Products & Tools

Hitachi Vantara Update Lets VSP One Query Data In Place, Up to Five Times Faster

Hitachi Vantara says VSP One's new in-place querying delivers analytical queries up to five times faster on 80% fewer CPUs, as it targets enterprises moving AI from pilots to production.

Hitachi Vantara speeds AI analytics with VSP One update
Hitachi Vantara speeds AI analytics with VSP One updateMarc PEZIN / unsplash
By Rebecca Stone7 min read

Why it matters

  • Hitachi Vantara says VSP One's direct querying of Iceberg, Parquet, JSON and CSV data delivers analytical queries up to 5x faster on 80% fewer CPUs.
  • VSP One Object now ships in a four-node starting configuration and offers up to 5.5x faster small-object access for AI, analytics and S3 workloads.
  • Hitachi iQ Studio adds Model Gateway, Model Context Protocol connectivity, ServiceNow support and role-based access control.
  • Chief Product Officer Octavian Tanase said: 'AI strategy is now inseparable from data strategy.'
  • The update was announced on 18 November 2026.

Hitachi Vantara says analytical queries on its VSP One storage platform now run up to five times faster on 80% fewer CPUs, after an update announced on 18 November 2026 that lets organisations query open-format data where it sits instead of moving it into separate analytics environments.

The change targets the most persistent bottleneck in enterprise AI: getting data from where it is stored to where it is processed. The update supports direct queries against Iceberg, Parquet, JSON and CSV data held on the system, removing a step that has historically required dedicated data pipelines and duplicate datasets. For companies building AI and analytics workloads, Hitachi Vantara positions this as a way to shorten the distance between storage and insight.

The announcement pairs the querying capability with a set of storage, availability and governance changes across VSP One and Hitachi iQ Studio, the company's software for building and managing AI applications and agents. Together, the releases aim at a specific market moment: organisations that have moved past AI pilots and are trying to scale the technology across the enterprise without rebuilding their data estates first.

What does the update change on the storage side?

The core technical claim is in-place querying. VSP One can now run analytical queries directly against open-format datasets stored on the platform. According to Hitachi Vantara, this supports queries that are up to five times faster while consuming 80% fewer CPUs than the alternative of moving data into a separate analytics environment.

The company frames the benefit in two ways:

  • Reduced dataset duplication, because data no longer needs copies in an analytics tier
  • Fewer supporting data pipelines, which cuts operational overhead and the risk of stale or inconsistent data

Hitachi Vantara positions VSP One as a unified data platform with a common control plane across its storage systems. Within that architecture, VSP One Object serves as the object storage layer for AI and analytics workloads. The new in-place querying function extends that role from holding data to serving queries against it.

The platform also gained small-object performance improvements. Hitachi Vantara says these support up to 5.5 times faster access to data for AI, analytics and S3 workloads. Small-object performance matters for AI workloads because training and retrieval systems frequently read large numbers of small files rather than a small number of large ones.

For availability, VSP One Object added multisite replication and regional features. Hitachi Vantara describes these additions as aimed at improving resilience and data availability, an acknowledgment that AI workloads promoted into production carry the same uptime expectations as other business-critical systems.

Why a four-node entry point?

Alongside the performance work, Hitachi Vantara introduced a four-node starting configuration for VSP One Object. The smaller entry point is intended to make deployment easier for organisations building storage for AI and analytics use cases.

The move lowers the floor for adoption. Object storage systems have typically demanded larger initial footprints, which puts them out of reach for departments or mid-sized organisations running early AI projects. A four-node configuration lets a customer stand up an AI-ready storage layer without committing to the scale a full production deployment requires, then expand as workloads grow.

That framing matches the broader market context Hitachi Vantara cites for the release. The company pointed to industry research showing that while many organisations now use AI regularly in at least one business function, fewer have scaled it across the enterprise. Infrastructure cost and complexity sit among the reasons. A smaller starting configuration attacks one part of that problem directly.

What is new in Hitachi iQ Studio?

Hitachi iQ Studio sits on top of the VSP One storage layer as the software element for building, deploying and governing AI systems. The latest version expands the data and governance features in three main directions.

First, it broadens access to structured and unstructured enterprise data sources for retrieval-augmented generation, including filesystem and NFS knowledge sources. Retrieval-augmented generation, or RAG, is the technique by which AI models draw on internal company information to answer questions or support workflows. By extending access to additional file and network storage sources, Hitachi Vantara aims to make more of that information available to AI applications under managed conditions, without requiring companies to move or restructure all of it first.

Second, the release adds what Hitachi Vantara calls Model Gateway, along with Model Context Protocol connectivity. These components are designed to give organisations more control over how AI applications and agents connect to models and tools.

Third, it adds governance and operational capabilities:

  • ServiceNow support
  • Role-based access control
  • Observability
  • Lifecycle management

Each addition addresses a distinct failure mode in enterprise AI deployments. Role-based access control governs who can reach which data. Observability surfaces how applications and agents behave in production. Lifecycle management covers the trajectory from build to retirement. ServiceNow support connects AI workflows to an operations platform that large enterprises already run.

The governance emphasis reflects the second obstacle Hitachi Vantara identifies in moving AI to production: the state of underlying corporate data. The company pointed to research indicating that data-quality problems remain widespread, complicating efforts to connect AI systems and agents to trusted information sources. Tightening controls over data access and model connections is the vendor's structural answer to that problem, rather than a one-off data-cleaning exercise.

What did Hitachi Vantara's leadership say?

Octavian Tanase, Chief Product Officer at Hitachi Vantara, said the changes reflect the growing link between infrastructure decisions and AI plans inside large organisations.

"AI strategy is now inseparable from data strategy. As AI becomes part of more business-critical workloads, the infrastructure underneath it has to make it easier to put enterprise data to work," Tanase said.

"With VSP One, we're giving customers a practical way to build the data foundation for AI, bring analytics closer to their data and scale with confidence as demand grows. That creates a stronger path from experimentation to production without sacrificing the resilience and control enterprises depend on," he added.

The two statements compress the company's argument. The first asserts that storage decisions are now AI decisions. The second claims that VSP One lets customers act on that premise, moving from experimentation to production while keeping the resilience and control that enterprise IT requires.

Why does this matter for the enterprise AI market?

The release lands in a market where the gap between AI experimentation and AI production remains wide. Hitachi Vantara's cited research describes that gap: regular AI use in at least one business function is common, while enterprise-wide scale is rare. Infrastructure vendors are competing to close it, and the competition increasingly turns on who can reduce data movement, widen governed access to enterprise information and control how AI agents interact with models and tools.

Hitachi Vantara's structural answer is a single stack. VSP One provides the storage foundation for Hitachi iQ, linking data infrastructure, analytics and AI software, according to the company. The latest changes focus on exactly the three pressure points the market research identifies: reducing data movement through in-place querying, widening access to enterprise information through expanded RAG sources in iQ Studio, and tightening governance through Model Gateway, Model Context Protocol connectivity, role-based access control, observability and lifecycle management.

The performance figures anchor the commercial case. Queries up to five times faster on 80% fewer CPUs translate directly into lower compute spend and faster analytical cycles. Small-object access up to 5.5 times faster shortens the data-retrieval path for AI, analytics and S3 workloads. The four-node entry point reduces the initial investment required to start.

The stakes extend beyond one vendor's product line. As AI workloads move into business-critical roles, storage platforms that merely hold data are giving way to platforms that serve it to models and queries directly. Hitachi Vantara's update is a bid to occupy that position, and the direction of travel it signals, less data movement, more governed access, tighter control over agent-to-model connections, will shape procurement decisions for enterprises building AI infrastructure in the coming cycle.

Source: itbrief.com.au

Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.

728 articles

Related articles

  1. Hitachi Vantara expands VSP One for governed AI data access
  2. Hitachi Vantara pushes AI analytics into storage with VSP One update
  3. Hitachi Vantara Brings Lakehouse Queries Directly Onto VSP One Object Storage
  4. Confluent launches Streamhouse and Lightning Tables for AI agents
  5. IBM Targets Mainframe Modernization With Z Atlas AI Package

« Previous articleNext article »