Equinix Sees AI Infrastructure Demand Surge as Enterprises Modernize Legacy Systems

Enterprise AI infrastructure demand is accelerating as companies modernize legacy on-premises systems and deploy new AI-native workloads. Equinix cited record backlog, interconnections and bookings, supporting its recently raised financial outlook. Customers are scaling four deployment patterns simultaneously: private AI stacks, sovereign workloads constrained by data residency rules, batch training and inference, and latency-sensitive applications located closer to users and data. Equinix is focusing on network orchestration and compliance through its Fabric Intelligence platform, while also addressing data center concerns around energy, water, land use and community economic benefits.

Enterprise AI infrastructure demand is being driven by the gap between companies’ future ambitions and the legacy infrastructure they continue to operate. The data center operator’s role has evolved through several major technology transitions, from internet scaling to multicloud connectivity. The current AI cycle is a shift that requires infrastructure to determine where data is processed, reasoned over and acted upon in real time, including at the network edge and within jurisdictional requirements.

Equinix serves 60% of the Fortune 500 and sees customers navigating both the challenges and opportunities associated with distributed, AI-driven workloads.

The company recently raised its financial outlook. The company’s confidence reflects demand across geographies, industries and customer types rather than a short-term AI investment cycle. Enterprises are modernizing legacy on-premises infrastructure that was not designed for distributed AI workloads, while new AI-native workloads are also creating demand. In many cases, the organizations pursuing those deployments are already Equinix customers.

Record backlog, interconnections and bookings over the preceding three or four quarters are factors supporting the company’s outlook. Equinix continues to evaluate infrastructure investments based on the same yield expectations investors have received historically.

Four AI use cases that Equinix is seeing scale at the same time among enterprise customers:

Stack: Deploying a company’s technology stack in an Equinix facility to run open models on private AI infrastructure, reduce token costs and connect to cloud and OEM partners.

Sovereign: Supporting data residency and compliance requirements. Equinix operates in 36 countries and has added software capabilities for geo-fencing workloads within a jurisdiction.

Batch: Deploying capacity for training and batch inferencing, supported in part by technologies such as liquid cooling.

Latency-sensitive workloads: Locating agentic workload nodes within a metro area to support low latency and, for some customers, reduce costs associated w...