A lack of capital is not the most pressing bottleneck for AI progress—it’s the power needed to fuel it.
The energy needs of AI are outpacing the current grid's capabilities, creating a significant hurdle for progress. Unique combinations of public and private market solutions have emerged to meet this unprecedented demand for energy and capital.
Powering the AI era requires creative solutions across capital, energy and technology.
Unlocking AI's potential will require historic amounts of capital, digital infrastructure, and power.Training more advanced models requires larger, power-hungry data centers built for AI workloads—but data center vacancy rates sit at record lows. The industry is racing to catch up, with over 50Msqft in planned developments—double the volume of five years ago.*
Today, AI server racks consume 10x more power than Cloud equivalents—and the power grid is unprepared for this spike in demand.**Securing the power to operate these new data centers is even tougher, especially with AI data centers housing tens of thousands of energy-intensive GPUs in dense clusters. By 2030, data center power demand is expected to surge +160%*** relative to 2023 levels—and after a decade of flat demand growth, the current grid was not designed for this future. With new natural gas plants taking 5-7 years to get online and renewables only able to provide intermittent power—the end-result is a critical power bottleneck constraining AI progress.
This power bottleneck requires multi-layered solutions—and capital is a critical, necessary accelerant of progress.The next five years will require an "all-in" approach as hyperscalers and data center operators leverage a combination of power sources like natural gas and renewables—or go "behind the meter"—to meet short-term demand, while investing in long-term solutions like nuclear energy. AI is reshaping dealmaking, underwriting, and capital flows. Efficiently sourcing, deploying and recycling capital will determine how quickly these bottlenecks get resolved.
Sources:
*Newmark, 2025 U.S. Data Center Market Outlook
**https://www.goldmansachs.com/insights/articles/rising-power-density-disrupts-ai-infrastructure
“
When we think about the scale of what’s happening with AI, the cloud era offers one of the closest parallels. Cloud required an enormous buildout of computational and physical infrastructure. What did that deliver? Nearly a trillion dollars of new revenue and the emergence of new companies, platforms, and applications.”
Connect with Investment Banking