AI’s Energy Question Deepens as Anthropic Considers Large-Scale Power Deals

The rapid expansion of artificial intelligence is beginning to reshape not only the technology sector, but the energy landscape that supports it. Anthropic, one of the leading developers of advanced AI systems, is exploring options to secure up to 10 gigawatts of power capacity, according to a report by The Information—an indication of the scale at which the industry now operates.

The discussions reflect a growing reality for AI companies: computing power is no longer just a question of chips and data centers, but of electricity. Training and operating large-scale AI models requires vast amounts of energy, and as demand for AI services accelerates, companies are increasingly focused on long-term access to reliable power.

While the reported figure represents potential capacity under consideration rather than a confirmed commitment, it highlights how leading AI firms are thinking in infrastructure terms traditionally associated with heavy industry. Securing power at this scale would likely involve a mix of long-term energy agreements, new data center developments, and partnerships with utilities or energy providers.

Across the technology sector, the race to build AI capability has triggered a parallel competition for physical resources. Cloud providers and AI developers have been expanding data center footprints at a rapid pace, often prioritizing locations with access to stable grids, favorable regulation, and abundant energy supplies. In some regions, utilities and policymakers have begun reassessing grid planning as technology demand rises faster than expected.

The issue also intersects with environmental and policy considerations. Many technology companies, including AI developers, have made public commitments to reduce carbon emissions or rely on low-carbon energy sources. Meeting those goals while supporting rapidly growing computing needs has added urgency to discussions around renewable power, nuclear energy, and grid modernization.

For Anthropic, which has attracted significant investment from major technology partners and cloud providers, securing long-term energy capacity would support continued scaling of its AI models and services. The company operates in an increasingly competitive field, where performance, reliability, and availability are closely tied to the underlying infrastructure.

Industry analysts note that energy planning is becoming a strategic differentiator in the AI race. Companies that can lock in power, land, and cooling capacity may gain an advantage as constraints emerge in certain markets.

The reported exploration does not guarantee a final agreement, and any large-scale power arrangements would likely unfold over multiple years. But the direction is clear: as artificial intelligence moves from research frontier to industrial platform, the question of how to power it is becoming as central as the technology itself.