors token have been bought back, keeping token supply clean and focused. Around $4M in USDT already bought back from investors, 80% early investors bought out = clean cap table.

Next targets:

* $40M revenue → $GPU buybacks = real value accrual.

The buyback program is not just a marketing line, it’s an operational policy. As we scale revenue, we scale buybacks. Every dollar earned is a step closer to sustained value for $GPU holders.

## 5️⃣ Marketing & Community Highlights (Q1 2025)

* 250K+ users onboarded via Questboard.
* 50K+ unique daily active users
* Multiple Spaces & X campaigns on compute, AI, DePIN.
* Deep KOL relationships, secured top-tier KOLs as contributors.
* Pushed major marketing campaigns with limited budget; drove real engagement.
* Pushed major marketing campaigns with limited budget; drove real engagement.\ <br>

  <figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcNnhJefKGrGEwJvTbhSsQMwZ0f2RFk4yUT5SB5huHrZggYf54D1P1Bv_GkqzPyc3dJCQOPoK8TgEhB4zIHD5ehRD7-Dx2BPbRVZIx_8AW1m7_U_4MJ__Ys5CjM2WIO0a2rd7hnyQ?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

  <figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXdc8H8kwWXJeR3xhHPmlLd7ba3tWtbuJymKj5Zvnsj0d4U8aH1sGEqNvcJ_dZCqh9m03G_7l1Tfmsn318CwmURBNtxZN0v6sRJ15L4Ohzqn2KtPMpXraZ1CtwOdL-AveJVGHxWEKQ?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>
* GPU preTGE Holders 40K
* More than 5,000 github verified devs on dapp contributing to this daily footfall.\ <br>

  <figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfCuAQDh11yV4kEk0etijxobDvUPoNtaYoQEFvLn-gswvkmZJUHS9cY0ZTsa3FZT_uLdomBNZd_B9N1XOQ18A9d9mly1H_-LuovrXslSBoB73QTla3POSlDkA6A0CPktgh3vsB_jA?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

## 6️⃣ Vision & Future Plans<br>

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcy16bEWymkAYjNkyFycNZno3eL5HkP3Poe50ydqqev3dwXcjz0Xn0BT5Sb6hPzciAHGBgDVFIQKcCQYT9bDhLVwfcQKITcygQtvcNNa1GRhQpNzgYKo_Zg3B7yDxW-radg6Gk1Bw?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

AI is here, 90% of jobs will vanish.Motorized cars replaced horses; motorized minds (AI) will replace humans. Compute is the fuel for this future, own compute, own $GPU, own the economy.

## Here what's next for GPUNet

🔹 TGE on 6 June – $GPU launches on Ethereum and Uniswap

<figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcWu689XK51QtomJJVZ1HN9h4rz6z_zJOE7su6wsMwUKpWSyLjTugTfFL1HugZWLAMto4mfVwATGwAptizrSMadams883zvLdOiYDOkYBoi5fw91jGi6ylao4x6N5TnMfEtLSIiQQ?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

&#x20;🔹 Cross-chain Roadshow:

* 20 June: Mint $GPU-ETH on Base via Aerodome

* 24 June: Mint $GPU-USDT on Hyperliquid

* 27 June: Mint $GPU-USDT on BSC (Pancakeswap)

* 30 June: Mint $GPU-USDT on Sui (Cetus)

* Upcoming: Mint $GPU-USDC on Solana (Jupiter)

* Upcoming: Mint $GPU-USDC on Solana (Jupiter)\
  \
  \ <br>

  <figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXcJodJ2CfK9txiwd-jxjRUsLBvet2zEQoih9IYFKATvUxyti8xmXxdsO1rH-NT0V7-MXfNEL5GLUTyfCoC_FovgSfv4JYLDZdniSe0_-MLjI91MASginyEdeO3KFAQ4QWKWe07S?key=zrw9dSnbyRc-X9GRiYwgog" alt=""><figcaption></figcaption></figure>

* Post-TGE Roadmap:

  * Questboard Season 2 and GXP > GPU conversion. More rewards, new campaigns.
  * Expand Subnet templates and roll out new features (Replit bots, compute rentals). Global Subnet & Provider growth → scaling globally.
  * GVEX (GAN Chain-based liquidity) → future is verifiable. Orderbook trading DEX.
  * Scale alliances, protocol integrations, and global partnerships.
  * Enhance dApp UI, staking dashboard, and analytics.

* Long-Term Vision:

  * Rollout multi-chain liquidity and compute access. $GPU as the currency of compute.
  * Drive compute demand via AI agents, subnets, and enterprise partners.
  * Scale $GPU buybacks as revenue grows.
  * Expand global data centers and partnerships across regions.

  &#x20; <br>


# EVM & SS58 Transfer

To transfer $GPU between your ss58 address and evm address, GPUnet offers a bridge at [https://subnet.gpu.net](https://subnet.gpu.net/)\
\
Goto the website and click on SS58 Transfer in sidebar.

<figure><img src="/files/nkbwSKBbJOV8eqUFLYjn" alt=""><figcaption></figcaption></figure>


# EVM to SS58

Just enter the amount and a valid ss58 address, and Bingo, your funds have been transferred.

<figure><img src="/files/4xxpDExx8igPf8xbhACg" alt=""><figcaption></figcaption></figure>


# SS58 to EVM

Please follow along to transfer your $GPU from your ss58 account to EVM account.

## Step 1

[Update Metadata](https://ganscan.gpu.net/#/settings/metadata) (only if you've not done it before)

Click on update button and sign the pop-up

<figure><img src="/files/6OtmPip4SmBJy2qag8B0" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/OODK3LbAWGo01JRi1QzU" alt=""><figcaption></figcaption></figure>

## Step 2

Enter the amount in the modal.

It will redirect you to GANscan, our explorer.\
Or just click on the "Open GANscan Transfer" button

<figure><img src="/files/qgDibVq2USLiii6TG1GY" alt=""><figcaption></figcaption></figure>

## Step 3

Click on Submission Tab, as shown in the picture below

<figure><img src="/files/zEkbdqwsJ0jvTNLKTiVF" alt=""><figcaption></figcaption></figure>

## Step 4

Select your account from drop button

<figure><img src="/files/le1Lg85klszOetfWXLMf" alt=""><figcaption></figcaption></figure>

## Step 5

Click on Submit Transaction in the bottom right corner of the screen.

<figure><img src="/files/Iimg2AViBaHuzHr8w5Ib" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/N2kwrT55hqRLnwj3HKk9" alt=""><figcaption></figcaption></figure>

## Step 6

Sign and Submit on your PolkadotJS wallet, there you go! Please wait for a few seconds and the balance would be reflected in your EVM wallet.

<figure><img src="/files/51AHMmHg1ppPhhrzywf0" alt=""><figcaption></figcaption></figure>


# SuperNova Upgrade

{% embed url="<https://drive.google.com/file/d/1CCAwFT-6J5bQAU3q3rx0ay6PP4zHjqCN/view?usp=sharing>" %}


# RWA

## Compute Is Already an Asset Class. Tokenization Decides Who Gets to Own It.

**Wall Street spent three years quietly rebuilding GPUs into collateral. The only question left is whether you're on the cap table or watching from outside.**

***

### The reclassification nobody voted on

On March 31, 2026, CoreWeave closed an $8.5 billion delayed draw term loan and something structural shifted. The facility was rated A3 by Moody's and A (low) by DBRS — the first investment-grade rated financing secured by high-performance compute hardware and a customer contract. Investment grade is the tier pension funds and insurance companies are permitted to buy.

Rewind to August 2023. The same company borrowed against materially the same class of hardware at roughly 15% floating. That is the rate a lender charges when it does not understand what it is holding.

Roughly nine percentage points of spread compression in under three years. That is not a negotiation win. That is a reclassification. Lenders stopped treating GPUs as depreciating electronics and started treating them the way they treat aircraft, power plants, and toll roads: as long-lived, income-producing infrastructure you can underwrite.

The follow-through came fast. In May 2026 CoreWeave closed a $3.1 billion DDTL 5.0 — the first *publicly syndicated* HPC-infrastructure-backed facility, rated Ba2 / BB+, oversubscribed, priced 50bps tighter during syndication. Public syndication means secondary trading. Secondary trading means a market.

JPMorgan's CMBS research team told Bloomberg in early February 2026 that data-center securitization could hit $30–40 billion annually in 2026 and 2027, up from about $27 billion in 2025. Oracle, Meta, xAI and CoreWeave have collectively moved roughly $120 billion of AI infrastructure debt into SPVs funded by Wall Street.

Independent analyst Dave Friedman put it best: CoreWeave is a leveraged infrastructure vehicle that finances GPUs like power plants, collateralizes them like aircraft, and backstops them with customer prepayments that behave like short-term loans.

Compute became a financial asset. It happened in credit markets, in private, in structures priced in billions. And the entry ticket was nine figures.

**That is the problem tokenization exists to solve.**

***

### Why GPUs are the best RWA candidate ever built

Most real-world asset tokenization is an exercise in dressing up something boring. Tokenized Treasuries are Treasuries with a wrapper. Tokenized real estate is an SPV with a Merkle root. The asset gains a blockchain and loses nothing — but it also gains almost nothing, because the underlying instrument was already liquid, already priced, already accessible.

GPUs are different, and the difference is not marketing. Four properties make compute the strongest RWA the sector has:

**1. Standardization.** An H100 is an H100 in Abu Dhabi, Mumbai, or Frankfurt. There is no "comparable sales in the neighborhood" problem, no appraiser's opinion, no bespoke legal wrapper per unit. Real estate tokenization dies on heterogeneity. Compute is fungible by design — a SKU, not a snowflake.

**2. Verifiable, metered cash flow.** A GPU either ran a workload or it did not. Utilization is a machine-readable fact, not a quarterly attestation from a property manager. This is the single most underrated property in the entire RWA category: the asset natively emits the exact telemetry an on-chain system needs to price it. Nothing else in RWA does this. A building does not report its own occupancy to a smart contract every six seconds.

**3. Clear residual value with an active resale market.** Yes, a high-end Nvidia GPU loses roughly half its resale value within three years. That is a real constraint — and it is also *knowable*, which is what underwriters actually need. Ampere-class chips launched in 2020 and remain heavily utilized six years later. There is a functioning secondary market, a depreciation curve you can model, and a floor under the collateral.

**4. Demand that is currently structural, not sentimental.** At GTC in March 2026, Nvidia unveiled Vera Rubin — 336 billion transistors on TSMC's N3P node — while confirming what the supply chain already knew: HBM4 is sold out through 2026, CoWoS packaging capacity is sold out through 2026, and GPU lead times run 36 to 52 weeks. Volume Vera Rubin production does not arrive until early 2027.

Hyperscalers have locked multi-year allocations. Everyone else — startups, mid-market enterprises, research labs, sovereign AI programs — waits until 2027 or pays a premium. The GPU-as-a-Service market sat at $5.7 billion in 2025 and is projected to reach $25.9 billion by 2031. OpenAI alone is estimated to need 4.8 billion GPU-hours this year.

Standardized. Metered. Residual-valued. Supply-constrained. If you were designing the ideal real-world asset to put on-chain from first principles, you would design a GPU.

***

### The access gap is the whole thesis

Here is the asymmetry, stated plainly.

An institution that wants exposure to AI compute economics can buy CoreWeave's paper at SOFR + 225bps, structured through a bankruptcy-remote SPV, with a $19 billion Meta backlog behind it.

An individual who wants the same exposure can… buy NVDA. That is a bet on a chip vendor's margin, not on compute utilization. It is correlated to the entire equity complex, it is priced by narrative, and it has nothing to do with whether the rack in the datacenter is earning.

That gap is not a market failure. It is a market *structure* — one that persists because the minimum viable ticket for GPU infrastructure finance is denominated in millions and the paperwork is denominated in months.

Tokenization collapses both.

***

### What tokenization actually changes — five things, concretely

**Fractionalization.** A rack becomes a position. Exposure to a $250,000 cluster becomes accessible at a size a normal person can write. This is not a philosophical point about "democratization"; it is the mechanical difference between an asset class with 200 participants and one with 200,000.

**Liquidity on an asset that has none.** This is the sharpest edge. Traditional GPU finance locks capital for 5–7 years — DDTL 4.0 matures in 2032. Your money is in until the facility unwinds. A tokenized position trades. You are not underwriting a seven-year hold on hardware with a three-year depreciation curve; you are holding a position you can exit into a bid. Tokenization does not remove duration risk from the asset. It removes it from *you*.

**Transparency that traditional structures cannot match.** The credit market's honest complaint about GPU-backed debt is that the structures are new, the collateral pools are opaque, and the secondary market for seized hardware is thin. Note that the same dollar of chip value can get counted three times — as a demand signal, as booked revenue, and as loan collateral. On-chain settlement makes double-counting structurally harder. Utilization, revenue, and distribution become auditable in public, continuously, by anyone — not disclosed quarterly to whoever holds the paper.

**Composability.** A tokenized GPU position is not a terminal state. It is a primitive. It can collateralize a loan, split into principal and yield tokens, feed a vault strategy, or seed a liquidity pool. GAIB's sAID already does exactly this across Pendle and Morpho. Traditional GPU debt sits in an SPV and does nothing else for anyone. This is the part TradFi structurally cannot copy.

**Faster capital formation for operators.** Look at it from the datacenter's side. Bank approval for hardware financing is a quarters-long process gated on relationships. On-chain capital formation is a days-long process gated on terms. In a market where lead times are 36–52 weeks and every week of latency is revenue that never existed, that delta is the business.

***

### The market has already voted

The numbers are not speculative anymore. Total distributed RWA value on public blockchains reached roughly $31–33.5 billion by early July 2026 per rwa.xyz — excluding stablecoins — held across 167 platforms by nearly 960,000 holders, up more than 400% since early 2025. Add stablecoins and you add about $299 billion.

Worth reading carefully, though: rwa.xyz separates *distributed* value (~~$33.5B, actually issued and tradable) from *represented* value (~~$345B, committed but not yet liquid). The gap is where most of the sector's inflated headlines come from. Judge platforms on the first number.

Meanwhile the DePIN sector runs around $19 billion in market cap, and the AI-compute slice of it is where the real assets live. The convergence is not a thesis anymore — it is a category. KPMG and Nuway published a paper in April 2026 examining physical compute as an alternative real asset, in the same frame as aircraft and real estate.

Two forces are moving toward each other: institutional credit is reclassifying GPUs as infrastructure, and on-chain capital is looking for yield that is not token emissions in a costume. They meet at tokenized compute.

***

### GPUnet's RWA Pool: owning the rack, not the narrative

This is where [**GPUnet's RWA Pool**](https://www.gpu.net/ganchain/rwa-pool) sits.

The proposition is deliberately unglamorous: **invest in real GPU hardware through tokenized contracts on Hyperliquid, with transparent returns, tradeable positions, and automatic profit splits.**

Break that into its four load-bearing claims:

**Real GPU hardware.** Not a governance token with a compute narrative attached. Not an index. Physical GPUs earning rental revenue from real workloads. This distinction matters more than it sounds: from January 2025 through March 2026, six of the seven top RWA project tokens posted negative returns ranging from -44.7% to -98.8% — even as the underlying RWA market quadrupled. The sector grew; the tokens bled. Owning the *asset* and owning the *narrative around the asset* are different trades, and only one of them worked.

**Tokenized contracts on Hyperliquid.** Position, not paperwork. Settlement, transparency, and secondary liquidity in one venue.

**Tradeable positions.** The exit. The thing seven-year term loans do not have.

**Automatic profit splits.** Distribution enforced by code rather than by a servicer's discretion and a payment calendar.

The reason to take GPUnet's version of this seriously is that the compute underneath it is not a roadmap item. GPUnet has been operating a live datacenter business — $10 million in 2024 revenue, roughly $9.8 million of it from enterprise partnerships — with virtualization across global datacenter partners including Northern Data, G42, and NTT, running on GANchain, a Proof of Compute L1. It raised a $5.25M Series A in April 2024 and later executed a $4 million buyback of $GPU from early investors, funded from treasury and operating revenue.

That is the part most tokenized-compute pitches are missing. Tokenization is a *financial* layer. It is only as good as the operating business generating the cash flow underneath it. A pool wrapped around real enterprise compute revenue and a functioning rental dApp is a different instrument than a pool wrapped around a pitch deck.

**→** [**Explore the GPUnet RWA Pool**](https://www.gpu.net/ganchain/rwa-pool)

***

### Read the risks, because they're real

Anyone selling you GPU RWA without this section is selling you something else.

* **Depreciation is the core tension.** Roughly half of resale value gone in three years, against financing structures that run six to nine. The math only works if utilization holds or the contracts behind the hardware keep paying. Ampere's six-year run is the bull case; it is not a guarantee.
* **Utilization is everything.** Many traditional providers run average GPU utilization below 70%. Idle silicon is a cost center wearing an asset's clothes. Yield is a function of hours sold, not chips owned.
* **Counterparty quality degrades as you scale.** CoreWeave demonstrated this in public: DDTL 4.0 got A3 on Meta's investment-grade paper; DDTL 5.0, backed by two non-investment-grade customers, priced materially wider. Ask who the offtakers are. Every time.
* **Secondary markets are thin.** Tokenization creates the *venue* for liquidity. It does not create the *depth*. Early pools trade thin; that is a real cost of being early.
* **Regulatory uncertainty and enforcement of on-chain claims over physical hardware** remain genuinely unresolved across most jurisdictions.
* **Yield claims across the sector vary wildly** — GAIB targets \~11–12% net from contracted financing payments; Compute Labs advertises 30–70% APY from direct leasing. Those are not the same product wearing different numbers. They are different risk positions. Read the structure, not the headline.

None of this is disqualifying. All of it is why the transparency and tradeability of a tokenized structure are features and not decoration — they are the tools you use to manage exactly these risks in real time, rather than discovering them at maturity.

***

### The close

Every generation of infrastructure gets financialized eventually, and the pattern is always the same: the asset becomes ownable long before it becomes ownable *by you*. Railroads, then real estate, then energy, then aircraft. By the time retail access arrived, the repricing had happened.

Compute is mid-cycle right now. The reclassification is done — investment-grade ratings settled that in March. The supply constraint is real through 2027. The demand is contracted, not projected. What has not been settled is distribution: who gets a position, at what size, with what exit.

Tokenization is the answer to that last question, and it is being answered this year.

[**GPUnet's RWA Pool →**](https://www.gpu.net/ganchain/rwa-pool) — real hardware, tokenized contracts on Hyperliquid, transparent returns, tradeable positions, automatic profit splits.

Own compute. Not the story about it.

***

*This article is for informational purposes and is not investment advice. Tokenized GPU exposure carries hardware depreciation, utilization, counterparty, liquidity, and regulatory risk. Do your own research.*


