Last month, I connected with other infrastructure leaders at the Aiken/Augusta Energy Summit hosted by E4 Carolinas. Below, I share a summary of my remarks on the past decade of growth in digital infrastructure, as well as takeaways about what’s next from a panel of peers that I had the pleasure of moderating.
The last decade of data center evolution has seen three distinct eras that have led up to our current AI boom. I’ve seen plenty of change across my 30 years in the technology industry, but we’re moving at warp speed in the data center world. We’ve grown from racks and rows to data halls and now to entire hyperscale campuses. It’s helpful to look back at where we have been in order to gain perspective on where we are now and to anticipate where we’re heading as an industry.
The Cloud Era
In 2016, everyone was talking about “digital transformation.” This marked the beginning of what I call the cloud era; companies wanted to get data centers out of their own buildings, but they needed a promise of resiliency. Data center buildings ranged from 50,000–100,000 square feet and 5–10 MW of power.
The Social and Video Era
Social media has been around for a while, but the COVID-19 pandemic kicked off an era of intensive usage as folks looked to connect with others while staying safe at home. Video via streaming apps became and remains many people’s preferred way to learn and be entertained. In this era, data centers expanded to 100,000–150,000 square feet and demanded 50 MW to more than 100 MW of power.
The Generative AI Era
That brings us to today, with generative AI largely driving the demand for new data center capacity. Buildings are reaching 1,000,000 square feet or more on sprawling campuses demanding 1 GW+ of power.
Each successive generation of data centers has incorporated the requirements of the preceding era while expanding to meet new demands.
Data Centers Today
There have been huge advancements in data center technology over the past 10 years. Think about the differences between a car built in 2016 and one built in 2026. You would notice improvements across almost every aspect of it, and the same is true for data centers.
Consumers tend to lump all data centers together into one category: AI factory. But data centers have different applications, just like different cars suit different drivers and purposes. Data centers today range in size and function, and each type requires different engineering. The DC BLOX portfolio contains examples of several types.
Comparison of Data Center Types
Data table version of type comparison
| Data Center Type | Power Capacity | Building Size (Sq. Ft.) | Typical Customers |
|---|---|---|---|
| Small Edge | 5–10 MW | 50,000–75,000 | Network carriers, wireless carriers, cloud companies, and social media |
| Medium Enterprise | 5–30 MW | 75,000–150,000 | Healthcare, manufacturing, banking, research, government, universities, and social media/web |
| Cable Landing Station | 10–25 MW | 100,000–200,000 | Global networks, internet, and governments |
| Large Hyperscale Cloud | 40–200 MW | 250,000–2,000,000 | Artificial intelligence, cloud companies, and social media |
| Extra Large AI | 400 MW–2 GW | 2,000,000–6,000,000 | Artificial intelligence, government, and cloud companies |
What’s Next for Data Center Growth
It’s no secret that data center demand continues to grow. The main drivers of demand today are the artificial intelligence boom, pervasive digital reliance, and a new generation of users of digital services. Platforms have high-density workloads requiring vast compute power, and our entire society operates with systemic dependence on cloud ecosystems. And people are relying on digital services earlier and earlier. Elementary schoolers represent a significant and growing demographic.
Here are some things I’m seeing as I look at the digital infrastructure landscape.
The Southeastern United States, where we started and continue to grow DC BLOX, represents 40% of the entire U.S. data center market thanks to a history of available power, plentiful land, and policies conducive to business. But opposition by governments and pushback from communities now rival power availability as the primary bottleneck for market entry. Public perception has shifted against our industry, despite widespread consumer and business adoption of AI. Data centers have become a ballot box issue.
Yet data centers aren’t going away, and they aren’t new, either; we’ve been living with them for more than 25 years. Across the panel of experts who spoke at the Aiken/Augusta Energy Summit, there was agreement that we want to be good stewards and partners, co-investing in the communities we join to overcome the obstacles of space, power, security, and availability of skilled labor. Tricia Pridemore, EVP of Energy, Regulatory and Policy at DC BLOX, calls these not challenges but opportunities.
We at DC BLOX believe in the importance of digital infrastructure as an engine of economic growth. We want to earn communities’ trust by taking the time to listen to their concerns and address them transparently, with factual information. We have pledged to uphold 10 commitments in every community we serve and hope to move past adversarial zoning fights to pave mutually beneficial paths forward. Further, we signed the White House Ratepayer Protection Pledge, committing to pay the full cost of the power and grid infrastructure our facilities require. We and our industry peers need to work with municipal, state, and federal officials to co-create policies that provide a structure for future growth. And with energy costs rising across the board, we need to encourage our power providers to develop rate structures and clean energy models to protect local ratepayers.
By approaching growth thoughtfully, acting with our core values of trust, tenacity, and teamwork, we hope to earn the social license to continue serving locally and connecting globally.
Learn more about the Whitehouse Ratepayer Protection Pledge and how it builds on the DC BLOX Community Commitment.
