For most of the past decade, clean energy investment was driven by policy. Tax credits, renewable mandates, and international climate commitments set the pace. That is still true. But something else is now driving capital into clean energy at a scale and speed that no policy framework anticipated: the power appetite of artificial intelligence.

The numbers are difficult to ignore. Global data center power demand is forecast to rise 27% in 2026 alone, reaching 132 gigawatts, up from 104 GW in 2025. Clean energy investment hit $2.2 trillion in 2026, nearly doubling fossil fuel spending. These two trends are not independent. They are the same story told from different ends of the same wire.

What AI Actually Demands From the Grid

The challenge is not simply volume. It is the nature of the demand.

AI data centers require 30 to over 100 kilowatts per optimized rack, compared to 5 to 15 kilowatts for traditional infrastructure. More importantly, AI inference workloads require power that is always on. A data center running large language models in real time cannot tolerate interruptions. It cannot draw down when the wind stops or the sun sets and wait for conditions to improve.

This constraint disqualifies intermittent renewables as a standalone solution. Solar and wind can contribute meaningfully to the overall energy mix, but they cannot provide the baseload, 24/7 power that AI infrastructure requires at the scale hyperscalers actually need. The result is a structural reconfiguration of how major technology companies think about energy, and where private capital is starting to flow in response.

The Nuclear Pivot

The clearest signal of this shift is the rapid repositioning of nuclear energy from an unloved legacy utility asset into a strategic infrastructure investment.

As of May 2026, 13 announced projects have committed over 9.8 GW of nuclear capacity to AI data center infrastructure. Every major US hyperscaler has signed at least one deal. Microsoft restarted Three Mile Island. Amazon committed $20 billion to convert the Susquehanna nuclear site, and X-Energy raised $1 billion in an IPO for small modular reactor development.

Small modular reactors are at the center of most of these arrangements. Nuclear is the technology showing the sharpest trajectory change in the clean energy investment landscape, with the center of gravity shifting away from Asia-Pacific toward Europe and the United States. Small modular reactors are emerging as a key catalyst, particularly as data center operators look for clean, dispatchable baseload power that solar and wind alone cannot provide.

For private market investors, the SMR story is particularly relevant. These projects are capital-intensive, long-duration, and not accessible through public equities in any meaningful way. They are being financed through structured private arrangements between hyperscalers and energy developers, with private credit and infrastructure funds playing an increasingly active role in the capital stack.

Where Private Capital Is Moving

The investment opportunity is not limited to power generation. The bottleneck, increasingly, is transmission.

Data centers face up to 14-year waits for grid connections in some markets. Without expanded grid capacity, cost declines in generation technology will not translate into reliable power reaching the system. Grid modernization is becoming one of the most capital-intensive infrastructure themes of the decade, and it is one where private investment is filling gaps that public utilities cannot address quickly enough.

Private markets are expected to play a growing role in funding grid upgrades and energy projects that are difficult to access through public equities. More than half of BlackRock’s surveyed institutional investors identified data center energy as an attractive investment theme, with 37% preferring energy infrastructure over exposure to large technology companies.

The capital is rotating downstream from AI software and chips toward the physical infrastructure that makes AI possible. Power purchase agreements, battery storage, grid interconnection assets, and long-duration energy storage are all attracting institutional attention that would have been unimaginable three years ago.

The Sustainability Dimension

There is a tension worth acknowledging. The same AI buildout that is accelerating clean energy investment is also driving electricity consumption to levels that strain existing grids and, in the near term, may increase reliance on fossil fuels to meet demand that renewables cannot yet fill.

Because hourly matching exposes the variability gaps in wind and solar, it is accelerating investment in technologies that can provide firm, low-carbon baseload power. Hyperscalers are now directing capital into advanced nuclear, including small modular reactors, and geothermal, precisely because those sources can deliver the consistent output that hourly matching requires.

The net effect on emissions will depend on how quickly the clean energy buildout can match the pace of demand growth. That race is the defining sustainability investment question of the next decade. And for the first time, it is being driven as much by private capital chasing infrastructure returns as by policy frameworks chasing climate targets.