Equinix (EQIX) TD Cowen 10th Annual Communications Infrastructure Summit summary

Executive insights and strategic focus

• Utility and AI are top priorities, with a focus on securing long-term platform value through land and power banking, especially in response to utility dislocations in Western Europe over the past 3-4 years.

• The company has shifted to larger, multi-hundred megawatt campuses, such as the recent Atlanta acquisition, reflecting a step change in scale and a more active approach to xScale in the U.S.

• xScale developments provide both hyperscale market access and future runway for power, with the flexibility for the retail business to become a tenant as needed.

• Pre-leasing activity has increased, with enterprise customers now reserving capacity 6-9 months in advance, a shift seen in the last 2-3 quarters.

• Data center construction is trending toward larger, more flexible facilities designed to support higher densities and future technology shifts.

Market demand and customer trends

• Despite a challenging macro backdrop, record gross bookings were achieved, driven by increased certainty and a shift in market dynamics, with most new capacity for 2025-2026 already pre-leased.

• Hyperscale demand has created tailwinds, pushing enterprises toward colocation and driving larger deal sizes, with new workloads rather than migrations underpinning growth.

• Financial services, insurance, and life sciences are leading in AI adoption, while other enterprises focus on data readiness and governance for future AI use.

• The end of on-premise to colo/cloud migration is near, with new workload creation in colocation now the main growth driver.

• Sustainability and density requirements, especially for AI, are accelerating the shift away from sub-scale enterprise data centers.

Technology, infrastructure, and product evolution

• Rack densities are rising, with the average moving from 4-5 kW to 8-10 kW, and high-density deployments (15-30 kW) becoming more common, especially for AI workloads.

• Facility designs now emphasize flexibility, supporting both current and future high-density needs, with liquid cooling and modular power augmentation as key strategies.

• Interconnection remains a strategic differentiator, with new large deployments consuming interconnection at rates similar to historical norms on a per deployment basis.

• The company is modernizing existing sites to increase power capacity and extend asset longevity.

• AI training and inference workloads are present but not yet a massive demand driver; inference is expected to resemble current edge compute deployments rather than ultra-high-density GPU clusters.