Bridging Speed-to-Power and Sustainability in the US Data Center Market

Over the past two days I had the opportunity to speak at two very different, but deeply connected, events focused on the future of power infrastructure for AI and data center growth in the United States.

The first was DICE near Washington DC, where I was fortunate to join speakers from companies including AWS, LinkedIn and other industry peers discussing the rapidly evolving realities of AI-driven infrastructure deployment. What stood out clearly throughout the discussions was that speed-to-power has now become the dominant constraint shaping decision making across the sector.

The challenge, however, is that different parts of the infrastructure ecosystem are often moving at different speeds and with different priorities. Operators are under pressure to deploy capacity rapidly. Cooling specialists are understandably focused on reducing water consumption and moving toward dry cooling solutions. Sustainability teams are targeting lower carbon outcomes. Yet in many discussions there still appears to be limited integration between these objectives at a systems level.

One observation I raised during the session was that combined cooling and power technologies are still not being sufficiently considered within many of these deployment pathways. In practical terms, there are opportunities to reduce electrical chiller demand, improve overall site efficiency and support lower carbon trajectories through integrated energy system design. Importantly, this does not need to conflict with the urgency around deployment timelines.

This is where structured transition thinking becomes increasingly important. Infrastructure deployed initially to solve immediate power availability challenges does not need to become stranded infrastructure. Systems can be designed with future integration in mind, allowing combined cooling and power, thermal recovery, lower carbon fuels and wider hybridization strategies to be incorporated progressively over time as operational requirements, technologies and grid conditions evolve.

The following day I attended the US Combined Heat and Power Alliance North East Chapter conference in Philadelphia, where discussions focused on how the CHP sector can support accelerating data center deployment across North America.

The atmosphere there reflected an industry that understands both the scale of the opportunity and the need to adapt to evolving customer expectations. One particularly important development discussed was the ongoing Federal evaluation of CHP eligibility under the 45Y tax credit framework. If implemented favorably, this could significantly improve the economic case for high-efficiency CHP deployment within data center infrastructure.

At the same time, several practical deployment hurdles remain. Space constraints within modern campuses continue to influence technology selection. Water consumption perceptions remain highly sensitive. And above all, developers continue to prioritize deployment certainty and speed.

For the CHP industry, this creates an important challenge — and opportunity. Success will depend not simply on promoting traditional efficiency arguments, but on delivering systems that align with how the market is actually evolving: rapidly deployable, operationally resilient, capable of supporting future decarbonization pathways and designed around the realities of modern data center development.

What became increasingly clear across both events is that the industry is moving toward a much more integrated conversation around power, cooling, resilience, efficiency and carbon trajectory. The technologies themselves already exist. The challenge now is coordinating them into infrastructure strategies that reflect both immediate deployment realities and long-term transition goals.

Frequently asked questions

What is DICE and what does it cover?

DICE, the Data Center Investment Conference and Expo, is a major US industry event focused on the investment, development, and operational dimensions of data center infrastructure. It brings together developers, operators, investors, and technology providers to discuss the evolving challenges of AI-driven infrastructure deployment including power availability, cooling strategy, sustainability, and capital deployment. The event reflects the broader shift in the data center sector from a primarily technology-focused conversation to one where energy infrastructure and investment strategy are increasingly central.

What is the 45Y tax credit and why does CHP eligibility matter for data centers?

The 45Y tax credit is a US federal clean electricity production tax credit introduced under the Inflation Reduction Act. It is designed to support investment in lower-carbon electricity generation technologies. Whether combined heat and power systems qualify under the 45Y framework has significant implications for the economics of CHP deployment in data center infrastructure. Favorable eligibility would materially improve the financial case for high-efficiency CHP systems at a time when the data center sector is evaluating distributed generation as a primary infrastructure option rather than emergency backup alone.

What is structured transition thinking in data center energy infrastructure?

Structured transition thinking is an approach to infrastructure design that recognizes systems deployed today to solve immediate power availability challenges do not need to become stranded assets as conditions evolve. Rather than optimizing for a single operating condition, structured transition thinking designs infrastructure with future integration in mind, allowing combined cooling and power, thermal recovery, lower carbon fuels, and wider hybridization strategies to be incorporated progressively over time as operational requirements, technologies, and grid conditions change. The objective is infrastructure that meets immediate deployment needs while preserving the ability to improve performance across its operational life.

Why are combined cooling and power technologies underutilized in current data center deployment strategies?

Despite practical opportunities to reduce electrical chiller demand, improve overall site efficiency, and support lower carbon trajectories through integrated energy system design, combined cooling and power technologies are not being sufficiently considered within many current deployment pathways. A primary reason is that different parts of the infrastructure ecosystem are moving at different speeds with different priorities. Operators are focused on deployment speed. Cooling specialists are prioritizing water reduction and dry cooling. Sustainability teams are targeting carbon outcomes. Limited integration between these objectives at a systems level means that technologies capable of addressing multiple challenges simultaneously are being overlooked.

What practical barriers does the CHP industry face in the data center market?

Several deployment hurdles affect CHP adoption in modern data center environments. Space constraints within campus developments influence technology selection and limit the footprint available for onsite generation and heat recovery equipment. Water consumption perceptions remain sensitive, particularly in regions where data center water use is under scrutiny. Above all, developers are prioritizing deployment certainty and speed, which creates pressure on any technology that requires additional design complexity or longer procurement timelines. Success for the CHP industry in this market will depend on delivering systems that align with how the market is actually moving rather than promoting traditional efficiency arguments that do not address immediate deployment realities.

Why is the data center energy conversation shifting from individual technologies to integrated systems?

The challenges now facing AI infrastructure development cannot be solved by optimizing any single variable in isolation. Speed-to-power, operational resilience, cooling efficiency, carbon trajectory, and long-term lifecycle performance are increasingly interconnected. Addressing one without considering the others creates infrastructure that performs well on a single metric but struggles under real-world conditions across a twenty-year asset life. The technologies needed to address these challenges simultaneously already exist. The barrier is coordinating them into infrastructure strategies that reflect both immediate deployment realities and long-term transition goals, which requires systems-level thinking rather than technology-by-technology evaluation.