Data Center Builds: Not your Standard Industrial Development

By Mitchell Osborne, AP Director of MEP

The growing interest in and use of artificial intelligence prompted Goldman Sachs to report that data center power demand will increase 50% globally and could expand up to 165% by the end of the decade. Meanwhile, McKinsey anticipated that data center capacity on a global scale could triple by 2030. The simple solution to this seemingly insatiable need is to build more data centers. But in our experience, it demands a detailed plan addressing site selection, contractors and strict timelines for successful operation.

A successful data center design-build requires the following:

#1 The Site

Data Center Chillers

An ideal data center location offers three things:

  • A large footprint to hold the center, parking and auxiliary requirements (like HVAC set-ups)
  • Efficient water connections to support chillers that prevent the overheating of racks and servers
  • A nearby energy source to provide enough power for ongoing operations

The above requirements mean that dense urban areas, such as those on the East Coast, might not be ideal for data center sites, due to land scarcity and strict zoning regulations. We’ve found that Arizona, Colorado, North Dakota, Northern California and Texas offer enough land to house these complex facilities. An ability to tap into power and water infrastructures is also a must for data center operations.

Long before we started dirt work on a 325,000-square-foot North Texas data center, we worked with GIS experts to determine size and infrastructure requirements. We also took the time to build relationships with local utilities and municipal and state governments.

Additional options to ensure enough power and water include cogeneration, in which utilities partner with one another, reducing reliance on a single local power and/or water supply.

Technologies are also available. For instance, immersive cooling, which is a server “bathtub” with non-conductive or corrosive fluid, absorbs server-generated heat. The result is a decrease in wasted cooling and utility overload prevention.

#2 The Organization Plan

Most design-builds don’t operate in a vacuum, and this is especially true with data centers. The process requires multiple stakeholders, including key MEP trade partners, engineers, CxA, general contractors, city councils, planning boards and community leaders. Sometimes, these groups have different agendas, which can lead to confusion and costly delays.

At AP, we made sure to get everyone together and on the same page before doing site work on the North Texas data center. As construction continues on the facility, we’ve found that frequent meetings and ongoing communication have eliminated confusion and assisted with hidden obstacles. One such challenge was the need for temporary power, lighting and roofing plans while going north on the center.

Additionally, forming the various stakeholders into a team was essential for deciding where spotters and signage should be placed.

Safety always comes first, and the team’s recommendations ensured the high-risk engagement with cranes and large equipment was well-managed.

#3 The Staff

By now, it’s clear that building a data center isn’t like creating a warehouse or logistics center. As such, the best option is to find trade partners and suppliers who are trained in the niche. It is possible to employ a quality-focused, technologically experienced construction team of fast learners. But your most valuable partners in a data center design-build are those with in-depth knowledge and expertise in this area.

For the North Texas data center build, we scheduled experienced data center trade partners and suppliers far ahead of time to ensure their availability. We also added others to the team who quickly picked up the ins and outs of data center construction. Also helpful is an internal MEP systems expert, who can provide a holistic pathway for the project.

#4 Speed to Market

There is no such thing as a delayed timeline when designing and building a data center. These facilities are built-to-suits covered by inflexible leases due to downstream effects on developers and owners. If that center isn’t ready for occupancy by the lease’s start date, the developer could end up paying millions of dollars a week in rent until construction completion.

To avoid this costly problem, use the lease commencement date as an in-stone deadline while paying attention to the following:

  • Develop and maintain a disciplined schedule that includes all imagined risks and contingencies (like weather or illnesses) that could lead to delays.
  • Properly manage owner-furnished, contract-installed (OFCI) materials.
  • Implement rotating schedules and develop makeup/overtime strategies.
  • Reduce building materials, such as mechanical and electrical materials, with other options. For example, removing a raised-access floor and implementing new routes for electrical or low-voltage distribution can free several weeks from a schedule.
  • Consider prefab building materials. This process means structural components are developed in off-site factories, then delivered to the site for quick assembly.

Start-up and commissioning is another element that must be considered from the start. Providing enough time and being thorough from receipt to start up on all the inspections and allowing enough time in the schedule for issue resolution and external vendor support.

#5 The Economy

Continually changing tariff policies, foreign trade confusion and economic volatility impacts material prices. This is true for data center construction, which is heavily dependent on component and equipment imports from Canada, Mexico and Korea. Certainly, manufacturing is moving back to the U.S.. However, raw materials continue to be mined and refined overseas.

Higher prices can wreck budgets and cause some developers to pull the trigger on construction. Delays in overseas shipping while trade details are being hammered out mean more change orders, which could delay the project.

In planning and building the North Texas data center, we handled these issues by:

  • Locking in equipment and building materials prices as far ahead as possible.
  • Buying in bulk to take advantage of discount pricing.
  • Putting a standard in place allowing off-the-shelf material purchasing.
  • Asking key trade partners to stockpile material for quick access when needed.

A Different Design-Build

There is plenty of justification for building more data centers. But doing so requires a different mindset than constructing a tilt-wall manufacturing or storage facility. Planning far enough in advance is essential to ensure the process is smooth, from site selection to construction completion. Similarly, adaptability is also necessary in successful data center projects, as change frequently happens. Examples include electrical/mechanical designs needing to change mid-construction or lead-time issues requiring utilizing different materials or equipment. We have to think outside the box to use available resources to keep the project moving forward.

More power will be needed to drive artificial intelligence, computing and other technologies. Developers interested in providing that power through well-constructed data centers must have the knowledge and capacity to meet this greater demand, or partner with those that do.

Article Author

As Director of MEP, Mitchell brings deep technical expertise and hands-on leadership to every project, applying his extensive knowledge of mechanical, electrical, and controls systems from preconstruction through commissioning. He is responsible for developing strategies that support early-access power and mechanical system energization, enabling efficient equipment checkout and setting projects up for a streamlined Level 3 startup once permanent power is achieved. Mitchell believes operational excellence is rooted in disciplined planning and collaboration, and he applies these principles to support his teams and reinforce AP’s core values.

Mitchell plays a critical role in AP’s project success, leading Level 4 and Level 5 commissioning efforts and guiding teams through integrated testing and project turnover. His fast-paced, detail-driven startup scheduling ensures on-time delivery and early resolution of vendor-related challenges, supporting smooth transitions into system operation. Mitchell’s commitment to excellence and his proactive, solution-oriented approach strengthens AP’s relationships with clients and partners across markets, contributing to reliable performance and repeat business.

Share this article.
Share On Facebook Share On Facebook Share On LinkedIn Share On LinkedIn

Related AP Insights