We are thrilled to introduce the OpenNebula 7.4 ‘Helix’ Beta. While previous updates focused on infrastructure scaling, this milestone release shifts the spotlight to absolute operational efficiency, an entirely reimagined user experience, and sophisticated automation for complex enterprise and AI workloads.

As a beta release, this version is intended exclusively for sandbox testing and validation. Please help us polish the final release by exploring the new features, reviewing the known issues, and reporting any bugs via GitHub.

Next-Generation User Experience

• Redesigned Sunstone Interface: A completely overhauled GUI providing cloud administrators with a sleek, modern, and highly intuitive user experience to streamline large-scale infrastructure management. 

Optimized Enterprise Workload & Resource Management

• OneSwap Batch VMware Migrations: Accelerate your transition away from legacy hypervisors with automated, bulk VMware migration capabilities designed to minimize friction.
• Dynamic VM Group Management: Manage VM groups dynamically directly through the CLI, increasing flexibility for complex, multi-tier application architectures.
• Improved OneDRS Datastore Selection: Enhanced migration logic automatically identifies optimal target datastores based on real-time space and usage requirements.
• Bulk Service VM Operations: Perform bulk deletions of scheduled actions across all service VMs, drastically reducing manual administrative overhead.

High-Performance Networking and Multi-Tenancy

• VLAN Group-Level Authorization: Delegate VLAN management directly to tenants, enabling secure Virtual Network self-provisioning in multi-tenant environments.
• Open vSwitch with Switchdev Mode: Native support for SR-IOV-capable PCI network interfaces running in Switchdev mode with Open vSwitch, unlocking maximum network performance for demanding workloads.

Advanced Backup & Storage Capabilities

• Expanded Backups via OneBEX: Leverage the OpenNebula Backup Exporter (OneBEX) to remove heavy Veeam integration storage requirements. Third-party tools can now interactively pull incremental changes on virtual disks (supporting both qcow2 and LVM-based disks). 

Expanded GPU and AI Accelerator Support

• Axelera AI Accelerator Integration: OpenNebula now supports the integration and lifecycle management of Axelera AI accelerators, enabling their allocation to virtual machines for AI inference and edge AI workloads.

Enterprise Operating System & Appliance Updates

• Ubuntu 26.04 Support: Production-ready packages fully supporting Ubuntu 26.04 (Resolute Raccoon) are now available.
• Enhanced Slurm Appliance: Deploy clusters for AI and HPC workloads effortlessly with out-of-the-box NVIDIA GPU support and native LDAP identity integration.
• Targeted Stability Fixes: A wide range of under-the-hood performance enhancements across HA, networking, storage, monitoring, and vCenter compatibility.

Expanding the Ecosystem: Advanced Enterprise Extensions

Available via the Enterprise Subscription, these extensions are designed for enterprise cloud and AI Factory deployments. They include:

• Automated AI Factory-Scale BMaaS (NVIDIA NICo): First version of native integration with the NVIDIA Infrastructure Controller (NICo) delivers automated Bare Metal as a Service (BMaaS) and seamless orchestration for GPU-heavy environments.
• Enhanced Kubernetes Provisioning (OneKS): Streamline container orchestration with multi-cluster configuration options and pre-deployment diagnostics designed to catch errors before provisioning failures occur.

Built for the Modern Enterprise Stack

OpenNebula 7.4 bridges the gap between modern usability and deep hardware automation. From a fresh visual interface to powerful backup and migration toolsets, 7.4 ensures your private, hybrid, or sovereign AI cloud is future-proof.

Dive into the new features, connect with the community, and help us refine the final release. Thank you to our incredible partners and users for your continued support.