Applied Digital has signed a $7.5 billion [1] lease with a high-investment-grade hyperscaler to provide AI-related capacity.
The agreement underscores the intensifying race for physical infrastructure to support artificial intelligence. As large-scale providers seek more data center space, the deal positions Applied Digital as a critical link in the AI supply chain.
Wes Cummins, chairman and chief executive officer of Applied Digital, said the strategic move during an appearance on CNBC’s ‘Squawk on the Street’ in New York on Wednesday. The lease is set to span 15 years [2], providing the company with long-term revenue stability while meeting the needs of hyperscale clients.
Cummins said the deal reflects the broader trend of massive capital investment in AI buildouts. The demand for specialized capacity is rising as companies scale their computational capabilities to handle more complex AI models.
Industry projections suggest a significant acceleration in spending. AI spending is poised to reach $700 billion [3] in 2026, driving the need for the high-density power, and cooling solutions that Applied Digital provides.
The company is targeting high-investment-grade partners to mitigate risk while scaling its footprint. This approach allows Applied Digital to capitalize on the current infrastructure gap — a shortage of data centers capable of supporting the power requirements of modern AI chips.
Cummins said the company is focusing on the rapid deployment of capacity to keep pace with hyperscaler demand. The 15-year term of the current lease provides a predictable framework for the company to expand its operations and manage the high costs associated with data center construction.
“Applied Digital has signed a $7.5 billion lease with a high-investment-grade hyperscaler.”
The scale and duration of this lease indicate that hyperscalers are moving away from short-term flexibility in favor of long-term infrastructure security. By locking in capacity for 15 years, these providers are betting on the sustained growth of AI workloads, suggesting that the current 'AI buildout' is a foundational shift in global computing rather than a temporary trend.
