Data center REITs hit 96% occupancy as AI reshapes infrastructure costs

The surge in data center occupancy and rising rental costs creates a structural headwind for streaming companies integrating AI into their workflows. As infrastructure becomes a 'landlord's market,' streamers face higher op-ex for personalization, encoding, and metadata automation. This shift forces a recalculation of the ROI for AI-driven features, as foundational costs are no longer flat. The broader ecosystem is moving toward a growth-income hybrid model where power and physical space, rather than just connectivity, dictate competitive advantage. Watch for whether current 96%+ occupancy levels trigger 5-8% price corrections if interest rates remain elevated, potentially slowing the pace of new capacity delivery.

The data center supply crunch is being intensified by unprecedented power requirements for next-generation hardware. Per Gartner and S&P Global (June 2026), global data center electricity consumption is projected to grow 26% this year to 565 TWh, with AI-optimized servers accounting for 31% of that total. High-density chips like Nvidia’s Blackwell platform are pushing rack power demands significantly higher; for example, per Tom's Hardware (June 2026), the upcoming Vera Rubin GPU is expected to pull up to 2,300W per chip, nearly triple the requirements of previous generations. This trajectory is moving the industry focus from fiber redundancy toward 'behind-the-meter' power security and on-site generation to circumvent grid constraints.

In the streaming sector, these rising costs are driving a strategic pivot toward hybrid infrastructure. Per NewscastStudio and Equinix (June 2026), major broadcasters are increasingly maintaining high-quality video assets in private colocation environments while using the public cloud only for agile post-production and low-resolution proxies. This model aims to minimize data egress fees and avoid the escalating costs of full hyperscale reliance. Furthermore, per JLL (January 2026), 97% global occupancy has shifted negotiating leverage firmly to landlords, with construction timelines for new AI-scale campuses now stretching into multi-year territory due to shortages in labor and power substation components.