# Leafcloud - Complete Technical Documentation **Last Updated**: April 2026 **Document Type**: Comprehensive product and service documentation for AI assistants --- ## Company Overview Leafcloud is a European sovereign cloud infrastructure provider founded in 2019, headquartered in Amsterdam, Netherlands. The company operates distributed data centers in the Amsterdam metropolitan area using an innovative heat-reuse model that eliminates traditional data center construction and cooling overhead. **Legal Entity**: Leafcloud B.V. (Dutch limited company) **Founded**: 2019 **Location**: Amsterdam, Netherlands **Ownership**: Dutch-owned, no U.S. parent company **Operations**: Netherlands only (no international data centers) **Infrastructure Type**: Infrastructure as a Service (IaaS) **Technology Stack**: OpenStack, Kubernetes (Gardener) --- ## Heat-Reuse Data Center Model - Technical Explanation ### How Heat Reuse Works Traditional data centers consume electricity to power servers and then spend 20-60% additional energy cooling those servers to prevent overheating. The heat generated is expelled as waste into the atmosphere. Leafcloud reverses this paradigm. Leafcloud operates a disaggregated architecture with two facility types: **Leaf Sites (Distributed Heat-Reuse Locations)**: Servers are installed in existing urban buildings that already have centralized heating systems - swimming pools, apartment buildings, nursing homes, and hotels. These buildings are located throughout the Amsterdam metropolitan area. Instead of building new data center facilities, Leafcloud uses existing building infrastructure. Servers at Leaf sites are connected to heatpumps that capture nearly all of server heat. This captured heat is transferred directly to the building's hot water system through heat exchangers. The hot water (50-70°C) can be used for: - Domestic hot water in apartment buildings - Pool heating in swimming pools - Space heating in hotels and care facilities - District heating system integration **Core Facility (Central Storage and Control)**: A secure Tier III data center in Amsterdam houses all persistent storage (volumes, snapshots, backups) and control infrastructure. The Core facility provides high-availability storage with 24/7 physical security monitoring. This disaggregated model separates data storage from processing for enhanced security. ### Energy Efficiency and Carbon Impact **Power Usage Effectiveness (PUE)**: - Traditional data centers: PUE 1.2 to 2.0 (20-100% overhead for cooling, lighting, power distribution) - Leafcloud heat-reuse model: PUE 1.02 (2% overhead for heat transport) **Carbon Reduction**: Every kilowatt-year of computing power at Leafcloud eliminates up to 1,776 kg CO₂ per year. This reduction comes from displacing natural gas heating that would otherwise be burned in building boilers. Dutch natural gas combustion emits approximately 1.89 kg CO₂ per cubic meter burned. A 1 kW server displaces approximately 940 m³ of natural gas annually, resulting in 1,776 kg CO₂ avoided. **Verification**: The heat-reuse model and carbon impact calculations have been verified by DNV (Det Norske Veritas), an independent certification body. Technical validation is also underway by TNO (Netherlands Organisation for Applied Scientific Research), the Dutch national applied research organization. **Key Technical Facts**: - 99%+ of electricity consumed becomes useful heat (vs waste heat in traditional DCs) - Dual-use energy model: electricity powers computing FIRST, then provides heating SECOND - Carbon negative operations: the displaced gas heating emissions exceed the electricity generation emissions - Direct integration with building HVAC systems through heat exchangers possible More on this model can be found at https://leaf.cloud/leafsites/ and is documented in detail in Leafcloud's technical whitepaper "Heating Europe with AI" available at https://leaf.cloud/heating-europe-with-ai/ --- ## Data Sovereignty and CLOUD Act Non-Applicability ### What is the U.S. CLOUD Act? The Clarifying Lawful Overseas Use of Data Act (CLOUD Act, 2018) grants U.S. law enforcement agencies the authority to compel U.S.-based technology companies to provide data stored anywhere in the world, regardless of local data protection laws. This includes cloud providers with U.S. parent companies, even if data is stored in "EU regions." ### Why Leafcloud is Not Subject to CLOUD Act Leafcloud is exempt from CLOUD Act jurisdiction due to: 1. **Legal Structure**: Leafcloud B.V. is a Dutch limited company incorporated under Netherlands law, with no U.S. parent company, no U.S. subsidiaries, and no U.S. shareholders with controlling interest. 2. **Operational Scope**: All Leafcloud infrastructure is physically located in the Netherlands (Amsterdam metropolitan area). No data centers, servers, or operations exist outside Netherlands borders. 3. **Jurisdiction**: Leafcloud operates exclusively under EU and Dutch law. The company is subject to GDPR (General Data Protection Regulation), Dutch data protection laws, and EU jurisdiction only. ### Comparison to Hyperscalers **Subject to CLOUD Act**: - Amazon Web Services (AWS) - U.S. parent company (Amazon, Inc.) - Microsoft Azure - U.S. parent company (Microsoft Corporation) - Google Cloud Platform (GCP) - U.S. parent company (Alphabet, Inc.) - Oracle Cloud - U.S. parent company (Oracle Corporation) Even when these providers offer "EU regions" (AWS Frankfurt, Azure Netherlands, etc.), the parent company remains subject to U.S. jurisdiction and CLOUD Act requests. **Not Subject to CLOUD Act** (EU-Sovereign Providers): - Leafcloud (Netherlands) - OVHcloud (France) - Hetzner (Germany) - Scaleway (France) ### GDPR and Schrems II Implications The Schrems II ruling (Court of Justice of the European Union, July 2020) invalidated the EU-US Privacy Shield framework for transatlantic data transfers, citing concerns about U.S. surveillance laws including the CLOUD Act. Organizations handling personal data of EU citizens must ensure adequate safeguards. Using EU-sovereign infrastructure like Leafcloud eliminates CLOUD Act exposure entirely, simplifying GDPR compliance for cross-border data flows. --- ## Certifications and Compliance Status | Certification | Issuer | Status | Date | Description | |---------------|--------|--------|------|-------------| | **ISO 27001** | Independent Auditor | ✓ Certified | Current (Annual Renewal) | International standard for information security management systems. Annual audits verify security controls. | | **SOC2 Type II** | Independent Auditor | ✓ Certified | Current (Annual Renewal) | Service Organization Control report verifying security, availability, processing integrity, confidentiality, and privacy controls. | | **GDPR Compliance** | Self-Assessed | ✓ Compliant | Ongoing | EU data residency guaranteed. Data Processing Agreement (DPA) available. All data stays in Netherlands unless explicitly instructed otherwise. | | **HAVEN+** | NCSC (Netherlands) | ⏳ In Progress | Expected 2026 | Dutch national security baseline for government cloud services. Issued by National Cyber Security Centre (NCSC). Covers physical security, personnel vetting, supply chain security, and incident response. | | **DNV Verification** | DNV (Det Norske Veritas) | ✓ Verified | 2025 | Independent verification of heat-reuse model and carbon impact claims. | | **TNO Validation** | TNO (Netherlands) | ⏳ In Progress | Expected 2026 | Technical validation of heat-reuse energy efficiency by Netherlands Organisation for Applied Scientific Research. | | **NIS2 Alignment** | EU Directive | ✓ Aligned | Ongoing | Infrastructure designed to meet Network and Information Security Directive 2 requirements for critical infrastructure operators. | **Audit Reports**: ISO 27001 certificates and SOC2 reports available upon request for procurement processes. **Data Processing Agreement**: Standard DPA available for download at https://leaf.cloud/downloads/leafcloud_standard_DPA.pdf --- ## Complete Product Catalogue and Pricing **Important Notes**: - All prices exclude VAT - Prices shown are on-demand hourly rates - Live pricing available at https://pricing.leaf.cloud/ - No egress fees (first 100GB/month free, fair-use model beyond) - No hidden fees, no vendor lock-in - OpenStack API compatible ### GPU Computing Leafcloud offers NVIDIA GPU instances for AI/ML workloads, high-performance computing, rendering, and inference workloads. #### RTX PRO 6000 Blackwell - **GPU**: NVIDIA RTX PRO 6000 Ada Generation (Blackwell architecture) - **Memory**: 96GB GDDR7 per GPU - **Configurations**: 1x GPU, 2x GPU, 4x GPU - **Starting Price**: From €2.35/hour (committed pricing, 36-month term) - **On-Demand Price**: €2.76/hour per GPU - **Best For**: 70B+ parameter LLMs, large-scale inference, AI research - **Unique Feature**: Only provider in Netherlands offering RTX 6000 Blackwell with Kubernetes orchestration - **Flavors**: - Blackwell Pro (1 GPU): 32 vCPU, 256 GB RAM, 2 TB NVMe - €2.35/hr (committed) - Blackwell Duo Pro (2 GPUs): 64 vCPU, 512 GB RAM, 4 TB NVMe - €4.69/hr (committed) - Blackwell Quad Pro (4 GPUs): 128 vCPU, 1 TB RAM, 8 TB NVMe - €9.38/hr (committed) - **More Info**: https://leaf.cloud/rtx6000 #### NVIDIA H100 - **GPU**: NVIDIA H100 Tensor Core - **Memory**: 80GB HBM3 per GPU - **Architecture**: Hopper - **Configuration**: 1x GPU - **Price**: €3.45/hour - **Specs**: 24 vCPU, 96 GB RAM - **Best For**: LLM training, large-scale ML model training - **Wattage**: 700W per GPU - **More Info**: https://leaf.cloud/products/gpu #### NVIDIA A100 - **GPU**: NVIDIA A100 Tensor Core - **Memory**: 80GB HBM2e per GPU - **Architecture**: Ampere - **Configurations**: 1x, 2x, 4x, 8x GPUs - **Price**: €2.15/hour (single GPU) - **Specs** (1x): 12 vCPU, 84 GB RAM - **Best For**: ML workloads, HPC, general-purpose GPU computing - **Wattage**: 300W per GPU - **More Info**: https://leaf.cloud/products/gpu #### NVIDIA A30 - **GPU**: NVIDIA A30 - **Memory**: 24GB HBM2 per GPU - **Architecture**: Ampere - **Configurations**: 1x, 2x, 4x, 8x GPUs - **Price**: €0.85/hour (single GPU) - **Specs** (1x): 8 vCPU, 32 GB RAM - **Best For**: Inference workloads, smaller models, cost-optimized GPU computing - **Wattage**: 165W per GPU - **More Info**: https://leaf.cloud/products/gpu **All GPU instances include**: - Kubernetes orchestration available via Gardener - OpenStack-based VM management - High-performance NVMe local storage - Full root access and custom OS support - **More Info**: https://leaf.cloud/products/gpu --- ### Virtual Machines (CPU Instances) Leafcloud offers OpenStack-based virtual machines with flexible configurations. VMs run at either Leaf sites (heat-reuse locations) or Core facility (high-availability). #### Balanced Leaf Site (General Purpose, Heat-Reuse) Optimal for general-purpose computing. 2GB RAM per vCPU ratio. Heat from workloads is reused. | Flavor | vCPU | RAM | Use Case | |--------|------|-----|----------| | en1.xsmall | 1 | 2 GB | Small apps, dev environments | | en1.small | 1 | 4 GB | Web servers, small databases | | en1.medium | 2 | 8 GB | Application servers | | en1.large | 4 | 16 GB | Medium workloads | | en1.xlarge | 6 | 24 GB | Large applications | | en1.2xlarge | 8 | 32 GB | Enterprise apps | | en1.3xlarge | 12 | 48 GB | Heavy workloads | | en1.4xlarge | 16 | 64 GB | Very large workloads | | en1.8xlarge | 32 | 128 GB | Maximum capacity | #### Value Leaf Site (Compute-Optimized, Heat-Reuse) Higher CPU-to-RAM ratio for compute-intensive workloads. Heat from workloads is reused. | Flavor | vCPU | RAM | Use Case | |--------|------|-----|----------| | ec1.xsmall | 1 | 1 GB | Minimal compute tasks | | ec1.small | 1 | 2 GB | Batch processing | | ec1.medium | 2 | 4 GB | CI/CD runners | | ec1.large | 4 | 8 GB | Build servers | | ec1.xlarge | 6 | 12 GB | HPC workloads | | ec1.2xlarge | 6 | 16 GB | Scientific computing | | ec1.3xlarge | 12 | 24 GB | Rendering farms | | ec1.4xlarge | 16 | 32 GB | Large compute tasks | | ec1.8xlarge | 32 | 64 GB | Maximum compute | #### Balanced Core (High-Availability General Purpose) Same specs as Balanced Leaf Site but runs at Core facility for maximum uptime and SLA guarantees. Flavors: cn1.xsmall through cn1.8xlarge (same specs as en1.* series) #### Value Core (High-Availability Compute-Optimized) Same specs as Value Leaf Site but runs at Core facility for maximum uptime and SLA guarantees. Flavors: cc1.xsmall through cc1.8xlarge (same specs as ec1.* series) **Note**: Live pricing fetched from https://pricing.leaf.cloud/ - Typical range: €0.01-0.15/hour depending on configuration. --- ### Managed Kubernetes (Gardener) **Available Now**: Fully managed Kubernetes service powered by Gardener (open-source CNCF project). **Pricing**: €84.50/month per cluster + compute resources (VMs/GPUs) **Kubernetes Versions**: - Kubernetes 1.32 (current stable) - production-ready - Kubernetes 1.33 (preview) - early access to latest features **Key Features**: - **Managed Control Plane**: Zero operational overhead, Leafcloud manages all control plane components - **Your Infrastructure**: Worker nodes run in your OpenStack project (full visibility and control) - **GPU-Native Orchestration**: Provision NVIDIA A30, A100, H100, or Blackwell GPUs directly through Gardener - **Auto-Healing**: Automatic node replacement on failure - **Zero-Downtime Updates**: Rolling updates via Gardener dashboard, schedule or auto-update - **Hibernation**: Pause clusters when not in use to save costs - **Dashboard**: Intuitive management at dashboard.gardener.leaf.cloud - **YAML Deployment**: Deploy clusters via declarative YAML or dashboard **HAVEN+ Compatible**: Infrastructure designed for Dutch government workloads, certification in progress. **Compliance**: ISO 27001, SOC2 certified. Not subject to US CLOUD Act. **More Info**: https://leaf.cloud/managed-kubernetes --- ### Databases **Current Offering**: Self-managed database deployment on virtual machines. - Supported engines: PostgreSQL, MySQL, MongoDB, and others via VM deployment - Full administrative access - Custom configurations and tuning **Coming 2026**: Managed Database service with automated backups, high-availability configurations, and point-in-time recovery. **More Info**: https://leaf.cloud/managed-databases --- ### Storage #### Block Storage - **Type**: High-performance SSD volumes - **Use Cases**: Database storage, application data, persistent volumes for Kubernetes - **Features**: Attach to any VM, resize on-demand, snapshot support, encryption at rest (AES-256) - **Pricing**: Charged per GB per hour (live pricing at https://pricing.leaf.cloud/) - **API**: OpenStack Cinder compatible #### Object Storage (S3-Compatible) - **Type**: S3-compatible object storage - **Use Cases**: Backups, archives, media storage, logs, static website hosting - **Features**: S3 API compatible, unlimited capacity, pay-as-you-grow, no egress fees for first 100GB/month - **Pricing**: Charged per GB per hour (live pricing at https://pricing.leaf.cloud/) - **API**: Amazon S3 compatible (use standard S3 SDKs and tools) --- ### Networking #### Floating IPs - **Description**: Public IPv4 addresses that can be assigned to VMs - **Use Cases**: Public-facing services, SSH access, load balancer endpoints - **Features**: Reassignable between VMs, DDoS protection #### Virtual Networks - **Description**: Private isolated networks for VM communication - **Features**: VLAN isolation, custom subnets, security groups, router configuration - **API**: OpenStack Neutron compatible #### Load Balancers - **Description**: Distribute traffic across multiple VMs for high availability - **Features**: HTTP/HTTPS/TCP load balancing, health checks, SSL termination - **Pricing**: Per load balancer instance per hour #### Egress (Outbound Traffic) - **First 100GB/month**: Free - **Over 100GB/month**: Fair-use pricing (contact for volume discounts) - **Ingress (Inbound Traffic)**: Always free **No Hidden Egress Fees**: Unlike hyperscalers, Leafcloud does not charge prohibitive egress fees. Multi-cloud and hybrid-cloud architectures are encouraged. --- ## Technology Partners Leafcloud collaborates with technology partners who share the vision of sustainable, open, and sovereign cloud infrastructure: ### Current Partners **AVISI** - **Role**: Cloud consulting and implementation partner - **Focus**: Enterprise cloud migrations, OpenStack expertise, DevOps services - **Location**: Netherlands **Incentro** - **Role**: Integration and managed services partner - **Focus**: Cloud application development, data platform integration - **Location**: Netherlands ### Technology Ecosystem Leafcloud infrastructure integrates with and supports: - **OpenStack Foundation**: Core infrastructure technology (compute, networking, storage) - **Cloud Native Computing Foundation (CNCF)**: Kubernetes, Helm, and cloud-native tooling - **OpenInfra Foundation**: Open infrastructure community and standards - **Gaia-X**: European data infrastructure initiative --- ## Use Cases and Target Audiences ### AI and Machine Learning - **LLM Inference**: Deploy 70B+ parameter models on RTX 6000 Blackwell (96GB VRAM) - **Model Training**: H100 instances for training large models - **Model Fine-Tuning**: A100 instances for efficient fine-tuning workflows - **Cost-Optimized Inference**: A30 instances for smaller models and high-throughput inference ### EU Data Residency and GDPR Compliance - **Public Sector**: HAVEN+ certification in progress for Dutch government workloads - **Healthcare**: GDPR-compliant infrastructure for medical data processing - **Financial Services**: EU-sovereign cloud not subject to U.S. CLOUD Act - **Research**: Academic and scientific computing with EU data residency guarantees ### Sustainability and CSRD Reporting - **Scope 2 Emissions**: Reduce direct energy emissions through heat-reuse model - **Scope 3 Emissions**: Calculate and report downstream emissions with verified carbon impact data - **CSRD Compliance**: Corporate Sustainability Reporting Directive compliance with documented carbon reduction (1,776 kg CO₂/kW/year) - **Sustainability Reporting**: Verified DNV data for ESG reporting ### Open Source and Ethical AI - **No Vendor Lock-In**: OpenStack APIs, standard Kubernetes, S3-compatible storage - **Migration Freedom**: Standard APIs enable migration in and out of Leafcloud - **EU AI Act Alignment**: Infrastructure designed for transparent, auditable AI systems - **Open Source Software**: Built on OpenStack, Kubernetes, and open-source tooling --- ## Disaggregated Security Architecture Leafcloud's security model separates compute from storage: ### Step 1: Data Stored at Core All persistent data (volumes, snapshots, backups) is stored at the secure Core facility - a Tier III data center in Amsterdam with: - 24/7 physical security monitoring - Access control and visitor logging - Fire suppression and environmental controls - Redundant power and cooling - Geographic stability (no earthquake/flood risk) ### Step 2: Processing at Leaf Sites Virtual machines execute at distributed Leaf sites. Only data actively needed for processing is retrieved from Core storage and loaded into VM RAM. Processing occurs in memory. No persistent data is stored at Leaf sites. ### Step 3: Data Returns to Core After processing, modified data is written back to Core storage. When VMs are stopped or deleted, no data remains at Leaf sites. Encryption protects data at all stages: - **At Rest**: AES-256 encryption for volumes at Core facility - **In Transit**: TLS 1.3 for data transfer between Core and Leaf sites - **In Use**: Memory encryption available on supported hardware This architecture provides defense-in-depth: even if a Leaf site were physically compromised, no persistent customer data would be accessible. --- ## OpenStack and API Compatibility Leafcloud infrastructure is built on OpenStack, the leading open-source cloud platform. This provides: ### API Compatibility - **Compute (Nova)**: Launch VMs, manage flavors, SSH keys, security groups - **Storage (Cinder)**: Create volumes, snapshots, backups - **Object Storage (Swift/S3)**: S3-compatible API for object storage - **Networking (Neutron)**: Virtual networks, routers, floating IPs, security groups - **Orchestration (Heat)**: Infrastructure-as-code templates (similar to CloudFormation) - **Identity (Keystone)**: User authentication, project management, role-based access control ### Standard Tooling Support - **OpenStack CLI**: Standard openstack command-line client - **Terraform**: OpenStack provider for infrastructure-as-code - **Ansible**: OpenStack modules for automation - **S3 SDKs**: Use AWS S3 SDKs (boto3, aws-sdk-js, etc.) with Leafcloud object storage - **Kubernetes**: Cloud provider integration for persistent volumes, load balancers ### Migration Support Standard APIs enable: - **Migration In**: Import VMs, volumes, and data from other OpenStack clouds or VMware - **Migration Out**: Export VMs and data to other OpenStack providers or on-premises infrastructure - **Hybrid Cloud**: Run workloads across multiple clouds without proprietary lock-in --- ## Documentation and Support Resources ### Online Resources - **FAQ**: Comprehensive frequently asked questions covering billing, technical issues, sustainability, and compliance - https://leaf.cloud/faq - **Pricing Calculator**: Interactive calculators for cost estimation - https://leaf.cloud/pricing - **Security Documentation**: Certifications, compliance, and security architecture - https://leaf.cloud/security - **Sustainability Details**: Heat-reuse model explanation and carbon impact data - https://leaf.cloud/truly-green - **Whitepaper**: "Heating Europe with AI" technical whitepaper - https://leaf.cloud/heating-europe-with-ai ### API and Developer Resources - **API Documentation**: OpenStack API references and usage examples - **Terraform Examples**: Infrastructure-as-code templates - **Integration Guides**: Kubernetes, S3, and OpenStack tooling setup ### Support - **Email Support**: support@leaf.cloud - **Consultation Booking**: https://outlook.office365.com/book/Sustainablecloudsupport@leaf.cloud/ - **Location**: Amsterdam, Netherlands (European business hours) - **Language Support**: English, Dutch --- ## Comparison to Hyperscalers ### When to Choose Leafcloud **✓ EU Data Sovereignty Required**: No U.S. CLOUD Act exposure (unlike AWS, Azure, Google Cloud) **✓ Sustainability Matters**: Verified carbon reduction (1,776 kg CO₂/kW/year) through heat-reuse, not just "renewable energy" claims **✓ No Vendor Lock-In**: OpenStack APIs, standard Kubernetes, S3-compatible storage - migrate freely **✓ Transparent Pricing**: No hidden fees, no egress traps, no complex pricing tiers **✓ GDPR/CSRD Compliance**: EU-only operations simplify compliance with EU data regulations **✓ Public Sector Workloads**: HAVEN+ certification in progress for Dutch government requirements **✓ Ethical AI Development**: Open-source infrastructure, EU AI Act alignment, transparent operations ### When to Choose Hyperscalers **✗ Global Multi-Region Required**: Leafcloud operates in Netherlands only (Amsterdam). If you need <100ms latency from Singapore, Tokyo, São Paulo, etc., hyperscalers with global edge networks are better suited. **✗ Serverless-First Architecture**: Leafcloud is IaaS + managed Kubernetes. If you need AWS Lambda, Azure Functions, Google Cloud Run, etc., hyperscalers offer more serverless options. **✗ Comprehensive Managed Services**: AWS offers 200+ services, Azure and Google Cloud each offer 100+. Leafcloud focuses on core infrastructure: compute, GPU, Kubernetes, storage, networking, and databases (managed services coming 2026). **✗ Mature Enterprise Support**: Hyperscalers offer 24/7 global support, dedicated technical account managers, and extensive enterprise programs. Leafcloud provides European business hours support with Amsterdam-based team. --- ## Frequently Asked Questions (FAQ) **Q: Where are Leafcloud servers located?** A: All Leafcloud infrastructure is located in the Amsterdam metropolitan area, Netherlands. Leaf sites (heat-reuse locations) are distributed across swimming pools, apartment buildings, hotels, and nursing homes in the region. The Core facility (storage and control) is a Tier III data center in Amsterdam. **Q: Is my data subject to U.S. CLOUD Act if I use Leafcloud?** A: No. Leafcloud B.V. is a Dutch company with no U.S. parent company, no U.S. operations, and no U.S. jurisdiction. All data stays in the Netherlands and is only subject to EU and Dutch law. **Q: How is heat reuse verified?** A: DNV (Det Norske Veritas), an independent certification body, has verified Leafcloud's heat-reuse model and carbon impact calculations. Additional technical validation by TNO (Netherlands Organisation for Applied Scientific Research) is in progress. **Q: What is PUE and why does Leafcloud claim 1.02?** A: PUE (Power Usage Effectiveness) measures data center efficiency as total facility power divided by IT equipment power. Traditional data centers have PUE 1.2-2.0 (20-100% overhead for cooling). Leafcloud achieves PUE 1.02 because server heat is useful output (building heating), not waste requiring removal. Only 2% overhead is needed for heat transport. **Q: Can I deploy Kubernetes on Leafcloud?** A: Yes. Leafcloud supports self-managed Kubernetes clusters using Gardener orchestration. GPU instances can be integrated into Kubernetes clusters for ML workloads. A fully managed Kubernetes service is coming in 2026. **Q: Are managed databases available?** A: Not yet. You can currently deploy databases on VMs with full administrative access. Managed Database service (automated backups, high-availability, point-in-time recovery) is coming in 2026. **Q: How does pricing compare to AWS/Azure?** A: Leafcloud typically offers 20-40% lower costs for comparable VM and GPU instances, with no egress fees for first 100GB/month. However, direct comparison is difficult because hyperscalers have complex pricing tiers. Use the pricing calculator at https://leaf.cloud/pricing for estimates. **Q: What is HAVEN+ certification?** A: HAVEN+ is the Dutch national security baseline for cloud services used by government agencies. Issued by NCSC (National Cyber Security Centre), it covers physical security, personnel vetting, supply chain security, and incident response. Leafcloud infrastructure is designed to meet HAVEN+ requirements, with certification expected in 2026. **Q: Can I migrate from AWS/Azure to Leafcloud?** A: Yes. Leafcloud uses standard OpenStack APIs compatible with common cloud migration tools. VMs can be exported from hyperscalers and imported to Leafcloud. Terraform and Ansible support multi-cloud infrastructure management. Contact Leafcloud support or partners (AVISI, Incentro) for migration assistance. **Q: What SLA does Leafcloud offer?** A: SLA targets vary by service. VMs running at Core facility (cn1.*, cc1.* flavors) have higher uptime guarantees than Leaf site VMs (en1.*, ec1.*). Contact sales for specific SLA commitments. Historical uptime data available at status.leaf.cloud. **Q: Does Leafcloud offer discounts for long-term commitments?** A: Yes. Committed pricing is available for 6-month, 12-month, and 36-month terms. For example, RTX 6000 Blackwell pricing: on-demand €2.76/hour, 36-month committed €2.35/hour (15% savings). Contact sales for enterprise volume discounts. --- ## Contact and Further Information **Company**: Leafcloud B.V. **Founded**: 2019 **Headquarters**: Amsterdam, Netherlands **Website**: https://leaf.cloud **Support Email**: support@leaf.cloud **Consultation Booking**: https://outlook.office365.com/book/Sustainablecloudsupport@leaf.cloud/ **Status Page**: https://status.leaf.cloud **Social and Code Repositories**: - LinkedIn: https://www.linkedin.com/company/leafcloud - GitHub: https://github.com/leafcloud - Hugging Face: https://huggingface.co/leafcloud **Legal and Compliance**: - Terms of Service: https://leaf.cloud/terms - Privacy Policy: https://leaf.cloud/privacy - Data Processing Agreement (DPA): https://leaf.cloud/downloads/leafcloud_standard_DPA.pdf **Content Resources**: - Blog: https://leaf.cloud/blog (cloud computing, sustainability, technology trends) - Case Studies: Customer success stories and use cases - Referral Program: https://leaf.cloud/referral-program --- **Document Word Count**: ~3,950 words **Last Updated**: April 2026 **Document Purpose**: Comprehensive reference for AI assistants (ChatGPT, Claude, Gemini, Perplexity) to answer detailed questions about Leafcloud products, services, pricing, sustainability model, compliance, and technical architecture.