As data center development accelerates across the United States, panelists at the Fort Worth Report “Informed Discussion: Data Centers in Fort Worth” converged on many ideas that cut through the noise. Featured panelists were Nikki Morris, TCU Ralph Lowe Energy Institute, Dr. Mohammad Islam, UT-Arlington, and myself. One of my major takeaways is that the information needed to bring transparency already exists. A challenge is organizing it into something communities and policymakers can actually use.
Some of my select highlights are offered and directions for the future follow. The video replay is at the end of the post.
From Repurposed Warehouses to Purpose-Built Campuses
Today’s data centers didn’t start out as data centers. From visits with two developers, originally with Digital Realty, I mentioned the industry’s roots in repurposing industrial-commercial real estate — defunct warehouses and old printing presses retrofitted to house early computing infrastructure. Today, a current wave of purpose-built, hyperscaler-financed campuses are reshaping local economies but other types of data centers exist in tandem.
The United States remains the epicenter of this build-out, capturing a disproportionate share of global data center development activity, nearly 80% according to a Cushman and Wakefield study. Hyperscalers—Amazon, Meta, Microsoft and Google are committing hundreds of billions of dollars annually to the build out, which are a primary engine behind the growth. Increasingly they are joined by a second category of builders, neo-clouds, but not exclusively so.
New Entrant Neo-Clouds
On the panel, I note a slightly more detailed case study centered on Iren, a firm that pivoted from Bitcoin mining to high-performance computing. Its project in Childress, Texas was cited as an example of the scale now typical in the sector — significant megawatt capacity paired with state-of-the-art GPU deployments. (Ironically, on 8/23, Nvidia said those specific systems would be 15% more expensive next year, owing to memory chip costs.) Iren is symbolic of this broader category gearing toward AI workloads and cloud computing. The model and its offshoots are much more complicated however. The trend is broadening as specialties emerge.
(Feature below details the evolution)
I noted that despite the competitive dynamic between traditional hyperscalers and newer neo-cloud entrants, both categories of company are, in practice, disclosing meaningful project information, if you’re willing to look. LinkedIn pages for firms like Applied Digital, Iren and Aligned Data Centers were pointed to as examples of where granular, real-time project detail is already surfacing publicly, in the absence of formal reporting requirements. Press releases across the industry have grown noticeably more detailed in recent years, a shift toward voluntary transparency predating current and evolving regulatory pushes. This is most apparent with higher quality developers that offer considerable disclosure.
Texas a Model
Texas is a case study in what attracts hyperscale investment — and why other regions, and even other countries, are studying it. These features listed explain some of the reasons why. Panelists pointed to the state’s competitive power market, its cost of power, and its availability of capital and manpower as the foundational advantages. One panelist, Morris, noted that countries in the Middle East are actively watching Texas’ approach as a potential model for their own infrastructure development. I have discussed this in other contexts, too.
I noted that Texas’ flexible-demand energy policy framework is a particular draw for those building projects. SB6 — legislation requiring large data centers to curtail power usage during grid emergencies and switch to backup generation — was described as a mechanism that lets the state’s competitive market accommodate new loads and support the grid. That is different than the capacity-market structure used in the Northeast, which is less flexible versus the Texas market.
Speed is a defining feature of the times. Speed-to-power has become a competitive necessity. While co-location of infrastructure assets often happens for reasons of speed, the goal for many hyperscalers is still to become a grid-connected asset over time.
A Transparency Checklist
I suggested the development of a best-practices checklist that would let policymakers and communities benchmark developers and projects. One suggested metric to use includes Power Usage Effectiveness (PUE), along with basic project fundamentals: IT capacity, megawatts consumed, project timelines, funding sources, and ownership structure.
The rationale wasn’t only about disclosure, but to begin to distinguish durable developers from speculative ones. Reliable developers plan on multi-decade time horizons, not just the next few years. The audit process suggested by Texas could help separate genuine long-term builders from opportunistic ones. I mention that one developer approaches these projects as if they intend to be part of their host community for a century. (The Compass Datacenter leader is cited in the feature above.)
The Texas solar development experience was raised as a cautionary tale: subsidies and incentive structures have previously encouraged speculative behavior and contributed to oversupply. I note that data center development reflected by the interconnection queue faces similar risks, ie., unintended consequences within the interconnection queue. Position in the queue may matter.
Leaving Communities Better
Running through the entire conversation was a community-benefit framing that panelists returned to repeatedly. I specifically mention Applied Digital’s investment in North Dakota and Microsoft’s in San Antonio as examples of data center development paired with local infrastructure and economic investment. Morris highlighted keeping it simple, leaving communities better off than before.
That standard is difficult to enforce through top-down regulation in a capitalist market where risk and infrastructure costs sit with private developers. My proposal was for better information architecture: nonpartisan, credible data that lets communities and policymakers make informed judgments.
Higher learning
There is a place for Texas’ research institutions. Rather than positioning any single regulator or company as the arbiter of best practices, I floated the idea of a university-anchored consortium — a nonpartisan clearinghouse for data center information that could aggregate the kind of granular disclosure already happening informally across LinkedIn pages and press releases into something structured and comparable.
Some of the transparency data center communities are asking for already exists in fragments. Additionally, the dynamics of urban areas versus rural are different as well. What works for Fort Worth may be different than Dallas, Childress, or Loudon County, Virginia, one of three ‘data center alley’ counties.
Fort Worth Report You Tube, August 20 at Texas Wesleyan University
FWR recaps day after panel
Addendum: Dallas is a top data center market, and the Texas footprint is growing, smarter.