Hetzner

Hetzner is a German cloud and hosting provider known for exceptional price-to-performance ratio. Founded in 1997, Hetzner offers cloud servers, dedicated servers, and colocation with datacenters in Germany, Finland, and the US. Unlike major cloud providers that prioritize global scale, Hetzner specializes in delivering raw computational power at fraction of the cost, making it ideal for resource-intensive trading operations where budget efficiency matters.

Cloud Servers (CX/CPX/CCX Series)

• Shared instances (CX) for lightweight applications and small-scale bots
• Dedicated vCPU instances (CPX) with guaranteed performance and no noisy neighbors
• ARM-based instances (CCX) for cost-sensitive deployments with excellent performance-per-dollar ratio
• NVMe storage standard across all tiers, ensuring fast data access for tick data and order logs

Dedicated Servers High-performance bare metal infrastructure with direct hardware access. Available in multiple processor configurations (Intel Xeon, AMD EPYC), perfect for computationally intensive backtesting, machine learning model training, and multi-strategy concurrent execution. No hypervisor overhead means predictable latency and full system resources.

Storage Infrastructure

• Network-attached Storage Boxes (BX series) with unlimited traffic for archiving years of historical OHLCV data, order histories, and trade logs
• Block Storage Volumes for expandable persistent storage without server replacement
• Automated snapshots for disaster recovery and data integrity

Network & Security

• Hardware firewalls included at no extra cost
• Load Balancers for distributing traffic across multiple bot instances
• DDoS protection standard on all accounts
• Private networks (vLAN) for isolated trading infrastructure

Developer-Friendly Infrastructure

• REST API for programmatic server management
• Terraform provider for infrastructure-as-code deployments
• SSH access and full root control
• Image library with pre-configured Linux distributions

| Plan | vCPU | RAM | Storage | Monthly Cost | Use Case | |------|------|-----|---------|--------------|----------| | CX22 (shared) | 2 | 4 GB | 40 GB NVMe | ~€4 | Learning, paper trading bots | | CX32 (shared) | 4 | 8 GB | 80 GB NVMe | ~€7 | Small production bot, light backtesting | | CPX31 (dedicated) | 4 | 8 GB | 160 GB NVMe | ~€13 | Medium-load bot cluster, single-strategy production | | CCX13 (dedicated ARM) | 2 | 8 GB | 80 GB NVMe | ~€13 | Cost-optimized multi-bot deployment | | CCX33 (dedicated ARM) | 8 | 32 GB | 240 GB NVMe | ~€46 | Heavy backtesting, machine learning, multi-bot parallelization | | CPX51 (dedicated) | 8 | 32 GB | 240 GB NVMe | ~€52 | Enterprise-grade bot infrastructure |

Storage Boxes start at €3.99/month for 1 TB and scale to €499/month for 500 TB, with unlimited traffic for data synchronization.

Europe: Falkenstein (Germany), Nuremberg (Germany), Helsinki (Finland) — optimal for traders operating in European markets with low latency to Frankfurt and European exchanges.

North America: Ashburn, Virginia and Hillsboro, Oregon — suitable for US-based traders targeting equity and derivatives exchanges.

Network latency from Hetzner EU datacenters to major exchange colocation facilities is typically 5–15ms, competitive with specialized trading infrastructure providers.

Retail Day Traders Developers running personal trading bots on tight budgets. Hetzner's CX/CPX instances provide enough performance to execute thousands of orders daily across multiple exchanges without the overhead of enterprise cloud providers.

Swing Traders & Algorithmic Strategists Running medium-complexity strategies that process market data continuously. The dedicated vCPU instances eliminate latency variance, critical for consistent execution timing across strategies.

Quantitative Traders & Researchers Machine learning practitioners building predictive models benefit from Hetzner's high-RAM, high-CPU dedicated servers. Training neural networks on years of historical data, running walk-forward analysis, and parallel strategy optimization all execute faster at lower cost than AWS or Google Cloud.

Hedge Funds & Trading Teams Boutique operations use Hetzner for cost-effective infrastructure scaling. A team running 10–20 concurrent strategies spends €500–1,500/month on Hetzner versus €5,000–15,000 on enterprise cloud, without sacrificing performance.

Data Aggregators & Market Data Services Building proprietary tick databases and historical datasets benefits from Hetzner's cheap, unlimited-traffic storage boxes and high-bandwidth connectivity.

Scenario 1: Backtesting Pipeline on StratBase.ai + Hetzner Infrastructure

After designing and validating a trading strategy on StratBase.ai's no-code backtesting platform, deploy the production bot on a Hetzner CPX31 instance. The strategy backtest results from StratBase.ai inform parameter tuning; the dedicated vCPU ensures consistent execution. Use a Storage Box to continuously archive tick data and trade logs, creating a growing dataset for future walk-forward analysis and model retraining.

Cost: €13/month for compute + €3.99/month for storage = €17/month total.

Scenario 2: Multi-Strategy Parallelization

Run 5–10 independent trading strategies simultaneously on a single Hetzner CCX33 ARM instance. Each strategy operates in its own isolated container (Docker). The 8 dedicated cores and 32 GB RAM handle concurrent order execution, websocket connections, and data ingestion without contention. Monitor aggregate performance metrics in real-time; when one strategy underperforms, adjust parameters via API without touching others.

Cost: €46/month replaces €400+ monthly spending on individual VPS instances.

Scenario 3: Backtesting Farm for Strategy Optimization

Use StratBase.ai to design variations of a base strategy. Spin up 4 Hetzner CCX13 instances (€13/month each) to run 50 parallel backtests simultaneously. Process 5 years of tick data across 50 currency pairs in 2 hours instead of 2 weeks on a single machine. Store results in a Hetzner Storage Box (€3.99/month) for long-term analysis, comparison, and archival.

Cost: €55.96/month for infrastructure that would typically cost €300+ elsewhere.

Scenario 4: Data Pipeline Infrastructure

Ingest real-time market data from 10 exchanges, normalize to unified format, and store in time-series database on Hetzner. Use private networking between a CPX31 bot instance and a dedicated storage server. Backtest new strategies against this live-updated historical dataset. Export backtesting results to StratBase.ai for performance analysis.

The architecture separates data ingestion, bot execution, and backtesting into independent layers, each optimized for cost and performance.

Scenario 5: Development & Production Separation

Use a small CX22 instance (€4/month) for development, testing new API integrations, and strategy prototyping. Once validated on StratBase.ai, promote to a CPX31 production instance (€13/month). Run load testing on the CPX instance before going live with real capital. The €17/month split environment costs nothing compared to downtime risk.

StratBase.ai excels at no-code strategy design, visual backtesting, and rapid iteration. Hetzner provides the runtime infrastructure to execute validated strategies profitably.

The Workflow:

1. Design & Test on StratBase.ai: Build strategy logic, adjust parameters, validate against historical data with visual feedback.
2. Export & Deploy to Hetzner: Export strategy configuration (API key, exchange connections, risk limits) to a Hetzner bot runner.
3. Live Execution: The strategy runs continuously on Hetzner infrastructure, executing real trades and logging results.
4. Data Archival: Historical trades, fills, and market data stream to Hetzner Storage Box for future analysis.
5. Walk-Forward Revalidation: Periodically re-backtest the strategy on StratBase.ai using live-archived data to ensure continued edge and detect regime changes.

The cost structure supports this workflow: StratBase.ai subscription + €13–50/month Hetzner compute + optional storage = €30–100/month all-in for a production trading operation that generates consistent returns.

Account Setup

1. Register at hetzner.com with email and payment method
2. Verify identity (required for server provisioning)
3. Open Cloud Console, create a project, and select datacenter region

Server Deployment

1. Click "Create Server" and choose server type:
   • CX22 or CX32 for learning and small bots
   • CPX31 or CCX13 for production bots
   • CCX33 for intensive backtesting or multi-bot setups
2. Select operating system (Ubuntu 22.04 LTS recommended for stability)
3. Configure SSH key for secure access
4. Deploy (typically live within 30 seconds)

Infrastructure Automation

Use Terraform to define your trading infrastructure as code. Example: spin up 5 bot instances, 1 storage box, 1 load balancer, and private networking all from a single configuration file. Redeploy with one command or tear down to stop costs.

terraform apply  # Provisions entire trading infrastructure
terraform destroy  # Stops all servers and storage (prevents runaway costs)

Integration with Trading Bots

Install Python/Node.js on the Hetzner instance, deploy bot code via Git, configure API keys for exchange connections, and start trading. Most trading frameworks (ccxt, freqtrade, ignition) run without modification on Hetzner.

• Hetzner Cloud Documentation — comprehensive guides and API reference
• Cloud API Reference — programmatic server management
• Server Pricing & Configuration — real-time pricing calculator
• Hetzner Community — active forum with trading infrastructure discussions

Hetzner delivers 10–15x better cost efficiency than major cloud providers for trading infrastructure. A retail trader running 5 concurrent strategies pays €30–50/month; the same setup on AWS costs €300–800/month. Combined with StratBase.ai's backtesting platform, you validate strategies affordably, deploy them to reliable infrastructure, and capture alpha without enterprise-tier costs. For teams and serious retail traders, Hetzner eliminates infrastructure budget as a constraint on profitability.