While the data center concept first took shape in the 1950s and 1960s, the growing number of large-scale data center construction projects currently underway across the US today are helping fuel the national economy. As more general contractors undertake these massive builds amid intense competition and pressure to get these centers up and running, understanding the hidden or unknown nuances of these types of projects can come with a steep learning curve.
With so much focus on the components that make the data center run, such as servers, equipment, and other infrastructure, it's imperative to consider challenges to the building itself. In effect, if the building fails, the internal components may be lost or severely damaged, which can lead to unexpected downtime and impacts to service. The result: catastrophic claims and lost business for the data centers as well as the scores of businesses and thousands of individual consumers who rely on the computing power.
To gain a clearer understanding of what's at stake and the related risk management measures needed to avoid failures and facilitate success, let's take a closer look at several key aspects of these projects, their related exposures, and remedies to protect these multimillion-dollar investments.
1. Building Size
The latest wave of mega-data-centers ranges in size from 20,000 to 100,000 square feet, with some projects having up to 10 times that footprint and more. While the building itself is large, the centers also require space; on average, today's data center requires an average of 224 acres of land. The sheer size of these facilities makes it impossible to monitor with circulating security teams alone for any one of the potential loss events that could occur, such as theft of valuable equipment or materials, or vandalism to water systems or the structure itself.
The massive scale calls for the careful installation of security cameras and motion detectors that can monitor for unapproved access to the jobsite and provide instant alerts of potential issues to security personnel and site management. Cameras can also be equipped with horns, strobe lights, or loudspeakers so 24/7 monitoring services can attempt to deter would-be trespassers. For extremely large projects, drones are also being used as a cost-effective option compared to perimeter cameras, as they can cover larger areas.
2. Lack of Qualified Workers
As one of the few construction industry segments expanding rapidly amid the current uncertain economic conditions, data center projects have attracted widespread attention from businesses seeking to capitalize on the financial opportunity.
The lure of these opportunities has brought numerous unqualified or underqualified contractors, subcontractors, and infrastructure and equipment manufacturers—among many others—to bid on, win, and accept jobs.
Labor shortages may also cause qualified contractors and manufacturers to hire underqualified workers to meet demand. However, the overall lack of experience can impact the project through such fundamental issues as poorly connected plumbing joints or incorrectly installed electrical lines.
To reduce the amount of rework and to mitigate against the heightened potential for exposures arising from inexperienced project participants, the adoption of Internet of Things (IoT) water prevention technology may be prudent as an added line of defense.
3. Massive Water Requirements
Data centers require immense quantities of water to provide critical cooling and heating, with large data centers using upward of 5 million gallons per day.
The miles of required water pipes and lines pose multiple threats to sensitive electronic equipment and servers: potential leaks or condensation on pipes causing drips or even conducting electrical current. These issues underscore the need for the installation of sophisticated water-detection sensors and devices that can quickly detect leaks and trigger instant notification to construction crews to address the situation. Small pucks can be placed in tight corners, smart tape with multiple probes can be placed in pits or up walls to monitor for quantities of water, and sensors with rope extensions can be placed on the floor around critical equipment bases.
As an added protection level, deployed water shutoff valves can be manually or remotely activated, stopping water flow when the potential for a loss event occurs. As multiple valves can be added to water lines, water flow shutoff can be isolated to specific areas without disrupting necessary water flow to the entire building.
4. Compressed Construction Schedules
With the spiraling demand for computing power throughout the US and in many areas of the world, along with the knowledge and experience accumulated through work on numerous projects in recent years, construction time frames for data center builds are getting tighter.
By some accounts, buildouts even for the largest projects have been reduced to 12–24 months on average compared to roughly twice as long in previous years. To keep these fast-tracked projects on schedule, every phase—from initial planning to design and permitting, procurement, construction (including installation of racking, computer hardware, and cooling systems), security and monitoring, and commissioning and testing—must be completed in strict accordance with accelerated timelines.
Under these intense pressures, the cost of data center construction delays can be exorbitant. According to Archilabs.ai, "… a one-month delay on a typical 60 MW data center can rack up roughly $14 million in extra costs." And with these abbreviated schedules, water is now live earlier than ever before in a build.
With schedules overlapping, properly testing water lines and plumbing is now occurring in parallel with other finishing work like drywall and the installation of expensive equipment and servers, decreasing the margin for error. When water is turned on, given widespread construction labor shortages and budget constraints, developers can't hire and position workers throughout the thousands of square feet of the facility to ensure no water leaks occur.
Deploying water, temperature, and humidity sensors allows the entire building to be monitored for direct water—such as a puddle—or indirect signs—such as increased humidity in a specific area. So, if an issue occurs in an obscure area after a routine site walk or during off-hours of construction, the team can be notified promptly and take action before it becomes an issue.
5. Type of Construction
Driven by hyperscale demand and condensed building timelines, data center construction has shifted in favor of modular and prefabricated delivery models. From fully engineered units to skid-mounted power and cooling assemblies, these modules are ready with critical mechanical, electrical, plumbing, and cooling systems, enabling repeatable quality for quick installation.
While these assemblies can speed up the data center operational timeline, the use of these units does not decrease the amount of risk; instead, they pose new potential issues. During build and transportation, moisture may be trapped within the structure, causing hidden damage or bringing unwanted moisture levels into the data center during installation. Modules may shift or settle, causing pipes, panels, or other connection points to not align as needed and create potential water pathways. Lastly, connecting and sealing the modules can be a complex project, as increased plumbing joints between connected modules adds more potential for leak points, often embedded within difficult-to-access structural elements.
As a solution, IoT sensors, such as smart mats, can be embedded during manufacturing, ensuring early risk detection. Other devices can be added before transit and storage to ensure continuous monitoring for adverse conditions, such as water and humidity. All these options provide future-proofing to ensure a quality build during the construction phase.
6. Heightened Risks of Renovations and Retrofits
Many data center projects involve renovating existing properties to take advantage of tax credits or incentives, or to speed up production. Yet, risk managers should beware of inheriting issues from the prior build.
This is an issue for any project; however, when data centers house several millions to billions of dollars in specialized equipment, a building failure can involve incredible costs. Placing IoT sensors throughout the shell facility in advance of the renovation work can provide architects with valuable insights on where to prioritize the redesign.
For instance, sensors can indicate where temperatures are hotter or cooler, or humidity is higher than expected, so further investigation can be done. Subsequently, the devices can be kept in place while the renovation construction work is being performed to continue to inform design and construction teams.
7. Unexpected Building Integrity Issues
Even though the building structure housing the data center may only be a shell, there is still the potential for damage to walls and foundation that can delay projects during construction or cause lingering issues postoccupancy.
For example, assume a 1 million square-foot bottling plant with a poured concrete floor and tilt-up walls experienced an issue during construction. While many of those involved in the project may have thought this basic shell was unlikely to experience any issues, an underground pipe broke over the weekend, causing nearly a foot of water to flood the structure, ultimately flowing out the door.
The concrete pad was compromised, even crumbling and broken in some areas, which resulted in a large insurance claim. The walls had already been placed; however, with a crumbling foundation, rework was required to maintain the overall integrity of the structure. It's worth noting that this overlooked exposure caused a 4-month delay on the project.
This type of risk applies to data centers as well, and the structural requirements are similar: a shell building with equipment added later. However, the costs of delays associated with data center projects can dwarf those of other types of facility builds. In this example, properly installed flow meters would have detected and alerted risk management and operations personnel of any water running during nonworking hours.
Furthermore, with the installation of shutoff valves, the team would have been able to remotely turn off the water within minutes, avoiding the issue altogether or minimizing the incident to water cleanup in a relatively small or isolated area instead of having to deal with flooding throughout the entire building and the disastrous consequences of related equipment damage.
8. Postconstruction Issues
Whether a project involves new construction or the renovation of an existing shell structure, undetected water leaks or water intrusion during the construction phase may have serious consequences months or years after a project is completed, fully equipped with technology, powered, and operational. When water flows and seeps into hidden spaces or overlooked areas, it may not be readily apparent to on-site workers or identified during visual inspections and walk-throughs.
The installation of appropriate IoT sensors can detect leaks as well as changes in humidity and conditions that would indicate water intrusion, flagging the potential issue before expensive equipment is installed and may be subject to damage. Not having a water event during construction leaves a healthier building when it's turned over for operations, bridging the gap between construction and occupancy. If an event does occur, additional resources can be added where necessary to continue monitoring to ensure a potential issue does not pop up again.
9. Reliability of Existing Monitoring Systems
Many data centers often have their own monitoring technology to alert to a potential issue. However, what happens if the equipment is faulty or nonoperational?
Once data centers are online, the multimillion-dollar equipment is operational 24/7; taking a data center offline has serious financial and operational consequences, not only for the operating company but for all businesses relying on those servers.
As recently as last year, issues at Microsoft and Amazon Web Services disrupted major systems across the globe. While not specifically caused by water, theft, or fire, these types of outages can be extrapolated to show the potential impact if a server system goes down due to one of these issues.
Leaving IoT technology in place helps to avoid a compromising situation. By having separate monitoring systems continue to function, the redundancy can help ensure continuous operations at optimal conditions. Protecting against a system failure in this manner may be a cost-effective option to ensure the continuity of critical business operations. In addition, the flexibility of IoT sensors enables them to be moved, added, or replaced as needed without any rewiring, so secondary system monitoring is always active.
10. Secondary Construction Needs
In addition to widely reported water needs, large-scale data centers and campuses have vast electrical power requirements. Builders and operators are tapping into nuclear power, solar, wind, thermal, oil, and gas to meet demands, which in some cases means constructing additional buildings and civil works on or near the data center site with their own mechanical, electrical, and plumbing needs.
These additional fuel sources also bring challenges that need to be monitored for potential issues, such as water, gas, or oil leaks. To mitigate potential exposures, specially designed IoT sensors can monitor not only water leaks but also particulates, oil, and gas, providing critical protection for the data centers at relatively modest costs.
Conclusion
The rapid development of numerous large-scale data centers and campuses in various parts of the US represents an exciting opportunity for all segments of the construction industry. However, these projects have unique, complex exposures that can involve exorbitant costs.
As they play a critical role in mitigating the varied and often hidden risks associated with these projects, risk managers can take advantage of the latest advances in IoT monitoring technology to keep projects on track and protect the high-value investments in computing power and artificial intelligence.
In this context, the value proposition is clear: An investment in a $200 sensor that detects a minor leak can pay huge dividends when compared with the alternative—an outage in a cooling system that can lead to millions of dollars in costs arising from a partial data center outage, as well as the substantial expense to replace damaged or destroyed parts. When measured against the cost of disruption, proactive monitoring is not just a safeguard; it's a strategic necessity.
Opinions expressed in Expert Commentary articles are those of the author and are not necessarily held by the author's employer or IRMI. Expert Commentary articles and other IRMI Online content do not purport to provide legal, accounting, or other professional advice or opinion. If such advice is needed, consult with your attorney, accountant, or other qualified adviser.
While the data center concept first took shape in the 1950s and 1960s, the growing number of large-scale data center construction projects currently underway across the US today are helping fuel the national economy. As more general contractors undertake these massive builds amid intense competition and pressure to get these centers up and running, understanding the hidden or unknown nuances of these types of projects can come with a steep learning curve.
With so much focus on the components that make the data center run, such as servers, equipment, and other infrastructure, it's imperative to consider challenges to the building itself. In effect, if the building fails, the internal components may be lost or severely damaged, which can lead to unexpected downtime and impacts to service. The result: catastrophic claims and lost business for the data centers as well as the scores of businesses and thousands of individual consumers who rely on the computing power.
To gain a clearer understanding of what's at stake and the related risk management measures needed to avoid failures and facilitate success, let's take a closer look at several key aspects of these projects, their related exposures, and remedies to protect these multimillion-dollar investments.
1. Building Size
The latest wave of mega-data-centers ranges in size from 20,000 to 100,000 square feet, with some projects having up to 10 times that footprint and more. While the building itself is large, the centers also require space; on average, today's data center requires an average of 224 acres of land. The sheer size of these facilities makes it impossible to monitor with circulating security teams alone for any one of the potential loss events that could occur, such as theft of valuable equipment or materials, or vandalism to water systems or the structure itself.
The massive scale calls for the careful installation of security cameras and motion detectors that can monitor for unapproved access to the jobsite and provide instant alerts of potential issues to security personnel and site management. Cameras can also be equipped with horns, strobe lights, or loudspeakers so 24/7 monitoring services can attempt to deter would-be trespassers. For extremely large projects, drones are also being used as a cost-effective option compared to perimeter cameras, as they can cover larger areas.
2. Lack of Qualified Workers
As one of the few construction industry segments expanding rapidly amid the current uncertain economic conditions, data center projects have attracted widespread attention from businesses seeking to capitalize on the financial opportunity.
The lure of these opportunities has brought numerous unqualified or underqualified contractors, subcontractors, and infrastructure and equipment manufacturers—among many others—to bid on, win, and accept jobs.
Labor shortages may also cause qualified contractors and manufacturers to hire underqualified workers to meet demand. However, the overall lack of experience can impact the project through such fundamental issues as poorly connected plumbing joints or incorrectly installed electrical lines.
To reduce the amount of rework and to mitigate against the heightened potential for exposures arising from inexperienced project participants, the adoption of Internet of Things (IoT) water prevention technology may be prudent as an added line of defense.
3. Massive Water Requirements
Data centers require immense quantities of water to provide critical cooling and heating, with large data centers using upward of 5 million gallons per day.
The miles of required water pipes and lines pose multiple threats to sensitive electronic equipment and servers: potential leaks or condensation on pipes causing drips or even conducting electrical current. These issues underscore the need for the installation of sophisticated water-detection sensors and devices that can quickly detect leaks and trigger instant notification to construction crews to address the situation. Small pucks can be placed in tight corners, smart tape with multiple probes can be placed in pits or up walls to monitor for quantities of water, and sensors with rope extensions can be placed on the floor around critical equipment bases.
As an added protection level, deployed water shutoff valves can be manually or remotely activated, stopping water flow when the potential for a loss event occurs. As multiple valves can be added to water lines, water flow shutoff can be isolated to specific areas without disrupting necessary water flow to the entire building.
4. Compressed Construction Schedules
With the spiraling demand for computing power throughout the US and in many areas of the world, along with the knowledge and experience accumulated through work on numerous projects in recent years, construction time frames for data center builds are getting tighter.
By some accounts, buildouts even for the largest projects have been reduced to 12–24 months on average compared to roughly twice as long in previous years. To keep these fast-tracked projects on schedule, every phase—from initial planning to design and permitting, procurement, construction (including installation of racking, computer hardware, and cooling systems), security and monitoring, and commissioning and testing—must be completed in strict accordance with accelerated timelines.
Under these intense pressures, the cost of data center construction delays can be exorbitant. According to Archilabs.ai, "… a one-month delay on a typical 60 MW data center can rack up roughly $14 million in extra costs." And with these abbreviated schedules, water is now live earlier than ever before in a build.
With schedules overlapping, properly testing water lines and plumbing is now occurring in parallel with other finishing work like drywall and the installation of expensive equipment and servers, decreasing the margin for error. When water is turned on, given widespread construction labor shortages and budget constraints, developers can't hire and position workers throughout the thousands of square feet of the facility to ensure no water leaks occur.
Deploying water, temperature, and humidity sensors allows the entire building to be monitored for direct water—such as a puddle—or indirect signs—such as increased humidity in a specific area. So, if an issue occurs in an obscure area after a routine site walk or during off-hours of construction, the team can be notified promptly and take action before it becomes an issue.
5. Type of Construction
Driven by hyperscale demand and condensed building timelines, data center construction has shifted in favor of modular and prefabricated delivery models. From fully engineered units to skid-mounted power and cooling assemblies, these modules are ready with critical mechanical, electrical, plumbing, and cooling systems, enabling repeatable quality for quick installation.
While these assemblies can speed up the data center operational timeline, the use of these units does not decrease the amount of risk; instead, they pose new potential issues. During build and transportation, moisture may be trapped within the structure, causing hidden damage or bringing unwanted moisture levels into the data center during installation. Modules may shift or settle, causing pipes, panels, or other connection points to not align as needed and create potential water pathways. Lastly, connecting and sealing the modules can be a complex project, as increased plumbing joints between connected modules adds more potential for leak points, often embedded within difficult-to-access structural elements.
As a solution, IoT sensors, such as smart mats, can be embedded during manufacturing, ensuring early risk detection. Other devices can be added before transit and storage to ensure continuous monitoring for adverse conditions, such as water and humidity. All these options provide future-proofing to ensure a quality build during the construction phase.
6. Heightened Risks of Renovations and Retrofits
Many data center projects involve renovating existing properties to take advantage of tax credits or incentives, or to speed up production. Yet, risk managers should beware of inheriting issues from the prior build.
This is an issue for any project; however, when data centers house several millions to billions of dollars in specialized equipment, a building failure can involve incredible costs. Placing IoT sensors throughout the shell facility in advance of the renovation work can provide architects with valuable insights on where to prioritize the redesign.
For instance, sensors can indicate where temperatures are hotter or cooler, or humidity is higher than expected, so further investigation can be done. Subsequently, the devices can be kept in place while the renovation construction work is being performed to continue to inform design and construction teams.
7. Unexpected Building Integrity Issues
Even though the building structure housing the data center may only be a shell, there is still the potential for damage to walls and foundation that can delay projects during construction or cause lingering issues postoccupancy.
For example, assume a 1 million square-foot bottling plant with a poured concrete floor and tilt-up walls experienced an issue during construction. While many of those involved in the project may have thought this basic shell was unlikely to experience any issues, an underground pipe broke over the weekend, causing nearly a foot of water to flood the structure, ultimately flowing out the door.
The concrete pad was compromised, even crumbling and broken in some areas, which resulted in a large insurance claim. The walls had already been placed; however, with a crumbling foundation, rework was required to maintain the overall integrity of the structure. It's worth noting that this overlooked exposure caused a 4-month delay on the project.
This type of risk applies to data centers as well, and the structural requirements are similar: a shell building with equipment added later. However, the costs of delays associated with data center projects can dwarf those of other types of facility builds. In this example, properly installed flow meters would have detected and alerted risk management and operations personnel of any water running during nonworking hours.
Furthermore, with the installation of shutoff valves, the team would have been able to remotely turn off the water within minutes, avoiding the issue altogether or minimizing the incident to water cleanup in a relatively small or isolated area instead of having to deal with flooding throughout the entire building and the disastrous consequences of related equipment damage.
8. Postconstruction Issues
Whether a project involves new construction or the renovation of an existing shell structure, undetected water leaks or water intrusion during the construction phase may have serious consequences months or years after a project is completed, fully equipped with technology, powered, and operational. When water flows and seeps into hidden spaces or overlooked areas, it may not be readily apparent to on-site workers or identified during visual inspections and walk-throughs.
The installation of appropriate IoT sensors can detect leaks as well as changes in humidity and conditions that would indicate water intrusion, flagging the potential issue before expensive equipment is installed and may be subject to damage. Not having a water event during construction leaves a healthier building when it's turned over for operations, bridging the gap between construction and occupancy. If an event does occur, additional resources can be added where necessary to continue monitoring to ensure a potential issue does not pop up again.
9. Reliability of Existing Monitoring Systems
Many data centers often have their own monitoring technology to alert to a potential issue. However, what happens if the equipment is faulty or nonoperational?
Once data centers are online, the multimillion-dollar equipment is operational 24/7; taking a data center offline has serious financial and operational consequences, not only for the operating company but for all businesses relying on those servers.
As recently as last year, issues at Microsoft and Amazon Web Services disrupted major systems across the globe. While not specifically caused by water, theft, or fire, these types of outages can be extrapolated to show the potential impact if a server system goes down due to one of these issues.
Leaving IoT technology in place helps to avoid a compromising situation. By having separate monitoring systems continue to function, the redundancy can help ensure continuous operations at optimal conditions. Protecting against a system failure in this manner may be a cost-effective option to ensure the continuity of critical business operations. In addition, the flexibility of IoT sensors enables them to be moved, added, or replaced as needed without any rewiring, so secondary system monitoring is always active.
10. Secondary Construction Needs
In addition to widely reported water needs, large-scale data centers and campuses have vast electrical power requirements. Builders and operators are tapping into nuclear power, solar, wind, thermal, oil, and gas to meet demands, which in some cases means constructing additional buildings and civil works on or near the data center site with their own mechanical, electrical, and plumbing needs.
These additional fuel sources also bring challenges that need to be monitored for potential issues, such as water, gas, or oil leaks. To mitigate potential exposures, specially designed IoT sensors can monitor not only water leaks but also particulates, oil, and gas, providing critical protection for the data centers at relatively modest costs.
Conclusion
The rapid development of numerous large-scale data centers and campuses in various parts of the US represents an exciting opportunity for all segments of the construction industry. However, these projects have unique, complex exposures that can involve exorbitant costs.
As they play a critical role in mitigating the varied and often hidden risks associated with these projects, risk managers can take advantage of the latest advances in IoT monitoring technology to keep projects on track and protect the high-value investments in computing power and artificial intelligence.
In this context, the value proposition is clear: An investment in a $200 sensor that detects a minor leak can pay huge dividends when compared with the alternative—an outage in a cooling system that can lead to millions of dollars in costs arising from a partial data center outage, as well as the substantial expense to replace damaged or destroyed parts. When measured against the cost of disruption, proactive monitoring is not just a safeguard; it's a strategic necessity.
Opinions expressed in Expert Commentary articles are those of the author and are not necessarily held by the author's employer or IRMI. Expert Commentary articles and other IRMI Online content do not purport to provide legal, accounting, or other professional advice or opinion. If such advice is needed, consult with your attorney, accountant, or other qualified adviser.