Nebul Becomes First Neocloud to Run NetApp Novus Storage

Nebul, a European AI-native cloud provider, has become the first neocloud to adopt NetApp's Novus storage architecture for sovereign AI infrastructure, NetApp announced on November 10, 2026. The architecture is generally available and designed for up to 100 TB per second of aggregate throughput. Nebul will deploy Novus inside its sovereign Intelligent Cloud platform.

Nebul describes Intelligent Cloud as a sovereign platform that combines accelerated infrastructure, enterprise data services and production AI capabilities while preserving customer ownership and control. Beyond the deployment itself, the two companies are exploring ways to accelerate sovereign-AI uptake by pairing Nebul's services with the NetApp Platform.

For Nebul's multi-tenant environment, NetApp says Novus can scale performance, capacity and concurrency independently beneath a single namespace as demand grows. Nebul founder and CEO Arnold Juffer said: "With this architecture, we can scale our AI Factory to serve growing customer demand without sacrificing the performance, security or compliance they depend on."

How does Novus separate metadata from data?

Novus splits metadata — the information that maps file locations — from the movement of the data itself. In NetApp's technical explanation, a client requests a layout from the Novus Data Director, then communicates directly with ONTAP-based storage. Data Director is an independently scalable, software-defined control layer running on standard x86 compute, which lets the metadata and data planes scale separately. That separation addresses a familiar bottleneck: small-file metadata operations and large checkpoint writes competing for the same controllers.

The product documentation says the system works with standard in-kernel NFSv4.2 and pNFS Flex Files clients using nconnect, plus NVIDIA GPUDirect Storage. Operators don't need custom client software across thousands of GPU servers. Additional ONTAP clusters add throughput without remounts or fragmentation, while every GPU continues to see the same namespace.

What does the ONTAP foundation carry over?

The ONTAP foundation also handles the data-governance side of the deal. NetApp says the design lets organizations extend existing data estates into sovereign cloud environments while preserving consistent security, protection and governance, including current jurisdiction and access policies. Willem Hendrickx, NetApp's senior vice president and general manager for EMEA and LATAM, named government, healthcare and biotech as regulated industries where providers must address compliance, security and multi-tenancy.

From product launch to general availability

NetApp introduced Novus on September 29, 2026, and made the architecture orderable that same day. The initial release combines Data Director software on qualified Supermicro infrastructure with ONTAP data services on AFF A90 systems. NetApp laid out a path toward software-defined deployments over time while maintaining the same namespace, operating model and data foundation.

Why storage throughput drives GPU economics

NetApp's economic argument is blunt: GPUs waiting on storage still depreciate without generating a return. At the September launch, the company said utilization can fall below 30% under traditional architectures when storage can't supply data fast enough. That framing makes storage throughput a direct constraint on GPU-cluster economics.

NetApp technical fellow Arindam Banerjee quantified the bandwidth demand: a single GPU can require as much as 2 GB per second, and 50,000 GPUs may need that rate simultaneously. He put conventional storage arrays at 40–80 GB per second. Reaching Novus's aggregate target with such arrays would take more than 100 systems — each adding a namespace, a failure domain and management overhead. That comparison ties the single-namespace design directly to the operational cost of scaling storage.

For Nebul, the adoption will test whether independently scaling metadata and data can support more tenants and more GPUs without fragmenting data across storage systems, while jurisdiction, access and governance controls stay consistent. The operational challenge ahead: expand shared AI infrastructure without weakening the isolation and control that regulated customers require.