Tony Qorri is vice president of construction at data center colocation provider Dallas DataBank. The opinions are the author’s own.
The artificial intelligence revolution has completely changed many aspects of virtually every industry today. Organizations have already achieved permanent and transformative improvements and continue to envision more innovations and benefits.
At the same time, AI is so transformative that it also presents new questions, as many companies must develop strategies to take advantage of all that it has to offer, outperform their competition, or both.
The data center industry is no different and one subset of our business, data center design and construction, could be affected even more than most. Like any other company, we are used to responding to market demands and other external factors to constantly pivot when building traditional enterprise data centers.
Tony Corri
Permission granted by DataBank
If these past conditions represented previously predictable winds of change, now AI has created a perfect storm of innovation, opportunity, and transformation in designing and building the next generation of AI data centers. Practices and approaches that were once accepted for a long time have become obsolete, seemingly overnight, forcing data center operators and their customers to come up with new ways of thinking related to data center construction.
Flexibility of AI infrastructure
This new environment requires infrastructure flexibility that adapts to changing customer needs. That means offering multiple power, space and cooling configurations to fit your future computing needs without complete redesigns.
While this flexibility has so far served enterprise and hyperscale customers effectively, AI workloads now present a whole new dimension. Data center builders are now experiencing “hard right and hard left” requests from different customers. For example, one may need full air cooling for network infrastructure, while another is looking for a large liquid cooling rig for GPU-intensive operations.
These are not minor variations. They represent fundamentally different infrastructure approaches and entirely new construction strategies.
AI data centers require specialized approaches that go well beyond traditional design considerations and project management approaches. In some cases, they present logistical and technical complexities that did not exist recently.
Liquid cooling
The switch to liquid cooling represents one of the most dramatic design changes. Until recently, most traditional data centers used raised floors and perimeter cooling, but more intensive AI workloads are causing facilities to replace computer room air handling units with cooling distribution units. These systems use secondary chilled water loops made of welded stainless steel components.
This change creates a ripple effect that now requires specialized welding capabilities and stronger structural designs to support heavier thermal storage systems. These implications are affecting earlier construction timelines and processes. Trying to fit more specialized work into the same schedules while designing new structural designs is more difficult.
Supply chain constraints add another layer of complexity to building an AI data center. The above procurement approaches may be inadequate when trying to source specialized materials such as high quality copper components or stainless steel piping and control valves. As a result, we face the risk of facing limited availability and extended delivery times.
In response, data center builders must go two to three levels deeper in the supply chain than in the past. For example, instead of working exclusively with switchgear vendors, build direct relationships with switch manufacturers and cable suppliers. It’s proof of more dominoes falling: we need to anticipate facility requirements years in advance and start negotiating letters of intent for raw materials before the actual procurement happens.
This deeper engagement creates a partnership model where suppliers are now actively involved in project success rather than simply fulfilling purchase orders. This new approach still requires careful attention to economies of scale, but helps improve supply chain security that AI infrastructure timelines now demand.
Labor challenges
Skilled mechanical and electrical trades are very scarce today, creating the most difficult aspect of building an AI data center.
Estimates suggest that in North America an 439,000 additional new workers they must meet pending construction demand for data center projects. Further investigation showed the total number of unfilled skilled trade jobs could reach two million by 2033.
If not managed effectively, this widening gap can force construction timelines and costs into unpredictable territory where previous relationships with general contractors and current project management processes could fall short.
Builders must anticipate labor needs up to three years in advance and, where appropriate, bring contractors into pre-construction activities three to six months in advance. Instead of following conventional bidding procedures, owners should engage contractors in design evolution and resource planning before they begin.
The scale of this challenge is evident throughout the industry. Projects that begin with predictable labor costs often face dramatic changes as construction progresses. It is not unusual to see price increases of 20% to 30% for mechanical and electrical trades between phases of the same development.
This is just the new reality of hiring skilled workers from other markets with higher daily costs and competitive wages. It has reached the point where labor escalation is now recognized in multiple active developments and markets.
Adapting to the speed of the AI
One thing is clear: data center construction must adapt quickly to keep pace with the AI revolution. While these new challenges are unprecedented, they can also present new opportunities for operators who can identify new solutions early in the process.
Success now requires treating all aspects of construction, from supply chains to labor partnerships, as strategic investments rather than transactional relationships. As AI workloads continue to evolve, the data center operators that thrive will be those that see adaptation not as an obstacle, but as a potential differentiator in an industry that is being redefined in real-time.
