Support the Building of Data Centers

Introduction

As a believing and practicing Christian, we are all called to steward the Earth that God has given to us. This was the initial calling that God gave to Adam after He gave authority to him over all the plants and animals of the Earth. (Genesis 1:26) This calls all people to be good stewards of the environment and it encourages us to be more mindful of our actions. Don't litter, try to drink out of metal water bottles instead of plastic ones, and conserve more energy. These are all things we've heard before and are constantly told by our peers and our government as the right thing to do.

So, it might shock some people when I say that I support the building of data centers in the United States. I believe that we should build as many data centers as possible to handle the compute power necessary for powering Artificial Intelligence, a technology that is changing the world as we know it. I believe that a lot of the concerns and protests that have erupted over the past few months are seriously misguided and are reacting based on unconfirmed statistics about environmental health.

This is exactly what we're going to be talking about today. Why it's necessary for the United States to be building more data centers and how we can do it sustainably.

What is compute?

Before we get started, we need to define our terms. According to the official engineering definition handbook of structural frameworks (ISO/IEC-22237-1) a data center is "A fault-tolerant facility engineered to maintain continuous operation of high-density IT equipment through redundant power distribution, precision thermal management, and secure network connectivity, operating within strict environmental parameters." It's essentially filled with IT equipment like servers, storage equipment, and cooling systems. These are essential for large IT companies such as Google, Meta, Microsoft, and Apple and they have been around for years, long before this debate has started.

When we talk about AI and data centers, the big topic is "compute" and how much it will cost to actually run Large Language Models. Compute is defined by "The computational resources-specifically processing power, memory, and architectural infrastructure-required for a computer, program, or machine learning model to function and execute software instructions." (Hewlett Packaging Enterprise, 2025) This is essentially how much computational power is necessary to perform. The majority of AI models require millions of mathematical calculations per second which can only be done through GPUs (Graphics Processing Unit), a specialized circuit unit that can handle multiple calculations with different cores at the same time.

These GPUs can be located in data centers and serve as the main power driver for the AI boom we are seeing right now. Companies like NVIDIA, AMD, and Intel are the ones designing these GPUs and dumping them onto the market for tech companies to buy up for more computational power. The measure of compute power is FLOP (Floating-Point-Operation) but we won't get into any of that today. The point is, there are millions of GPUs processing trillions of mathematical equations every single day, and we are only at the start of the AI revolution.

ChatGPT currently reports having over a billion users weekly, but only about 5% of all users actually have a paid subscription whether it's plus, pro, or enterprise. (OpenAI, 2026) This means that if we are going to see more people purchasing higher tier subscriptions, this means more compute power is needed, more GPUs needed to process the information, and then ultimately, more data centers to hold these essential machines. Stopping our building of data centers right now would halt the rise of AI in our world, something that should not be suppressed, especially for something this world-changing.

The numbers behind the power.

There's no denying that data centers take up a lot of power and energy. The Lawrence Berkeley National Laboratory released a study recently on data centers and how much power they use currently. They stated how energy usage has increased exponentially on data centers since 2022. Another study from a government agency said that data centers "Currently consumed approximately 176 terawatt-hours of electricity in 2023, representing 4.4% of all U.S. electricity consumption ... .data-center use reaching 649 TWh by 2030, or about 11.8% of U.S. electricity consumption." (Center of Enterprise for Data Center Usage, 2024)

The ways that these data centers actually run and are actually powered is through a variety of different methods. The energy sources that power these massive buildings stem from natural gas (40%), renewable energy (wind/solar, 25%), nuclear power (20%), and coal (15%). (International Energy Agency, 2026) Nuclear energy is becoming more popular with large tech firms like Microsoft who are actually investing more money to reopen closed nuclear reactors and power the data centers they are invested in.

Now, that's all fine and good. But why is it that we, as the American people, are always hearing that all of our drinking water is going to be used up by these centers? Why wasn't water one of the main energy sources listed above? This is due to the fact that water being used is not powering the centers, they are cooling down the overheating servers. Currently, the majority of data centers are using clean drinking water from public taps (Friends of the Earth, 2026), but tech companies are actively switching over to unclean and undrinkable water to cool their systems. The water being used is evaporated and used to cool the servers in the data centers, and the amount of water being used is actually decreasing throughout the last few years.

This is important. According to Berkeley Lab, compute demand in data centers increased 550% from 2010 to 2018. However, energy consumption within data centers only increased 6% in that same time period, with energy used per compute fell by 20% in the same period. (Berkeley, 2023) This is unprecedented. Just because there is going to be higher demand for compute in the next 10 years, this does not automatically mean that there is more energy that will be required at the same rate of increase in demand. Processes are becoming more efficient, energy is being used more effectively, and data centers are becoming cheaper than ever before. It's almost like this is what a free market society does when implementing new technologies, they make it cheaper.

The solution to more energy demand is more energy.

We are seeing unprecedented amounts of investment being made in AI and data centers, especially from companies such as Oracle, literally going into debt building these data centers. The Department of Energy reported that the United States added "53 GW of utility-scale generating capacity in 2025." (DOE, 2026) The last time we added that much energy infrastructure to our systems was in 2002, when the Internet was taking over everything and cloud computing was beginning to become a serious part of our economy. In 2026, developers are planning to add another 86 gigawatts of utility-scale generating capacity, which would be the largest annual increase in United States history if completed. More than half of that planned capacity comes from solar, along with major additions in battery storage, wind, and natural gas. (U.S. Energy Information Administration, 2026)

This is how I believe we should be looking at the data center debate. If electricity demand is increasing because we are building more data centers, manufacturing more semiconductors, and expanding Artificial Intelligence, our first reaction should not be to stop those industries. We should be figuring out how to produce more electricity and strengthen the infrastructure around it. The Department of Energy has already stated that there is a "pressing need" for additional transmission infrastructure because of growing demand from data centers, manufacturing, and other large industrial users. (Department of Energy, 2026) To me, that sounds like an infrastructure problem that can be solved, not a reason to slow technological advancement.

Obviously, this does not mean that every data center should be approved regardless of where it is built. A massive facility proposed in an area already struggling with electricity or water availability should have to demonstrate how it plans to operate responsibly. Companies building these facilities should also help pay for the generation, transmission, substations, and other infrastructure needed to support them instead of simply placing those costs on the surrounding community.

Putting the Environmental Impact Into Perspective

This is where I believe a lot of the conversation loses perspective. Berkeley Lab estimated that electricity used by American data centers produced approximately 61 million metric tons of CO2-equivalent emissions in 2023. The Environmental Protection Agency reported approximately 6.2 billion metric tons of total greenhouse-gas emissions in the United States that same year. (Berkeley Lab, 2024) (EPA, 2025) That puts the electricity associated with data centers at roughly 1% of total American greenhouse-gas emissions. One percent is not zero, but it is far different from the idea that data centers are becoming one of the country's largest environmental threats.

Water requires the same kind of context. Berkeley Lab found that the amount of water consumed by a computing workload can vary by more than 10,000 times depending on where the data center is located, the cooling system it uses, the efficiency of its servers, the surrounding climate, and the electricity source. (Berkeley Lab, 2025) This means there is no single answer to how much water "AI uses." The better question is where these facilities are being built and how efficiently they are being designed.

Stewardship Does Not Mean Stagnation

This brings me back to where I started. Being a good steward of the Earth does not mean refusing to use the resources available to us. It means using them responsibly. Hospitals consume enormous amounts of electricity. Factories require energy and water. Farms require land and irrigation. Transportation requires fuel and infrastructure. The question has never realistically been whether human activity can have zero environmental impact. The better question is whether the value being created justifies the resources required and whether we continue improving how efficiently those resources are used.

I believe compute is quickly becoming one of the most valuable resources in the modern economy. Artificial Intelligence is already helping researchers study new medicines, analyze enormous amounts of scientific information, improve weather forecasting, develop new materials, increase productivity, and give small organizations access to capabilities that previously required much larger budgets. And we are still at the beginning.

The "cloud" is not actually floating somewhere above us. Every AI model, cloud service, search engine, streaming platform, and digital tool eventually comes back to physical computers sitting inside real buildings connected to real electrical infrastructure. Those computers are going to exist somewhere in the world, and I would rather see more of that infrastructure built here in the United States than China.

Environmental concerns should absolutely influence how we build data centers, but I do not believe they should become an argument for whether we build them at all. Build the data centers, expand the grid, invest in more generation, improve cooling systems, protect local water supplies, and continue making every generation of chips and facilities more efficient than the one before it. If Artificial Intelligence becomes even remotely as important as I believe it will be, then countries with abundant energy, advanced semiconductors, reliable infrastructure, and large amounts of computing power will have an enormous advantage. I want the United States to remain one of those countries.

Responsible stewardship should encourage us to build better and more efficiently. It should not convince us to stop building.