Digital Realty, Dell, and DXC Partner to Accelerate Private AI Adoption: Turnkey AI Factory on PlatformDIGITAL®

The race to move artificial intelligence (AI) from proof of concept to real business impact advances on a new front. Digital Realty, Dell Technologies, and DXC Technology have announced a strategic collaboration to deploy and operate “end-to-end” private AI infrastructure on PlatformDIGITAL®, Digital Realty’s global data center, colocation, and interconnection platform. The agreement combines technology (the Dell AI Factory with PowerEdge servers, Dell AI Data Platform, and Dell Networking), location (high-density, carrier-neutral data centers distributed across more than 50 metro areas in 25+ countries and six continents), and operations (run and continuous transformation) managed by DXC. The goal is to reduce friction, accelerate time-to-value, and bring AI closer to client dataon their terms.

This isn’t just another piece of the puzzle; it’s an integrated package: validated use cases, reproducible architectures, accelerated deployment, and end-to-end management over infrastructure colocated near the data gravity centers where clients’ critical data resides and generated. Connected via ServiceFabric®, Digital Realty’s interconnection and orchestration solution, “Success in AI requires more than infrastructure: it demands the right people, processes, and technology,” emphasizes Holland Barry, Global Field CTO at DXC. “With Dell’s validated designs and Digital Realty’s infrastructure, DXC provides secure, end-to-end deployment and management that helps turn ambition into faster business results with greater impact.”

What each actor brings (and why together)

  • Digital Realty provides the physical and interconnection layer: PlatformDIGITAL®, a network of over 300 carrier-neutral facilities, with high density, power, and cooling tuned for GPU/AI, and ServiceFabric® to interconnect with clouds, carriers, SaaS, and data ecosystems where it matters (near-cloud, near-edge, near-AI).
  • Dell contributes the AI factory (Dell AI Factory): PowerEdge servers optimized for acceleration, Dell AI Data Platform for ingest, prep, governance, and lifecycle management of data/AI, and networks (high-performance switching, RoCE, leaf-spine) to channel training and inference pipelines.
  • DXC takes charge of planning, deployment, and operation: blueprints that are repeatable, deployment tools, observability, security, and transformation services to operationalize use cases and scale with business metrics.

AI is transforming how organizations operate and innovate, but realizing its potential requires infrastructure, strategy, and know-how that work seamlessly,”

explains Colin McLean, Chief Revenue Officer at Digital Realty. “With Dell and DXC, we help solve real problems, gain speed, and scale AI initiatives more securely, delivering actionable intelligence directly to the client’s data and operations.”

From PowerPoint to production: a library of cases and proven architecture

The initiative begins with a library of validated use cases —designed for private deployment with Dell AI Factory on PlatformDIGITAL®— covering common value patterns such as:

  • Enterprise copilots and augmented search (RAG) on internal data with governance and traceability;
  • Predictive analytics and operational optimization (maintenance, routing, inventory);
  • Security and compliance (classification, redaction, policy enforcement);
  • Multimodal processing (text, image, audio) with mixed pipelines of CPU/GPU for sensitive data that stays within the defined perimeter.

This library not only accelerates the phase 0 (valued in weeks), but also serves as a technical and operational template to scale from a pilot to SLA-governed environments. The high-density architecture available in PlatformDIGITAL® allows a transition from a few GPUs to entire pods, with ready power and cooling and low-latency connectivity to public clouds and corporate systems. “Ecosystem collaborations like this accelerate the adoption of hybrid, data-centric, AI-enabled infrastructure,”

summarizes Courtney Munroe, research vice president at IDC, highlighting that practical solutions, real use cases, and know-how shared together simplify operations and unlock the strategic value of data.

Why private AI (and why “with the data”)

Companies facing AI bottlenecks often recognize three friction points: dispersed data, latency, and compliance. Training, fine-tuning, and inference conducted far from data sources multiply timelines and costs; moving massive volumes between regions or clouds conflicts with Data Gravity, and regulatory requirements (privacy, residency, sectoral sovereignty) limit what can leave the perimeter. That’s why the core condition of this proposal is: bring AI to the data, not the data to AI.

With PlatformDIGITAL®, organizations deploy the AI factory locally, interconnected with their sources (ERP, CRM, data lakes, sensors, medical imagery) and with nearby clouds when managed services or elastic capacity are needed. ServiceFabric® orchestrates secure routes and connectivity policies so that training and inference flow through the correct lane (private, encrypted, QoS-enabled), and DXC adds the operational layer: monitoring, scaling, MLOps, FinOps, and SecOps tailored to each vertical.

From PoC to run: people, processes, and SRE for AI

The statement emphasizes that collaboration does not end at the rack. DXC provides enablement services for client teams (training, governance frameworks, MLOps lifecycle), deployment, and continuous operation. This run—often underestimated—includes:

  • Versioning of models and data, drift and bias monitoring, alignment with policies;
  • Tuning for cost and performance (when to use CPU, when GPU, queue sizing, batching);
  • End-to-end observability (latency, SLO/SLA, traces, audit);
  • Security (isolation, zero trust in north-south and east-west, tokenization, redaction);
  • FinOps applied to AI (more cost-effective results, reappropriation of consumption, overage from egress and traffic).

Reproducible architectures and dense blueprints: the “shortcut” to production

The consortium has prepared a deployment toolkit that standardizes planning, implementation, and operationalization of Dell AI Factory on PlatformDIGITAL®, with DXC as the integrator. These blueprints include everything from bill of materials (GPU/CPU, memory, networking, storage) and placement policies to CI/CD templates for data pipelines and models, security and governance rails (roles, RBAC, lineage), as well as acceptance testing and SRE for AI. The outcome is a streamlined path from concept to production, minimizing operational complexity and anchoring AI where the data and client’s hybrid systems are located.

Density, power, and networks: the “machine room” demanded by AI

The high-density colocation is at the heart of the proposal. The demands of training and inference—which consume kilowatts per rack—require stable power, liquid/air cooling, and a network fabric with low latency and high throughput. PlatformDIGITAL® has been densifying rooms for GPU-intensive loads for years, with interchange points (IX) and cloud nodes within the same facility or campus, reducing hops and jitter. Dell Networking completes the circle within leaf-spine, DC-to-DC, and edge-to-core domains; ServiceFabric® ensures logical proximity to clouds, carriers, and ecosystems as needed by the application.

What it means for regulated sectors (finance, healthcare, public sector)

In financial services, healthcare, or government, private AI isn’t just a preference: it’s a requirement to meet data residency, privacy, retention, and audits. The physical and logical proximity of PlatformDIGITAL®, combined with the control Dell AI Factory offers over lifecycle, data, and models, along with DXC’s management capabilities, provides a viable path to operate AI without exfiltrating sensitive information or restructuring the client’s core architecture.

Speed with control: the value of “validated”

One common barrier to AI adoption is repeating architectural errors: selecting inappropriate GPU for patterns, over-sizing networks, underestimating storage or egress, or neglecting governance from the start. The collaboration emphasizes validated designs: patterns that have been tested (and fine-tuned) in production environments, with documentation and realistic SLOs. This means the client teams inherit a stable, extensible starting point, rather than building “custom” solutions from scratch.

Market statements and perspective

This collaboration demonstrates how ecosystem partnerships accelerate the adoption of data-driven, AI-enabled hybrid infrastructure,” notes IDC via Courtney Munroe. “By combining global infrastructure with practical solutions, real use cases, and technical expertise, companies simplify operations, speed up deployments, and unlock the potential of their data as a strategic asset.” This aligns with Digital Realty’s vision of Data Gravity: as volumes grow, they tend to attract applications and services into their environment — and not the other way around.


Frequently Asked Questions

What exactly is Dell AI Factory deployed on PlatformDIGITAL®?
It is Dell’s AI factory PowerEdge servers, Dell AI Data Platform, and Dell Networkingintegrated within Digital Realty’s high-density data centers (PlatformDIGITAL®) and operated end-to-end by DXC. The client gains a secure, scalable environment for private AI “close” to their data, with validated use cases and tools to move into production.

How does ServiceFabric® reduce friction with legacy clouds and data?
ServiceFabric® is Digital Realty’s interconnection and orchestration layer. It enables secure, predictable connections between the AI factory and public clouds, carriers, neutral points, and enterprise systems. It applies security policies, QoS, and routing to ensure that training and inference traffic flow through private, low-latency, high-bandwidth circuits.

What’s the difference between a cloud PoC and this private AI proposal?
A cloud PoC often moves AI away from data (more egress, latency, lax compliance) and does not involve continuous operations. This proposal brings AI to the data (in high-density colocation), with validated designs, observability, security, and SRE to operate at SLA, with DXC managing deployment and run.

How does the validated use case library accelerate deployment?
It provides reproducible architectures, templates, and scaled learning (size, cost, latency, guardrails), avoiding common design mistakes and enabling the client to launch their private AI in weeks, with business metrics from day one.


Notes: Digital Realty operates over 300 data centers across more than 50 metros and 25+ countries. PlatformDIGITAL® and ServiceFabric® are trademarks of Digital Realty. All statements regarding plans and anticipated benefits are subject to risks and conditions typical of corporate communications.

via: digitalrealty

Scroll to Top