Hitachi iQ upgrade targets on-prem agentic AI at scale

Hitachi Vantara has expanded its Hitachi iQ portfolio with new software capabilities, additional NVIDIA infrastructure options, and tighter data integration aimed at running agentic AI in on-premises and virtualised environments.

The update adds enhanced AI blueprints and multi-agent coordination in Hitachi iQ Studio, along with broader support for new NVIDIA GPU platforms. It also deepens integration with Hammerspace for data access and orchestration within the iQ Studio environment.

Enterprises have increased investment in AI pilots over the past two years, but many have struggled to move systems into production. As organisations look to run AI against business data, data governance, sovereignty requirements, and security controls have become more prominent.

Hitachi Vantara pointed to research suggesting many organisations still lack mature data foundations. A recent report found only 42% of organisations in the US and Canada were considered data-mature. It also reported that 84% of data-mature organisations saw measurable AI return on investment, compared with 48% of organisations with weaker data foundations.

Octavian Tanase, chief product officer at Hitachi Vantara, said the new iQ capabilities are aimed at customers moving from experimental use to deployed AI systems.

"AI is moving into production faster than many organisations' data foundations are ready to support. With these latest enhancements to the Hitachi iQ portfolio, we are expanding across software innovation, high-performance infrastructure and intelligent data integration to give customers greater flexibility and control as they move agentic AI from pilot to production."

Infrastructure options

Hitachi iQ combines compute, networking, and storage in a validated stack built on Hitachi Vantara's Virtual Storage Platform One data platform. Hitachi also positions the product line alongside HMAX by Hitachi, a suite of next-generation solutions focused on applying AI to social infrastructure.

On the hardware side, Hitachi iQ now supports NVIDIA Blackwell GPUs in air-cooled configurations and NVIDIA Blackwell Ultra GPUs in both air-cooled and liquid-cooled versions. The portfolio also includes a 2U NVIDIA MGX-based system with up to four NVIDIA RTX PRO6000 Blackwell Server Edition GPUs.

Hitachi iQ is also expected to add support for the NVIDIA RTX PRO 4500 Blackwell Server Edition GPU. The aim is to offer options for different AI workloads-including model development, fine-tuning, inference, and agentic applications-while addressing constraints such as cooling, power, and space.

Hitachi says the design keeps data close to compute, improving utilisation and efficiency for data-intensive AI workloads.

Studio updates

Hitachi iQ Studio is the software component of the iQ portfolio, used to design, deploy, and govern AI agents within enterprise environments. It is built on the NVIDIA AI Data Platform reference design.

The new release introduces expanded AI blueprints and multi-agent coordination, adding more structure for teams moving from prototypes to deployed systems.

A key addition is a blueprint model with defined roles for different agents: supervisor models and worker models. Worker agents execute tasks, while supervisor agents coordinate multi-agent workflows and adjust based on outcomes.

iQ Studio also expands support for NVIDIA Nemotron models, which NVIDIA positions for tool-using agentic AI systems. Another addition is a "time machine" function, which Hitachi says lets AI systems navigate historical datasets with more context and speed.

Data integration

Hitachi iQ now includes tighter integration between Hitachi iQ Studio and Hammerspace. The integration provides access-within iQ Studio-to data managed by Hammerspace via Model Context Protocol, an open standard for connecting AI systems to external data sources.

The approach is intended to let customers build AI agents in iQ Studio that work with Hammerspace-managed data while keeping governance within VSP One. It also reduces the need to relocate data and improves observability across distributed datasets.

Hitachi also highlighted VSP One Block infrastructure as part of the stack supporting the data layer, citing consistent performance and 100% data availability while maintaining hybrid cloud flexibility.

Jason Hardy, vice president of storage technologies at NVIDIA, said consistent software and trusted data are central to scaling AI in enterprise settings.

"As enterprises continue to scale AI, the ability to combine accelerated computing with consistent software and trusted data becomes essential. Full-stack AI infrastructure optimised for enterprise demand enables organisations to support a wider range of AI outcomes while maintaining the performance, governance, and operational consistency enterprises require."

Storage roadmap

Alongside the iQ changes, Hitachi Vantara said it will support the NVIDIA STX reference architecture as part of its work on what it calls AI-native storage. It said the architecture is powered by NVIDIA Vera Rubin, BlueField-4, and Spectrum-X networking, alongside NVIDIA AI software.

Hitachi Vantara plans to showcase Hitachi iQ and Hitachi iQ Studio at NVIDIA GTC 2026 in San Jose, presenting the latest updates and outlining next steps for the iQ portfolio.