As artificial intelligence accelerates demand for computing power, the U.S. data center market is entering a period of rapid growth and heightened complexity. In a recent webinar, our panel of industry leaders examined the opportunities, constraints, and strategic inflection points defining this next phase, from energy availability and infrastructure strain to regulatory friction and capital risk.
The discussion didn’t end when the presentation did. Participants raised thoughtful questions that went to the heart of the industry’s most pressing issues: whether power shortages could slow development, how supply‑chain bottlenecks and policy choices are influencing timelines, and how operators can balance global workloads with local community concerns.
Below, we explore those questions and responses in greater depth, offering practical insight into how developers, operators, investors, and policymakers are thinking about risk, resilience, and growth in an AI‑driven data center economy — and what it means for the deals and developments taking shape today.
Q. What is your opinion about Elon Musk’s future view of data centers in space to solve for space and energy issues. If valid, how far in the future?
A. While it is difficult to put anything past Musk, we view data centers in space as a remote possibility.
Space-based data centers would face significant technical hurdles, including the significant costs of launching equipment into orbit, thermal management challenges in the vacuum of space, latency issues for ground-based users, and the difficulty of maintenance and repairs. Current satellite technology focuses on connectivity rather than compute. While solar energy is abundant in space (no weather interference), the infrastructure challenges make this a very long-term proposition at best.
If such technology were to become viable, it would likely fall beyond a 15-20 year horizon, which is the maximum length of current data center leases. Data from the Foley Data Center Report, as well as other publicly available information, indicates that advanced nuclear solutions like SMRs are likely 8-10 years from commercialization, and nuclear fusion is projected on a 2035-2040 timeline. Space-based data centers are likely to require technological leaps beyond those timelines.
Q. Where are the supply chain shortages? What needs policy support?
A. The scale of newer HPC and AI-hyperscale data centers is 10-50 times larger than facilities built 5-10 years ago, placing enormous stress on suppliers. Manufacturing capacity constraints, offshore manufacturing dependencies, and tariff/trade policy confusion compound the problem. Below are some of the key areas and components in which developers and operators report shortages and/or delays:
| Component | Challenge |
| Specialized Computing Hardware (GPUs, AI chips) | 68% of providers cite this as most difficult to source |
| Cooling Systems | 62% of providers report sourcing difficulties |
| Transformers | 44% cite sourcing challenges |
| Batteries and Rare Earth Materials | Often sourced from outside the U.S.; tariffs and trade issues create delays |
| Power Equipment (generators, switchgear) | Long lead times impacting construction schedules |
Useful Policy Support:
- Tariff relief for critical components
- Tax incentives for domestic manufacturing expansion
- Improved permitting processes for data center development, with 53% of developers citing permitting as the top obstacle in the Report
- Expansion of domestic manufacturing capacity, which more than six in 10 providers believe will improve equipment availability
Q. How much is the shortage of power over the next two years?
A. Quantifying the shortage of available power relative to demand is difficult and somewhat speculative. What most industry stakeholders agree on is that energy availability and redundancy is “far and away the number one obstacle” to successful data center development, both today and projected through 2030. Microsoft’s CEO recently framed the main barrier to data center development as one of power supply, rather than an excess of compute, citing more demand than the company’s data centers could currently handle. Power demand from data centers is projected to double by 2035 to almost 9% of all U.S. demand, which some describe as potentially the biggest spike since the advent of air conditioning.
There is also something of a power cliff anticipated by many in the industry. For example, one executive director at an international bank investing in data centers recently put it this way, “For the next year or two, it’s all growth — maybe even a hockey stick curve. Once power runs out in 2027 or 2028, that’s where we think deal flow will start to slow down.”
Q. Is there any industry concern that the “scaling” hypothesis behind generative AI may shift as models get smaller, advancements in AI move away from machine learning, or compute demand decreases? Or, do you see the future of tech (e.g., quantum) continuing to drive energy demand and the need for more data center expansion?
A. The consensus view favors continued demand, with comments like the one from Microsoft’s CEO mentioned in the question above providing anecdotal validation and support for the view. Quantum computing advancements are identified as a risk in the Report but ranked very low (6%) as a current obstacle, suggesting that quantum computing is not expected to fundamentally disrupt the data center demand model in the near term. While it is impossible to predict the future, and while some consolidation or “right-sizing” may occur in the industry, there has never been a time in human history where the usage of storage and/or compute has contracted in any meaningful way. To the contrary, demand tends to expand to meet supply.
Q. Since hyperscaler workloads and customers are global, how do operators ensure local data centers deliver tangible benefits to the host communities rather than just serving remote users? Is this a factor or constraint?
A. Recent community opposition to data center development has led to increased permitting issues and regulatory red tape. To that end, regulatory and permitting is cited in the Report as the phase where deals most commonly break down (48% of respondents). Public outcry has caused local governments to reject or delay data centers even in traditionally business-friendly areas, driven by concerns over water consumption, grid reliability, and general AI opposition.
The reality is somewhat different. “Big tech and data center companies are an easy target for politicians, utilities, and local groups to blame for rising energy prices, even when it is not necessarily true,” says Jeff Atkin. “Navigating this issue and correcting the narrative will therefore become increasingly important.” Many operators are actually the largest investors in local communities. These activities can produce a boost to local tax rolls and create construction jobs, for example. Most developers are also moving from evaporative-style cooling to closed-loop systems that use considerably less water, and some newer developments will actually put more water back into the local community. Utility upgrades, substations, road improvements, and other infrastructure investments can extend benefits to previously undeveloped areas.
Q. Are any of you concerned that some of these circular funding strategies pose a leveraged risk to the industry overall?
A. This is a heavily debated issue in the industry and in financial markets. Concerns about circular deals involving the sector’s biggest players have cause some to posit that there is ample default risk given the amount of structured finance now backing data center projects. There is also a growing number of “phantom data centers” now haunting utilities’ electrical grids.
Mitigating factors exist, however. The Wall Street Journal recently ran an article noting that companies making the lion’s share of data center investments are “mature and profitable companies” with “incredible balance sheets.” 63% of stakeholders in the Report anticipate a strategic correction but not necessarily a collapse, which suggests the industry expects some rationalization but not a dot-com-style bubble burst. Industry leaders like Mark Zuckerberg have acknowledged the risk of misspending “a couple of hundred billion dollars” but argue that “the risk is higher on the other side” (i.e., underinvesting in AI infrastructure is the greater risk to hyperscalers). While concern exists, the prevailing view is that solid company fundamentals and genuine demand for AI computing power provide meaningful protection against systemic leveraged risk.
