Artificial intelligence is transforming how we work, learn, communicate, and solve problems. From accelerating medical research to improving weather forecasting and optimizing business operations, AI has the potential to become one of the most impactful technologies of our lifetime.
But every technological revolution comes with tradeoffs.
Behind every AI generated answer, image, recommendation, and prediction is an increasingly expanding network of data centers operating around the clock. According to the International Energy Agency, electricity demand from data centers is expected to more than double by 2030, driven largely by AI workloads.
How that electricity is generated will determine not only climate outcomes, but also the quality of the air people breathe. The challenge is not whether AI should continue to advance. It is how we ensure the infrastructure powering it grows in a way that protects public health, the environment, and our air quality.
At IQAir, we are at the intersection of AI technology and air quality. We leverage AI across our business while remaining conscious of our environmental footprint. That's why we're focused on creating products that consume less energy, are built with more sustainable materials and processes, and are designed to last for well over a decade.
The Missing Metric in AI Sustainability: Air Quality
AI infrastructure is increasingly an air pollution issue, not just an energy issue.
In many regions, rising electricity demand is still met by fossil fuels, it’s how around 80% of the world’s total energy is produced. That means additional emissions of nitrogen oxides and fine particulate matter (PM2.5), pollutants linked to asthma, heart disease, respiratory illness, and premature death.
PM2.5 is responsible for nearly 90% of the health impacts associated with air pollution. In other words, the smallest particles create the largest consequences.
This is where AI infrastructure becomes tangible. At one of the world’s largest data center hubs in Northern Virginia, researchers estimate that air pollution linked to a proposed AI-related data center could generate $53 million to $99 million annually in health damages, including premature deaths and respiratory illness.
This is the blind spot in most AI sustainability discussions. Emissions are not only a global carbon problem. They are also a local air quality problem.
Environmental Justice Is Becoming an AI Issue
The geography of infrastructure matters.
Harvard research shows that proposed AI-related energy projects with the highest air pollution impacts are often located in communities with higher social vulnerability, lower incomes, and higher baseline rates of respiratory illness (1).
This is not a new pattern. Fossil fuel power plants in the United States have historically been 31% more likely to be located near and upwind of historically redlined neighborhoods.
The concern is not that AI is uniquely harmful. It is that AI-driven energy demand may reinforce existing infrastructure patterns where communities with the least political and economic power absorb a disproportionate share of environmental burden.
If AI is reshaping society, its physical footprint cannot replicate past and current inequities.
A Different Way to Think About AI Infrastructure
The conventional view treats data centers as digital infrastructure. A more accurate view is that they are air quality infrastructure by proxy. This framing matters because it changes the design problem.
The question is not simply how to make data centers more efficient. It is how to ensure that AI growth does not unintentionally increase pollution exposure in already burdened communities.
That requires moving beyond carbon-only metrics and incorporating air quality and health outcomes into infrastructure planning.
What Responsible AI Growth Actually Requires
The path forward is not about slowing AI but aligning strategic investment with social impact. It is about building AI infrastructure differently: with public health and air quality in mind.
That includes:
- Requiring AI growth with clean energy deployment, not just energy procurement
- Prioritizing AI data centers powered and funded by renewable energy and accelerating the transition away from fossil fuel-dependent grids
- Designing data center operations that reduce peak grid stress, when fossil fuel use is often highest
- Increasing transparency around electricity source, emissions, and local air quality impacts
- Offsetting air pollution caused by AI data centers and investing resources in communities impacted by data centers
- Designing more energy-efficient AI supercomputers and adopting industry-wide energy performance standards that make efficiency as important as computing power
It also means recognizing that efficiency alone is not enough if total demand continues to rise without planning for where the energy comes from.
The Impact Most People Aren’t Thinking About
AI is often framed as a productivity story. It is also an air quality story, and increasingly, a climate and public health story.
But it is not a story of tradeoffs between technology and sustainability. It is a design question about whether we build systems that externalize environmental costs or internalize them from the start.
The most important shift is conceptual: AI infrastructure is not just digital infrastructure. We need to start treating AI infrastructure as air quality and environmental infrastructure with real-world consequences.
The opportunity now is to ensure the intelligence we are building does not come at the expense of the air people breathe, but instead helps create a cleaner, more equitable energy future.
If we acknowledge this early, we still have time to shape our future.
