The need for data centers has grown significantly in recent years to meet demand for critical IT applications, data storage, network infrastructure, and other computing resources ubiquitous in digital commerce and social networking. Gartner, an IT research company, projects that more than $200 billion will be spent on data center infrastructure worldwide in 2021. The market is expected to continue to grow for the foreseeable future.
The expanded use of the public cloud and data center technology advancements, as well as long-term technology advancements in areas such as artificial intelligence and autonomous cars, are driving increases in data center capital spending, and in particular, in hyperscale facilities. According to a January 2021 report by Synergy Research Group, “the total number of large data centers operated by hyperscale providers increased to 597 at the end of 2020, having more than doubled since the end of 2015.” According to the same report, “The companies with the broadest data center footprint are the leading cloud providers – Amazon, Microsoft, Google and IBM. “
The massive scale of investment on data centers globally is on par with that of other capital-intensive industries. However, compared to the long history of other capital-intensive industries—energy, chemicals, consumer goods, etc., where it is not out of the question to encounter a 100 year old facility— the data center construction industry is still nascent. The data center construction industry has an opportunity to learn from the successes and challenges of more established industries. Since its founding in 1987, IPA has worked with owner companies in capital-intensive industries to capture lessons learned, to correlate statistical relationships between project practices and results, and to improve the performance of project systems around the world. From opportunity shaping, project governance, cost estimating, project development and execution, team development, and contracting, IPA understands the issues that can improve the on-time and cost efficient delivery of data center construction projects. Some critical areas include:
Predictability of Project Outcomes
As investment in hyperscale facilities continues to grow, the pressure for accurate cost estimates and schedules will only increase. If a small investment falls behind schedule, sometimes additional resources can be added to the project to recover schedule, or the demand can be met by a different facility. Recovering schedule on a multi-billion project is much more challenging. Similarly, 10 percent cost overrun on a 30 MW facility may not be meaningful for a large company, but a 10 percent cost overrun on a 300 MW facility is much more likely to be scrutinized. IPA data on Best Practices for developing early, conceptual-level cost and schedule targets have helped many owner companies set and meet competitive schedule and cost targets.
Knowledge of Local Conditions
Data center developers are already well aware of the need for reliable and cost-effective power when siting their projects. As owners build data centers in areas of the world where they have not constructed assets before (and even in familiar parts of the world), they need to understand context around local conditions beyond just power supply. Such conditions include the clarity and stability of the local regulatory/political environment, availability of local labor, the contractual preferences of the local construction market, and the presence of unions. IPA has seen many projects underestimate the cost and complexity of developing a greenfield project in a new-to-company country.
Data centers have not experienced the same degree of community objection/protest as projects in industries such as mining or pipelines. But some communities have registered data center construction concerns regarding the noise and emissions of diesel generators, the increased presence of above ground power lines and large diesel storage tanks, and various right of way (ROW) and land use issues. In a few cases, community protests have led to rejected or modified operating permits. As data centers grow in size and cost and spread to new locations, owners should expect community awareness, regulatory scrutiny, and (potentially) community protest/activism, to grow as well. IPA data show that early identification and alignment with all key stakeholders is a critical success factor.
As facility size grows, field labor exposure to safety incidents grows. Owner and contractor safety strategies, practices, and required resources may have to change or be supplemented to reduce safety risk. Safety performance influences not only the bottom line but has very strong reputational effects as well.
Hyperscale data centers have the opportunity to use pre-fabricated power and cooling modules. The increased use of modularization has several benefits. Often cited benefits include increased productivity associated with a better-controlled environment and a compressed construction schedule given the use of multiple work fronts. But the experience of other industries shows that data center owners have to anticipate the challenges that come with modularization strategies. For example, productivity savings are often offset by additional transportation and structural steel costs, more upfront engineering is required (e.g., envelope sizing, logistics planning, etc.), and the engineering/procurement sequence must be developed earlier and is less amenable to late change.
Data center owners also have started to embrace facilities standardization as a way to accelerate schedule, increase investment predictability, and reduce capital costs. IPA’s research has found significant benefits (10 to 15 percent cost savings) from the use of standardization in the oil and gas world. These projects, on average, also incur less schedule slip. The primary challenges data center owners are likely to face with standardization include: ensuring continuity of both the design and the owner/contractor team from one project to the next; avoiding execution changes; and ensuring that the scope that is not standardized (often civil or supporting utility elements) is thoroughly defined for each successive project.
As owners invest larger sums in data centers, planning will become more difficult, yet the rewards of improved planning will be magnified as even small gains in efficiency tied to better planning (and the avoidance of late change) will lead to significant construction schedule and cost savings. Bigger data centers will mean more people, more material, more complicated logistics, and more risk. Data center owners will have to become even stronger at developing robust schedules, detailed logistics and procurement plans, and execution plans.
IPA’s database of more than 20,000 projects executed by hundreds of global owners put us in a unique position to apply the lessons learned from the energy, chemicals, mining, pharmaceutical, transportation, and other industries to the data center market. The very large spenders of capital dollars with whom IPA routinely works need their projects to be on time, on budget, and safe. They routinely expand their markets to new areas and face knowledge deficits of local conditions. They face significant regulatory scrutiny and often encounter community activism and protest toward their intended investments. They too are increasingly seeking efficiency improvements through the use of both modularization and standardization. In other words, many key issues that affect data center project results are not industry-specific. They are common in one form or another to almost all large construction projects.
The data center industry, as a nascent and growing industry, has the opportunity to learn from the challenges of other industries and leapfrog some of these issues. IPA’s breadth and depth of lessons learned from other capital-intensive industries uniquely positions us to counsel data center owners and investors, to improve their organizations, their teams, their projects, and their results.