Data centres are the critical rate-limiting factor in hyperscaler growth and, in turn, power supply is the rate-limiting factor for those data centres.
So who is financing that power infrastructure? How are they doing it? And what happens if a market correction occurs, or if the flood of capital ultimately leads to overcapacity? Finally, what historical capital cycle offers the best lens through which to examine this unprecedented infrastructure buildout?
Our guest argues that the answer lies in the shale revolution. Joining our host Paul Chapman in our Houston office is Hill Vaden, Executive Director of the Energy Capital Insights Group at S&P Global.
S&P Global is also hosting the Financing US Power Conference in Houston from 28 to 30 September, covering many of the themes discussed in this episode.
Podcast Briefing: an Edited Q&A
The following Q&A has been adapted from the HC Commodities Podcast and edited for clarity and length.
AI's Real Bottleneck
Paul Chapman: Data centres are expanding rapidly to support AI, but power supply increasingly appears to be the real constraint. How do you view the relationship between AI demand and the energy sector?
Hill Vaden: Much of the public discussion around AI focuses on models, software and productivity gains. The physical reality is that AI ultimately depends on infrastructure. Data centres are becoming the critical bottleneck to hyperscaler growth, and power supply has become the critical bottleneck to data centre development.
What we are seeing is an unprecedented demand pull from hyperscalers. Their appetite for compute is creating an equally large appetite for power. That demand is now shaping investment decisions throughout the energy sector and creating opportunities for entirely new generation projects to be financed and built.
The Financial Engine Behind the AI Buildout
Paul Chapman: How does the financing process work when a hyperscaler decides it needs another major data centre?
Hill Vaden: It starts with a long-term contract. A hyperscaler agrees to purchase power over an extended period, giving developers confidence that there will be a reliable source of revenue behind a new project.
That contract then becomes the foundation upon which financing is raised. Lenders and investors are far more willing to support a project if the offtaker is a highly creditworthy company with a long-term need for electricity.
At the same time, enormous amounts of infrastructure capital have been raised globally. Investors are searching for long-duration assets that can generate predictable returns, and power infrastructure increasingly fits that requirement. The intersection between AI growth and energy demand has therefore become one of the most attractive themes in infrastructure investing today.
The Race for Speed to Power
Paul Chapman: We keep hearing the term "speed to power". Why has it become so important?
Hill Vaden: Data centre developers want capacity online quickly, but grid connections often take years. That gap has encouraged increasingly creative solutions, including behind-the-meter generation using gas engines and fuel cells. In many cases, the ability to secure power quickly has become one of the most important competitive advantages in the entire AI infrastructure buildout.
Why Natural Gas Sits at the Centre of the Story
Paul Chapman: What role does natural gas ultimately play in enabling AI-driven growth?
Hill Vaden: The issue is less about AI itself and more about reliability. Data centres require power around the clock, often with extremely demanding uptime requirements. While renewable generation continues to grow, hyperscalers still need access to dispatchable power that is available whenever it is required.
That reality has placed natural gas in a particularly strong position. It remains one of the most practical ways to provide large amounts of reliable power at scale, which is one reason why regions such as Texas continue to attract investment.
Interestingly, the hyperscalers do not generally want to become energy companies themselves. Their preference is to secure long-term power through contracts while leaving generation and operational risk to utilities and developers.
Why This Looks More Like Shale Than Dot-Com
Paul Chapman: Many people compare AI to the dot-com boom. Why do you think shale is the more useful analogy?
Hill Vaden: Shale was the last energy investment boom that combined technological innovation, abundant capital and intense competition. It was location-specific, infrastructure-intensive and characterised by a race to secure strategic positions before the opportunity became fully understood.
You can see similar dynamics emerging today. Developers are competing for land, power access and customers in much the same way that shale operators once competed for acreage. There is a clear winner-takes-most mentality in parts of the market.
The other similarity is the importance of demand pull. Hill points to George Mitchell's early pipeline commitments as an example of how guaranteed demand helped unlock investment and innovation in shale. In today's market, hyperscaler demand is playing a similar role by providing the commercial rationale for new power projects.
The details may be different, but the combination of capital, infrastructure development, innovation and competitive urgency gives this cycle a distinctly shale-like feel.
One of the major constraints to the AI ambitions being realised is the access to power. And specifically dispatchable generation, dispatchable power.
The Risk of Overbuilding
Paul Chapman: What happens if AI demand fails to meet the market's most optimistic expectations?
Hill Vaden: There is certainly a possibility that parts of the market become overbuilt. History suggests that major infrastructure cycles rarely stop at exactly the right point. Capital enters quickly, supply expands aggressively and capacity eventually exceeds immediate demand.
The more important question is whether that infrastructure remains useful over the long term. Railroads, highways and power systems have all experienced periods of excess capacity. In many cases, demand eventually grew into the infrastructure that had been built ahead of time.
Permitting, Politics and Public Resistance
Paul Chapman: What are the biggest obstacles slowing development today?
Hill Vaden: Financing may dominate headlines, but project execution remains a major challenge. Power projects are encountering permitting delays, infrastructure constraints and a growing level of public scrutiny.
Data centres are increasingly facing the same political and community challenges that have long affected pipelines, transmission projects and industrial facilities. Concerns around water use, electricity prices and local impacts are becoming more visible as the sector expands.
AI, Productivity and the Future of Work
Paul Chapman: If AI delivers the productivity gains many people expect, what does that mean for jobs?
Hill Vaden: Technology tends to change the nature of work rather than eliminate it entirely. Some roles may become less important, but new specialisms will emerge around the development, management and use of AI systems. The jobs may evolve, but the demand for skilled people is unlikely to disappear.
What Happens If the Bubble Bursts?
Paul Chapman: If there is a correction, what survives?
Hill Vaden: History suggests that capital cycles can create both extraordinary value and significant losses at the same time. The dot-com era produced many failures, but it also produced some of the world's most valuable companies. Shale followed a similar path, generating substantial capital destruction before ultimately transforming global energy markets.
If AI delivers even a portion of the productivity gains many expect, the demand for power, data centres and digital infrastructure is likely to remain significant. The real challenge for investors is identifying which companies and assets emerge as the long-term winners once the cycle matures.
The technology may survive even if parts of the capital structure do not. That is one of the most important lessons from previous investment booms, and one reason why the shale analogy remains so compelling.
HC Group is a global search firm dedicated to the energy and commodities markets.
Learn more about our Power & Environmental Products Talent Practice
Explore the full HC Commodities Podcast archive