WE CAN BUILD A BETTER DATA CENTER.
Innovation shouldn’t stop at the technology inside the building.
Sentinel Metalworks
The modern world runs on digital infrastructure.
The applications we use, the businesses we operate, the communications we depend on, scientific
research, cloud computing, and the rapid advancement of artificial intelligence all rely on an
increasingly vast network of physical infrastructure.
And that infrastructure is growing.
The International Energy Agency projects that global electricity consumption from data centers is
projected to double to approximately 945 terawatt-hours by 2030 in its Base Case. That growth presents enormous opportunities—but it also creates legitimate questions about power, water, land, infrastructure, and the communities where these facilities are built.
Those questions should not be dismissed.
But neither should they force us into a false choice between technological progress and responsible development.
We can do both.
In fact, we believe we have to.
We can build a better data center.
Not by pretending that every challenge has a simple solution. Not by claiming that every facility should look the same. And certainly not by assuming that technological growth comes without consequences.
We can do it by asking better questions—and bringing more people into the process of answering them.
The Conversation Should Be Bigger Than the Building
When people think about data centers, they often think about servers, processors, artificial intelligence, and the technology operating inside the facility.
But a data center is also a highly complex physical system. It requires:
• Power infrastructure
• Cooling systems
• Structural support
• Electrical equipment
• Mechanical systems
• Equipment enclosures
• Piping and fluid systems
• Backup power
• Fire protection
• Physical security
• An enormous network of manufactured components
Every one of those systems presents an opportunity to think differently.
The question shouldn’t simply be:
How quickly can we build the next data center?
It should also be:
How can we build the next one better than the last?
Water: Can We Cool More Intelligently
Water consumption has become one of the most significant concerns surrounding data-center
development.
That concern is legitimate, particularly in regions already experiencing water stress.
However, there is an important distinction that often gets lost in broad public conversations: there is no single number that accurately describes the water impact of every data center.
Water use can vary significantly depending on factors including:
• The cooling technology being used
• The climate
• The efficiency of the equipment
• The utilization of the servers
• The source of electricity
• The local infrastructure
• The specific design of the facility
Recent research from Lawrence Berkeley National Laboratory found that workload-level water use can vary by more than 10,000-fold, with major determinants including server efficiency, electrical-grid water consumption, server utilization, cooling-system type, infrastructure efficiency, climate zone, inactive server percentage, and server refresh cycle. The study concludes that there is no single recipe for minimizing water use; optimal outcomes depend on site-specific combinations of these factors.
That should not discourage innovation.
It should encourage more thoughtful design.
Opportunities being explored across the industry include improved cooling efficiency, direct liquid
cooling, closed-loop systems, water reuse, reclaimed water, captured condensate, rainwater harvesting, and other approaches appropriate to specific locations and operating conditions.
The U.S. Department of Energy has also identified multiple opportunities to improve cooling-water
efficiency, including air-side and water-side economizing, maximizing cycles of concentration, reverse-osmosis treatment in appropriate applications, and direct liquid cooling.
The answer will not be the same everywhere.
And that’s precisely the point.
Better infrastructure requires solutions designed for the environment in which that infrastructure actually exists.
Energy: Efficiency Must Be Part of the Design
The conversation surrounding data centers and electricity is equally complex.
Digital infrastructure requires significant amounts of power, and rapid growth can create real challenges for utilities and electrical infrastructure—particularly because large data centers can concentrate substantial demand in relatively small geographic areas.
The IEA projects strong growth in data-center electricity demand through the end of the decade and notes that the broader energy system requires significant planning, infrastructure investment, and time to accommodate that growth.
But increasing demand doesn’t mean efficiency should become an afterthought.
It should become a design principle.
The Department of Energy’s approach to efficient data-center design prioritizes improving the efficiency of IT equipment, electrical systems, and cooling systems before considering additional opportunities such as heat reuse and water-saving heat rejection.
That principle deserves attention beyond the data center itself.
• How efficiently are components manufactured?
• How efficiently can they be transported and installed?
• Can systems be designed to reduce unnecessary material?
• Can maintenance and replacement be simplified?
• Can infrastructure be designed for future upgrades instead of complete replacement?
Efficiency is not one decision.
It is the cumulative result of thousands of design decisions.
What Are We Doing With All That Heat?
A data center’s fundamental engineering challenge is also remarkably straightforward:
Technology generates heat, and that heat must go somewhere.
For decades, the primary goal has been to remove that heat as efficiently as possible.
But what if, in the right circumstances, some of it could become a resource?
The Department of Energy has identified opportunities for recovering and reusing data-center waste heat, particularly when a suitable nearby heat user exists. Depending on location and system design, recovered heat could potentially support applications such as water heating, district heating, or other low-temperature heating needs.
This is not a universal solution.
Distance matters. Temperature matters. Economics matter. Local infrastructure matters.
But it is exactly the type of question we should be asking.
Instead of viewing every byproduct as waste, can we identify opportunities to create additional value?
Sometimes the answer will be yes.
Sometimes it won’t.
Innovation requires being willing to investigate the question anyway.
What If We Designed Data Centers to Be Built Better?
This is where we believe manufacturers have an important voice.
Too often, manufacturing expertise enters the conversation after the fundamental design decisions have already been made.
An engineer designs a component.
A project requires it.
Then a manufacturer is asked:
Can you build this?
But there is another question that can be just as valuable:
What would happen if manufacturers were involved earlier?
Manufacturers understand the practical realities of turning a design into something physical.
We understand materials.
Fabrication.
Assembly.
Tolerances.
Transportation.
Installation.
Maintenance.
Production capacity.
And, perhaps most importantly, the difference between something that can be built and something that can be built efficiently, consistently, and at scale.
Earlier collaboration between designers, engineers, contractors, operators, and manufacturers could create opportunities to explore:
• Modular infrastructure
• Prefabricated assemblies
• More efficient material utilization
• Standardized components where appropriate
• Designs that simplify maintenance
• Easier replacement and upgrades
• Reduced construction waste
• Improved transportation and installation efficiency
The goal shouldn’t be to standardize everything.
Some infrastructure requires highly customized solutions.
The goal should be to recognize that manufacturability is part of good design.
Responsible Growth Doesn’t Mean Stopping Growth
There are legitimate concerns surrounding rapid infrastructure development.
Communities should be asking questions.
They should understand how proposed facilities could affect local resources and infrastructure.
Industry should listen.
Government, utilities, developers, technology companies, manufacturers, and communities all have
perspectives worth bringing to the table.
But responsible development and technological progress are not mutually exclusive.
They cannot be.
The world is becoming increasingly digital, and the physical infrastructure supporting that digital world will continue to evolve.
The real opportunity is not simply to build more infrastructure.
It is to continuously improve how we build it.
To conserve resources where possible.
To improve efficiency.
To invest in better technology.
To design with long-term operation in mind.
To listen to communities.
And to bring together people with different expertise before decisions become permanent.
We Can Build a Better Data Center.
That statement isn’t a claim that one company has all the answers.
We don’t.
It is a belief that the people responsible for designing, building, operating, and supporting critical
infrastructure have an obligation to continue looking for better answers.
The technology inside a data center may evolve rapidly.
The infrastructure surrounding it must evolve too.
At Sentinel Metalworks, we believe manufacturers have an important role to play in that evolution.
We work in the physical world.
The world of materials, structures, fabrication, equipment, and production.
We understand that innovation isn’t limited to software, processors, and artificial intelligence.
Sometimes innovation is a better bracket.
A more efficient enclosure.
A smarter structural design.
A component that lasts longer.
A system that is easier to maintain.
A manufacturing process that creates less waste.
Or simply asking a question that hasn’t been asked often enough.
What if we challenged ourselves to build it better?
We believe that question is worth asking.
And we believe the answer can be yes.
We can build a better data center.
SOURCES & FURTHER READING
The factual claims in this article are supported by the following primary or authoritative sources.
Accessed September 2026.
International Energy Agency (IEA)
Energy and AI — Energy Demand from AI
https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai IEA analysis of global data-center electricity demand, including the Base Case projection of
approximately 945 TWh by 2030.
International Energy Agency (IEA)
Energy and AI — Executive Summary
Overview of the report’s findings concerning AI, data-center electricity demand, energy infrastructure and system planning.
Lawrence Berkeley National Laboratory
The Water Use of Data Center Workloads: A Review and Assessment of Key Determinants
2025 peer-reviewed study examining workload-level water use and the factors driving substantial
variation among data-center workloads and sites.
U.S. Department of Energy — Federal Energy Management Program
Cooling Water Efficiency Opportunities for Federal Data Centers
DOE guidance covering cooling-water efficiency, economizing strategies, cycles of concentration,
reverse osmosis and direct liquid cooling.
U.S. Department of Energy
Best Practices Guide for Energy-Efficient Data Center Design
DOE guidance addressing IT, electrical and cooling efficiency, water efficiency and opportunities for
data-center waste-heat recovery.
U.S. Environmental Protection Agency
Water Reuse for Industrial Applications Resources
EPA resources addressing industrial water reuse, including data-center cooling and potential alternative water sources.
U.S. Environmental Protection Agency
Water Reuse Case Study: Quincy, Washington
EPA case study documenting a large-scale data-center cooling-water reuse project in Quincy,
Washington.
Sentinel Metalworks
Building the physical infrastructure behind what powers the future.
SENTINEL METALWORKS™ | THOUGHT LEADERSHIP
Comments