The Data Center Capacity Crunch – How Bad Is It and How Did We Get Here?

The OCOLO team spends our days connecting all parties in the digital ecosystem, working to enable more frictionless connections, relationships and commerce. We listen to the biggest challenges that IT enterprises and data center service providers face as demand for digital infrastructure skyrockets in this new era of AI and automation. And we do what we can, with the resources we have – deep industry knowledge, a comprehensive database of market intel and colocation inventory and strong relationships that span dozens of countries – to help them navigate them. Right now, the biggest challenge everyone faces is unprecedented capacity constraints.

Today, we’ll look at which markets are most constrained, what’s causing the crunch and how providers are coping. And we leveraged AI to help us take a deeper and more efficient dive into our analysis – with very impressive results!

Here’s what we did: we asked ChatGPT to help us with a blog analyzing which US Tier 1 data center markets are most capacity constrained, what major factors have driven the constraints over the past five years, and how providers are coping, using public data.

Here’s what ChatGPT did: All of that, succinctly and quickly, plus it created a “Severity Ranking” and its progression from 2019-24 for the top five most constrained Tier 1 markets. While we didn’t get all the underlying data and actual calculation used, AI did provide the rationale for the Severity Ranking, list of source materials and data used for charts and tables. Did we mention it was quick?

So, let’s jump in with the primary drivers of the capacity crunch, which of course start with everyone’s favorite blessing and curse: AI. In fact, the rest of the drivers on the list below strike us as simply follow-on effects of accelerating AI adoption and implementation:

Drivers of the Capacity Crunch

1. Artificial Intelligence Workloads: The explosion in AI demand, particularly for training large language models, has driven massive compute cluster requirements. GPU clusters draw 10x the power of traditional workloads, rapidly outstripping the capacity of existing infrastructure.
2. Power & Utility Delays: Grid interconnects and substation upgrades take years to plan and permit. Utility coordination has not kept pace with the speed of demand.
3. Land Use and Zoning Barriers: Community backlash (NIMBY), environmental review processes and zoning delays have made new approvals harder to secure.
4. Supply Chain & Construction Challenges: Long lead times for transformers, switchgear and specialized construction labor continue to slow deployment timelines.

We can’t count the number of conversations our OCOLO team has had with current and prospective clients and partners on all these topics, and they are covered exhaustively on the industry and investor conference circuit as well.

Of course, the bigger the data center market is, the harder it’s going to be hit by all these challenges, which brings us to the Top 5 Tier 1 Markets by Severity of Capacity Constraints, below:

Tier 1 Markets Ranked by Severity of Capacity Constraints (Worst to Least)

1. Northern Virginia (NoVa) – Severity Score: 10.0
   The world’s largest data center market has now become a victim of its own success and massive bottleneck for hyperscalers and colocation service providers alike. Power availability in Loudoun County is severely constrained, with multi-year timelines for new substation buildouts. Dominion Energy has imposed restrictions on new hook-ups, and local opposition to data center sprawl is intensifying.
2. Silicon Valley – Severity Score: 8.5
   Land is scarce, power is expensive and utility timelines are long. AI workloads have drastically intensified the strain, and local municipalities are increasingly resistant to new large-scale developments.
3. Dallas-Fort Worth (DFW) – Severity Score: 7.0
   DFW has gone from boasting abundant power and land to significant power allocation delays and grid pressure due to overlapping growth in residential and industrial demand.
4. Phoenix – Severity Score: 6.5
   Heat and water issues, along with growing sustainability scrutiny, are slowing the pace of new builds. The local grid is also under strain, especially in key industrial corridors.
5. Chicago – Severity Score: 5.0
   Still better off versus other Tier 1 markets, but large-scale AI deployments are driving up power pricing and interconnection delays. Fiber congestion is also becoming a concern.

Here’s how they all look together on a chart:

It’s even more shocking to see how quickly some of these markets went from stable to Defcon 5. The table below compares the Severity Scores of the Top 5 in 2019 vs. 2024:

It’s easy to see why the top two markets – NoVa and Silicon Valley – are scrambling to find capacity in alternative markets like Northern Florida and Charlotte, NC (NoVa), as well as Salt Lake City, UT and Reno, NV (Silicon Valley). Data center developers are now looking to Omaha, NE as an alternative to DFW, and to San Antonio, TX instead of Phoenix for land and power as well.

OCOLO’s partners on both sides of colocation transactions would agree with the markets on this list, and the order and rapid escalation of severity. At this point, single-location multi-megawatt deals are the most challenging assignments, and stringing together multi-location deals requires ingenuity, interconnections and patience on the part of everyone involved. And that’s before even considering specialized connectivity or cooling needs some AI installations require.

Conclusion

AI is not just a workload trend—it’s a full-scale infrastructure challenge. The current capacity constraints facing U.S. Tier 1 data center markets are forcing a reckoning. To stay competitive, operators must evolve site selection, permitting and energy strategy approaches. The winners in this next phase will be those who can align long-term vision with tactical agility, whether in core markets or the emerging edge.

We see OCOLO among these winners because we’re able to take on the work of finding the capacity – particularly those highly desirable and elusive megawatts – wherever they are and connecting our clients wherever they need to be.

---