How vCluster Labs and Netris are transforming GPU cloud architecture with hard multi-tenancy

At NVIDIA’s GTC 2025 conference, vCluster Labs (formerly LoftLabs) and Netris announced a partnership that could reshape how enterprises manage AI infrastructure. The two companies introduced what they describe as the first “full-stack Kubernetes multi-tenancy” platform built specifically for GPU cloud providers and enterprise AI factories.

By combining Kubernetes-level and network-level isolation, the collaboration automates the deployment of secure, scalable, and compliant AI environments—reducing the complexity of managing high-performance GPU clusters. This joint approach addresses one of the industry’s biggest pain points: how to maintain hard tenant isolation while maximizing GPU efficiency.

The vCluster–Netris integration unites software and network layers into one orchestrated workflow, allowing enterprises to deploy isolated Kubernetes clusters across private data centers, edge sites, and large GPU facilities. The result is faster provisioning, greater control, and improved GPU utilization for AI workloads that demand both flexibility and security.

Why is full-stack Kubernetes multi-tenancy crucial for AI infrastructure in 2025?

The surge in enterprise AI adoption has exposed a critical need for secure, high-performance, and cost-effective GPU infrastructure. Public cloud hyperscalers offer scalability but often come with trade-offs in data control, compliance, and cost. Many enterprises are now turning to hybrid and private GPU clouds, known as AI factories, to balance flexibility and sovereignty.

Multi-tenancy lies at the heart of these environments. Traditional Kubernetes implementations separate workloads logically but rely on complex, manual network segmentation to enforce security. This can lead to inefficiency and slower scaling.

vCluster Labs and Netris have solved this by merging Kubernetes virtualization with automated network-level isolation. The integration creates true hard multi-tenancy, ensuring each tenant has its own isolated compute and network resources. This makes it possible to safely share GPUs among multiple users or departments while keeping workloads fully secure and compliant.

The timing is ideal. As AI models grow in size and cost, GPU utilization efficiency has become a major economic driver. The full-stack solution allows organizations to extract maximum performance from every GPU, directly impacting ROI.

How does the vCluster Labs and Netris integration work across compute and network layers?

At the compute level, vCluster Labs provides Kubernetes virtualization that automates the creation and scaling of tenant clusters. Each virtual cluster behaves like a standalone Kubernetes environment while leveraging shared infrastructure resources, significantly reducing overhead.

At the network level, Netris automates hard isolation through its Network Automation and Multi-Tenancy (NAAM) platform. It dynamically provisions isolated network paths for each tenant across diverse fabrics such as Ethernet, InfiniBand, NVLink, DPUs, and edge nodes.

This integration delivers a unified automation pipeline—creating new tenant environments in minutes, with end-to-end segmentation and compliance built in. It replaces legacy manual scripts and VLAN-based management with declarative automation validated for production use.

According to vCluster Labs Chief Executive Officer Lukas Gentele, the collaboration represents “a full-stack approach to multi-tenancy for AI operators,” extending Kubernetes simplicity to the network layer.

What advantages does this offer for GPU cloud providers and AI factories?

The benefits for GPU cloud operators and AI infrastructure teams are multifold. First, it accelerates time-to-revenue by transforming idle hardware into productive resources. Operators can launch new tenants quickly and securely, enabling faster customer onboarding and workload execution.

Second, it enhances GPU utilization. Hard isolation ensures GPUs can be dynamically reassigned between tenants without risking interference or compliance breaches.

Third, it reduces operational complexity and risk. Automation replaces manual configuration, eliminating errors that commonly occur when managing complex multi-fabric networks.

Netris Chief Executive Officer Alex Saroyan emphasized that “hard multi-tenancy” is essential for predictable performance at scale. He explained that the joint platform allows operators to “run securely, onboard instantly, and keep every GPU productive.”

How does this partnership reflect broader trends in AI infrastructure and GPU markets?

The collaboration mirrors a larger transformation in how enterprises build and operate AI infrastructure. As organizations move away from cloud-only deployments, hybrid and edge-based GPU factories are becoming the norm.

This shift is driven by the rising cost of public cloud GPUs and the need for compliance with local data regulations. The vCluster–Netris solution offers a way for companies to combine cloud agility with on-premises control—bringing hyperscaler-grade efficiency to private environments.

Industry analysts see this as part of a decentralization trend in AI infrastructure. Enterprises such as Adobe, Deloitte, and CoreWeave have already embraced virtualized Kubernetes models for AI workloads. The addition of network automation at this scale could accelerate that momentum further, enabling new players to compete with established GPU cloud providers.

What could this mean for enterprise adoption and scalability?

For enterprises, the joint platform simplifies operations while ensuring full compliance. Platform teams can deploy secure, isolated AI environments faster and more reliably. The architecture’s scalability supports dynamic GPU allocation and real-time scaling for fluctuating AI workloads—from training large models to inference pipelines.

The impact extends beyond technical performance. By lowering the complexity barrier, more organizations can participate in the AI infrastructure economy. Some may even repurpose excess GPU capacity as rentable AI cloud resources, creating new revenue streams.

Analysts expect future updates to include deeper integration with observability tools, enhanced policy enforcement, and broader support for edge AI workloads—making the partnership a foundational layer for next-generation AI operations.

What should infrastructure teams and investors watch next?

The next milestones will involve real-world validation. Case studies demonstrating improved GPU utilization rates, reduced deployment time, and measurable cost efficiencies will help quantify the benefits of the platform.

Competitors in the AI infrastructure ecosystem may respond with similar integrations, and potential acquisitions could follow if hyperscalers or enterprise cloud vendors seek to strengthen their multi-tenancy capabilities.

For investors, the key takeaway is that the AI infrastructure market is entering a phase where orchestration and automation are strategic assets. As AI scales, the winners will be those who can deliver both efficiency and isolation at every layer of the stack.

Key takeaways from the vCluster Labs and Netris partnership

• The partnership delivers the first full-stack Kubernetes multi-tenancy model integrating both compute and network isolation for AI infrastructure.

• Enterprises can automate secure and scalable GPU cloud environments across private data centers, edge sites, and multi-node clusters.

• The collaboration eliminates manual network configuration by uniting vCluster Labs’ Kubernetes virtualization with Netris’s NAAM automation.

• It enables faster tenant onboarding, higher GPU utilization, and stronger compliance enforcement.

• The partnership reflects the growing trend toward hybrid and private GPU clouds as enterprises seek control beyond hyperscaler platforms.

• Market observers view this as part of a broader decentralization in AI infrastructure—where on-premises automation rivals public cloud scalability.

• Future deployments and performance metrics will determine its adoption trajectory, with potential acquisition interest from major cloud or GPU players.