Nutanix and ChronoScale have announced a strategic partnership to build enterprise AI infrastructure together, pairing Nutanix’s hybrid cloud software with ChronoScale’s accelerated computing, GPU as a service (GPUaaS), and AI services. The deal also folds in NVIDIA-based technology, though a lot of what was announced is still in planning or development.
The Nutanix-ChronoScale alliance in 20 seconds
- ChronoScale plans to use Nutanix software to manage infrastructure, Kubernetes, and AI services.
- Customers will be able to mix on-premises infrastructure with external GPU capacity on demand.
- ChronoScale Foundry can run inside the customer’s environment, keeping agents and data local.
- The planned infrastructure includes NVIDIA HGX B300 systems and Spectrum-X networks.
- Several announced integrations aren’t commercially available yet.
The pitch targets a common hurdle when companies move AI from testing to production: getting GPUs, on-premises infrastructure, management tools, and a platform to oversee workloads all at once.
ChronoScale mainly brings the accelerated computing, while Nutanix supplies much of the software layer to manage it. The two also plan to run shared demo and proof-of-concept environments for enterprise customers.
On-demand GPUs to extend on-premises infrastructure
One of the more interesting parts of the deal is the option to extend a Nutanix installation on the customer’s premises or in their data centers using GPU resources from ChronoScale.
The company offers two modes.
For predictable workloads, there’s reserved capacity through the GPUaaS service. For ad-hoc or experimental needs, it points to ChronoScale Token Factory, built on prepaid inference tokens and open-source models.
The plan is to tie both services into Nutanix’s enterprise AI offerings, including Agent Gateway and Private Inference.
The technical aim is a shared control plane for local resources and the capacity ChronoScale provides. If it works as planned, an organization could run certain applications and data on its private infrastructure and reach for extra GPU capacity when demand spikes.
That model fits the rise of hybrid AI. Deploying large numbers of accelerators costs so much that not every company can justify keeping that capacity on site all the time.
At the same time, some applications need data residency, security, or control that make moving all the processing to outside infrastructure impractical.
Combining both environments is meant to cover that middle ground.
ChronoScale Foundry can run inside the customer’s environment
The second piece is ChronoScale Foundry, an enterprise platform for building, running, and managing AI agent workflows.
Nutanix plans to let it deploy inside the customer’s own infrastructure.
According to the partners, agents, enterprise data, and workflow states stay within the organization’s controlled boundaries. The platform will also plug into the Nutanix Kubernetes Platform catalog to speed up deployment.
That addresses a growing worry in enterprise generative AI: a capable model for text or code isn’t enough on its own.
Agents need access to internal applications, databases, documents, and other systems to get work done, which raises the bar on permissions, isolation, auditing, and data control.
Keeping that layer inside private infrastructure can matter a lot for companies handling sensitive information or bound by strict data governance and sovereignty rules.
Keep in mind that not everything mentioned is available yet. Nutanix states plainly that many of the products, integrations, and features are still in planning, development, testing, or rollout, and that timing can change.
NVIDIA will play a big role in the accelerated infrastructure
The collaboration also leans on both companies’ ties to NVIDIA.
ChronoScale is an NVIDIA Cloud Partner, and Nutanix acts as a technology partner and independent software vendor with validated solutions for enterprise AI infrastructure.
ChronoScale plans to deploy NVIDIA HGX B300 systems connected over NVIDIA Spectrum-X. Its platform will also use NVIDIA AI Enterprise, including NIM microservices and NVIDIA NeMo.
That choice mirrors how enterprise AI projects are shifting from one-off accelerator purchases to full architectures with GPU servers, storage, high-speed networking, Kubernetes, models, and management tools.
Nutanix wants a big slice of that management layer.
Over the past few years, Nutanix has stretched beyond hyperconverged infrastructure and hybrid cloud into tools for deploying and running AI workloads.
ChronoScale supplies the other half: accelerated capacity you can reach without every customer having to physically install all the GPUs they might need at peak demand.
The two also expect joint sales, technical collaboration, and shared solution development. Their targets include Global 2000 organizations and large enterprises after AI infrastructure and services with more control over their data.
There is an important caveat, though. The announcement says the collaboration sets up a framework expected to be formalized through one or more definitive agreements. So it’s worth separating the planned scope from what customers can actually use today.
The approach points to a model that keeps gaining ground in enterprise infrastructure: hold data, applications, and some AI workloads in a private cloud, and connect to large external pools of accelerators when capacity needs climb.
Frequently Asked Questions
What have Nutanix and ChronoScale announced?
A partnership that combines Nutanix’s hybrid cloud and infrastructure software with ChronoScale’s accelerated computing and AI services.
Will you be able to use external GPUs from a Nutanix setup?
That’s one of the announced goals. ChronoScale intends to provide capacity through GPUaaS, while the two companies work on integration to manage local and external resources from one environment.
What hardware will ChronoScale use?
It plans to deploy NVIDIA HGX B300 systems with NVIDIA Spectrum-X networks, plus NVIDIA AI Enterprise components such as NIM and NeMo.
Are all these features available now?
No. Several integrations and services are still in planning, development, testing, or rollout. Nutanix stresses that availability and timing can change.
via: nutanix
