The Surging U.S. Demand for Compute and Power — and How Galaxy Plans to Deliver

The exponential growth of AI is driving unprecedented power demand, projected to double or triple data center consumption by 2028, severely straining the aging U.S. electrical grid and leading to multi-year interconnection delays. Despite efficiency gains, the Jevons paradox ensures that increased AI adoption outpaces these savings, compelling hyperscalers to commit over $360 billion in capex and pushing data center vacancy rates to a record low of 1.6%. This acute supply-demand imbalance underscores the strategic value of large-scale, power-ready infrastructure, exemplified by Galaxy's Helios campus in West Texas, which offers 800MW approved capacity and 2.7GW expansion potential, positioning it as a critical hub to meet the escalating needs of AI workloads amidst significant grid expansion initiatives in the ERCOT region.

Analysis

Exponential growth in AI is fueling an unprecedented power demand, which is severely straining the aging U.S. electrical grid and creating significant bottlenecks for data center development. Projections from the U.S. Department of Energy indicate data center power consumption could triple by 2028, while McKinsey forecasts U.S. capacity needs will surge from 25 GW in 2024 to over 80 GW by 2030. This has resulted in critical supply constraints, evidenced by a four-year average wait time to connect to the grid and an all-time low data center vacancy rate of 1.6% in North America. Despite recent AI model efficiency gains, the Jevons paradox is taking effect: lower compute costs are accelerating adoption and fueling more complex workloads, such as reasoning models, which ultimately drives net demand higher. This trend is validated by the sustained and increasing capex commitments from hyperscalers like Amazon, Microsoft, and Meta, which are collectively expected to exceed $360 billion in 2025. Consequently, markets with available power and supportive regulatory frameworks, such as the ERCOT region in Texas, are becoming key strategic locations. The recent approval of new high-voltage transmission lines in West Texas underscores the region's emerging role in accommodating this growth, making assets with pre-approved, large-scale power access, like the 800 MW Helios campus, exceptionally valuable.

Key Decisions for Investors

• Investors should consider long positions in companies exposed to electrical grid infrastructure, data center real estate, and power generation, particularly those operating within high-growth regions like Texas's ERCOT market where significant grid expansion is underway.
• The sustained high capex from hyperscalers (AMZN, MSFT, GOOGL, META) signals continued strength for their supply chains; therefore, opportunities may exist in component manufacturers, engineering firms, and data center operators poised to capture this spending.
• Be cautious of narratives suggesting AI efficiency gains will temper power demand; the evidence points to a Jevons paradox where efficiency drives greater adoption and overall consumption, reinforcing a long-term bullish thesis on compute and energy demand.
• Prioritize investments in companies or assets that possess rare, pre-approved access to large-scale power, as these entities hold significant strategic value and pricing power due to multi-year interconnection queues and acute supply shortages.