Virtuozzo unveils AI infrastructure system for GPUs

Virtuozzo has unveiled the Virtuozzo Infrastructure System for AI infrastructure, combining compute, storage and networking in a single architecture.

The hyperconverged stack also includes an operating system, orchestration, management, automation and protection tools. It is aimed at organisations delivering AI and GPU-based services, as well as businesses running AI training and inference workloads.

The launch outlines a full-stack approach at a time when companies are trying to control AI deployment costs and make better use of scarce GPU resources. Virtuozzo argues that tighter integration across infrastructure layers can reduce operational complexity and lower total cost of ownership compared with rival setups.

At the centre of the system is V/OS, a Linux-based operating system designed to support both virtual machines and system containers. It is paired with V/Orchestration for coordinating compute, storage, networking and Kubernetes environments, and V/Management for central oversight of infrastructure and services.

Other components include V/Automation for provisioning, billing and customer management, and V/Protection for backup, disaster recovery and cybersecurity. The aim is to remove the need for separate external systems and fragmented infrastructure tools.

Chief Executive Officer Kurt Daniel said AI now influences how the company develops and runs its products.

"AI is pivotal to how we develop our product today, how it operates and how our users benefit from it," said Kurt Daniel, Chief Executive Officer of Virtuozzo.

He also outlined the company's case on cost and hardware efficiency.

"Our Infrastructure System is co-designed to ensure maximum performance and efficiency from modern hardware, helping our customers and partners run and monetize AI with better utilization, lower overhead and 60-80% less total cost of ownership (TCO) with a typical hosting setup versus competitive offerings," said Daniel.

Customer view

Virtuozzo also included backing from Leapswitch, which uses its technology to supply GPU services. The customer comment highlights one of the launch's main commercial themes: faster rollout of on-demand AI infrastructure with usage-based pricing.

"Virtuozzo enabled us to meet growing AI demand by bringing GPU services to market much faster in a flexible and scalable way," said Ishan Talathi, Chief Executive Officer of Leapswitch.

Talathi said pricing and tenancy features have shaped the economics of that offering.

"With usage-based pricing and multi-tenant capabilities, we can offer on-demand GPUs while keeping costs under control and staying highly competitive with hyperscalers. It's the combination of efficiency, simplicity and quick time to value that really makes the difference," said Talathi.

The system is intended to serve a wide range of users, from early-stage AI businesses to larger production environments. Integrated metering and billing are intended to help providers turn AI infrastructure into commercial services more quickly.

Virtuozzo is drawing on a long history in container technology. It said it holds 132 patents and described its engineering approach as closely tied to AI and open source development.

The latest product release also adds GPU metering and management to the company's admin panel. That points to a broader effort by infrastructure suppliers to give service providers more precise visibility into GPU consumption as demand for AI computing grows.

Virtuozzo also linked its pitch to wider changes in the virtualisation market, saying customers are under pressure from rising hardware costs and VMware price increases following Broadcom's acquisition. It positioned infrastructure efficiency as a way to offset those pressures.

The system is available to partners and customers worldwide.