Dell and NVIDIA’s Strategy to Expand AI Data Centre Stack

Dell Technologies is extending its AI Factory with NVIDIA, placing data centre infrastructure at the centre of enterprise AI deployment.

The latest updates focus on how organisations build, run and scale AI workloads across on-premise environments, with particular attention on servers, networking and integrated systems.

First introduced in March 2024, the Dell AI Factory with NVIDIA now reaches more than 4,000 customer deployments.

The platform combines NVIDIA GPUs, networking and software to support workloads such as retrieval-augmented generation – a method that enhances AI responses using external data – and multimodal search, which allows systems to process different types of data including text and images.

The latest updates are structured across four areas, spanning desktop development through to full-scale data centre operations.

Data centre infrastructure for AI workloads

At the core of the announcement is Dell’s continued investment in PowerEdge servers, which form the backbone of AI deployments inside data centres.

These systems are designed to handle both training and inference – the two key stages of AI workloads, where models are first built and then applied in real time.

PowerEdge XE9812 integrates the NVIDIA Vera Rubin NVL72 platform, enabling high-performance processing for large-scale AI tasks.

Alongside this, the XE9880L, XE9882L and XE9885L servers use the NVIDIA HGX Rubin NVL8 architecture and incorporate liquid cooling to manage heat generated by dense GPU configurations.

Liquid cooling is increasingly important in modern data centres, as it allows operators to run high-performance systems more efficiently without increasing overall power consumption.

This approach supports AI workloads that require continuous processing while maintaining operational stability.

Dell is also introducing the PowerEdge R770, R7715 and R7725 servers, which allow organisations to integrate AI capabilities into existing infrastructure.

By adding NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, these systems bring accelerated computing into more traditional enterprise environments without requiring a complete rebuild of data centre architecture.

Networking and high-performance connectivity

As AI workloads grow, networking becomes a critical component of data centre design.

Moving large volumes of data quickly between servers is essential for both training and inference tasks.

Dell’s PowerSwitch SN6000-series addresses this requirement, delivering speeds of up to 1.6Tbs and supporting liquid cooling.

These switches connect AI clusters and enable high-throughput data transfer across the data centre.

Additional flexibility comes from the PowerSwitch SN5610 and SN2201, which support multiple network operating systems including Cumulus Linux and Enterprise SONiC Distribution by Dell Technologies.

This allows data centre teams to tailor network management to their operational needs.

NVIDIA Quantum-X800 InfiniBand Q3300-LD switches are also part of the ecosystem.

InfiniBand is a high-speed networking technology designed for low latency and high bandwidth, making it suitable for AI and cloud environments where rapid data exchange is required.

Together, these networking solutions ensure that compute performance is matched by data movement capabilities, avoiding bottlenecks that can limit AI application performance.

From desktop development to production scale

While much of the focus remains on data centres, Dell is also extending AI capabilities to developers through desktop systems.

The Dell Pro Max with GB10 and GB300 introduces high-performance computing at the workstation level, powered by the NVIDIA GB300 Grace Blackwell chip.

These systems allow developers to build and test AI models locally before deploying them into production environments.

This creates a more streamlined pipeline from development to deployment, reducing friction between teams.

However, the transition from experimentation to production remains a key challenge for many organisations.

Arthur Lewis, President of Infrastructure Solutions Group for Dell Technologies, highlights the importance of integration across the stack.

“We’re the first OEM to ship the GB300 Grace Blackwell Ultra Desktop Superchip with our Dell Pro Max desktop, bringing enterprise-grade AI computing directly to developers’ desks,” he says.

“For enterprises still stuck between pilot and production, the lesson is simple: integration matters, data readiness matters, deployment expertise matters.”

“A partner who delivers all three is the difference between AI as an experiment and AI as a business driver.”

Integrated systems for scalable deployment

To support end-to-end deployment, Dell is also offering Integrated Rack Scalable Systems, which bring together servers, networking and supporting infrastructure into a unified rack design.

These systems simplify deployment within data centres by providing pre-integrated configurations optimised for AI workloads.

This approach reduces the complexity of building AI infrastructure from multiple components, allowing operators to deploy capacity more quickly while maintaining consistency across sites.

The updates, announced at NVIDIA GTC 2026 in San Jose, reflect a continued focus on aligning compute, networking and physical infrastructure within the data centre.

By combining these elements into a single ecosystem, Dell and NVIDIA aim to support enterprises as they expand AI from initial development into full-scale production environments.