# 0G Compute Network

> 0G Compute Network is a decentralized GPU marketplace for AI inference, fine-tuning, and model training — pay only for what you use.


In today's world, AI models are transforming industries, driving innovation, and powering new applications. However, running advanced AI models for your application faces several obstacles:

- **High Costs**: Enterprise AI services require significant monthly commitments
- **Complex Setup**: Cloud GPU configuration requires technical expertise
- **Vendor Lock-in**: Limited flexibility when switching providers

**The result?** AI computing remains inaccessible for many developers and startups.

## What is 0G Compute?

0G Compute is a decentralized framework that provides AI computing capabilities to our community. It forms a crucial part of the 0G stack. 0G Compute is a decentralized marketplace where GPU owners sell computing power to developers who need it - think Uber for AI computing.

**Key difference**: Instead of renting from AWS/Google with high costs and lock-in, you access a global GPU network that's 90% cheaper with pay-per-use pricing.

## How It Works

### For AI Users

1. **Deposit Funds**: Pre-fund your account with pay-as-you-go credits
2. **Request Service**: Send your AI request (inference, training, etc.)
3. **Get Results**: Receive output from the best available GPU
4. **Automatic Payment**: Pay only for actual compute used

### For GPU Owners

1. **Register Hardware**: List your GPU specs and availability
2. **Set Your Price**: Competitive marketplace pricing
3. **Process Requests**: Automatic job allocation
4. **Instant Earnings**: Get paid immediately upon completion

### Technical Implementation

- **Smart Contract Escrow**: Funds secured until service delivery
- **Signed Transactions**: Cryptographic verification of all interactions
- **ZK-Proof Settlement**: 100x lower transaction costs through compressed proofs

<details>
<summary>**What makes it trustless?**</summary>

Like eBay with automatic escrow - the smart contract ensures:

- Payment only after service delivery
- Both parties must fulfill obligations
- No intermediary can interfere

This means no one can censor your AI usage, freeze your account, or change terms suddenly.

</details>

## Why Choose 0G Compute?

### 💰 Key Advantages

| Feature          | Traditional Cloud   | 0G Compute         |
| ---------------- | ------------------- | ------------------ |
| Pricing Model    | Fixed monthly costs | Pay-per-use        |
| Provider Options | Limited vendors     | Global GPU network |

### 🌐 Access From Anywhere

- **Blockchains**: Direct integration with Ethereum, Solana, any chain
- **Traditional Apps**: Simple REST API using 0G SDK

### 🔐 Your Data, Your Control

- No data retention by providers
- Verifiable computation proofs - Supports TEEML, OPML & ZKML

## Common Questions

<details>
<summary>**What about reliability?**</summary>

Built-in redundancy:

- Automatic failover to next provider
- Thousands of providers globally

</details>

<details>
<summary>**Can I run proprietary models?**</summary>

Yes. Upload any model, set requirements and pricing, and start serving requests. Perfect for specialized use cases.

</details>

<details>
<summary>**How does pricing work?**</summary>

Pure pay-per-use:

- No subscriptions
- Competitive market-driven pricing
- Transparent costs visible upfront

</details>

## Get Started

### 📚 Technical Deep Dive

Architecture and implementation details  
→ [Technical Documentation](/developer-hub/building-on-0g/compute-network/overview)

### 🚀 For Developers

Start using AI compute in 5 minutes  
→ [Quick Start Guide](/developer-hub/building-on-0g/compute-network/router/quickstart)

### 💎 For GPU Owners

Turn idle hardware into income  
→ [Become a Provider](/developer-hub/building-on-0g/compute-network/inference-provider)

---

*0G Compute: Making AI accessible to everyone.*