Digital Realty & Vultr launch global GPU AI infrastructure
Digital Realty and Vultr have entered a strategic partnership to provide GPU-accelerated AI infrastructure for enterprises across major global data hubs.
The collaboration leverages Vultr's cloud-based GPU clusters, powered by NVIDIA HGX B200 and AMD Instinct MI325X chips, and deploys them on Digital Realty's PlatformDIGITAL data centre platform. This integration aims to position high-performance AI computing resources near where enterprise data is generated and stored, supporting a range of business needs from compliance to scalability.
Infrastructure combined
With this partnership, organisations will be able to access GPU resources and AI workloads through Vultr's infrastructure, which is now directly integrated with Digital Realty's ServiceFabric, a fabric interconnection service. This permits direct access to the AI Private Exchange (AIPx), presented as a secure, low-latency channel for hybrid and multi-party AI operations. According to both companies, access is already live in Atlanta, Dallas, London, and Singapore, with future connectivity planned for San Francisco, Frankfurt, Mumbai, Sydney, and Tokyo.
The solution is intended to serve enterprises requiring tailored configurations for compliance and sovereignty, as well as those seeking to scale AI deployments with flexible, usage-based consumption models. The architecture, both firms stated, has been optimised for environments demanding high-density and GPU-intensive computing.
Integration focus
The announcement highlights several intended benefits for global businesses. Among them, enterprises gain expedited access to both GPU clusters and the AI Private Exchange, as well as the ability to deploy in pre-validated environments supporting AI training, inference, and other workloads. The design also aligns local infrastructure to assist with regulatory and data sovereignty obligations, offering a consumption model that purportedly does not require long-term commitments.
"Our partnership with Vultr is about putting AI where enterprise data lives," said Chris Sharp, Chief Technology Officer of Digital Realty. "By combining Vultr's disruptive price-to-performance model with our global PlatformDIGITAL footprint, we're enabling instant AI infrastructure that empowers enterprises to activate Private AI - close to their data, within compliance boundaries, and without the complexity of standing up bespoke environments. This trusted foundation allows customers to focus on achieving AI outcomes at scale. ServiceFabric and AIPx extend this value, offering secure orchestration pathways to distributed compute and multi-party AI collaboration."
Vultr's GPU clusters are aimed at organisations looking to support both immediate and scalable AI workloads globally. The companies underscored the ability for enterprises to align infrastructure locally, meeting a variety of compliance criteria depending on jurisdiction, while still benefitting from low-latency performance for AI programmes such as agentic AI, model training, and inference operations.
Market objectives
Kevin Cochrane, Chief Marketing Officer of Vultr, commented on the focus of the joint venture for enterprises seeking both performance and regulatory assurance.
"Enterprises need AI infrastructure that's not only powerful, but production-ready responsive to compliance considerations," said Kevin Cochrane, Chief Marketing Officer of Vultr. "Our collaboration with Digital Realty delivers exactly that: immediate access to high-performance GPU infrastructure in the world's most strategic digital hubs. Together, we're trusted, long-term partners helping organizations scale AI with confidence."
Both companies indicated plans to further expand deployment and connectivity within key global digital markets to meet the growing demand for AI infrastructure. This includes ongoing extension to strategic cities and continued development of low-latency, secure interconnection solutions.
The partnership comes as enterprises increasingly seek dedicated, production-grade infrastructure capable of supporting the computational demands of modern AI workloads, alongside requirements for data compliance, security, and scalability.
Through this alliance, clients are anticipated to have access to a global footprint of data centre resources equipped to run high-performance AI jobs, with orchestration options that allow for integrated, compliant, and responsive deployment across borders.
