Tech Giants' $400B AI Power Push Faces Energy Bottleneck, Space Solar Solutions Explored
November 10, 2025
The push by Dominion Energy and other utilities to expand data-center capacity in Virginia and beyond could lift U.S. electricity use for data centers to roughly 7-12% by 2030, though some forecasts warn this may be overstated and that many planned centers might never be built.
Some firms are exploring unconventional power options, including repurposed turbines and even space-based solar concepts, with collaborations like Google and Starlink testing in development.
There is growing tension between climate commitments and rapid AI-scale ambitions, signaling a shift in energy strategy among hyperscalers and tech firms.
Beyond securing chips, the next bottleneck is powering and provisioning facilities near power sources to avoid inventory and build delays.
Tech giants such as Google, Microsoft, AWS, and Meta are collectively investing hundreds of billions—roughly $400 billion in 2025 with more in 2026—to build out AI infrastructure and data-center capacity.
Longer-term energy strategies include pursuing nuclear options like Small Modular Reactors, expanding solar and battery storage, and exploring large-scale projects in Texas and California.
The AI race hinges not just on chips but on reliable, affordable electricity to power massive data centers, a bottleneck emphasized by leaders at Microsoft and OpenAI.
In the near term, utilities are leaning on quick power solutions such as natural gas and retained coal plants, with some data centers even purchasing used turbines to accelerate supply.
Projected supply strains could create a 45-gigawatt gap by 2028, risking AI rollout if new power sources don’t come online.
Space-based computing concepts, powered by solar energy, are being tested or proposed by Starlink and Google, signaling a futuristic avenue for data-center power.
Efforts include grid upgrades and ambitious visions for space-based computing, illustrating a broad, diversified approach to meeting growing energy needs.
Some hyperscalers have softened climate pledges, shifting focus toward practical power projects and diversification of energy sources, with reports of changes to net-zero commitments.
Summary based on 3 sources
Get a daily email with more Tech stories
Sources

Digital Journal • Nov 10, 2025
The AI revolution has a power problem
Economic Times • Nov 10, 2025
The AI revolution has a power problem
AFP • Nov 10, 2025
The AI revolution has a power problem