In our rapidly evolving digital world, the rise of generative artificial intelligence (genAI) has revolutionized technology, unleashing new possibilities for creativity, productivity, and problem-solving. Amidst the marvels of generative AI lies an often-overlooked concern: the environmental impact of the technology that powers it. While many have become accustomed to the convenience and efficiency of cloud-based services, we must look beyond the intangible facade of the "Cloud" to understand the consequences of this often ignored modern reality.
Contrary to its ethereal namesake, the Cloud is anything but immaterial. It is a complex and immense network of hardware and infrastructure that enables the seamless transfer and storage of vast amounts of data. To understand the environmental impact of genAI, we must begin by unraveling the material threads of the Cloud – the physical data centers around the world that genAI draws information from.
The data centers powering the Cloud are voracious energy consumers. According to a Forbes report back in 2017, data centers in the United States consumed more than 90 billion kilowatt-hours of electricity per year. Expand the scope to a global level, and the total energy demand is 416 terawatts – or roughly 3% of the world’s electricity. It is estimated that today, the technology sector now accounts for approximately 20% of the world's electricity consumption, with data centers serving as a significant portion of this demand. Data centers often rely on non-renewable sources of energy, straining power grids and contributing to carbon emissions. It is estimated that a single data center can consume the equivalent electricity of 50,000 homes, making it a major carbon emitter.
It is estimated that today, the technology sector now accounts for approximately 20% of the world's electricity consumption, with data centers serving as a significant portion of this demand.
Heat is the byproduct of computation, and data centers require immense cooling capacities to maintain optimal performance. Water has become the preferred cooling agent over air, leading to substantial water usage in regions where data centers are concentrated. While some may tout innovative cooling strategies such as moving data centers to colder regions like Scandinavia, the majority of data centers still require significant amounts of energy and water to operate – things that were not top of mind when communities in areas such as Virginia and Utah welcomed data centers in during times of struggling economic conditions.
Unfortunately, a battle over resources is emerging at the local level, as a warming climate and booming data center demands strain resource allocation. A breakthrough MIT exploration points to Utah as one example where the data center, resource, social dots are connecting, “In Bluffdale, Utah, residents are suffering from water shortages and power outages, as a result of the nearby Utah Data Center, a facility of the U.S. National Security Agency (NSA) that guzzles seven million gallons of water daily to operate.” NBC has also reported on similar tensions, as communities in Arizona experience the effects of the environmental impact of data centers, have begun protested their development, noting that the long term environmental costs do not outweigh the short term financial benefits of welcoming big tech in.
As ‘post-pandemic life’ takes root, society’s dependence on data centers has come into full view, as a myriad of technologies and industries become more reliant on the cloud. Where remote work was, just a year ago, reliant primarily on video conferencing and white-boards, it now sees genAi and cloud storage built into tools like Zoom. In media, while Millennials and GenZers may note that they ‘cut the cord’ ages ago, it was just last year that cloud-reliant streaming services officially eclipsed cable. And in the tech world itself (expanding and far reaching as it is), genAi is just starting to move beyond the training and beta stages, with companies like Microsoft poised to monetize the technology by including its Copilot into Microsoft 365 – and others are poised to do the same.
Across almost all facets of life, the cloud is central in enabling our daily activities. The explosive growth in data storage infrastructure over the next decade could triple energy usage, exacerbating environmental concerns further – buoyed by the explosion of Ai.
According to a recent Wall Street Journal article, “Spending in the global AI infrastructure market—including data centers, as well as networks and other hardware that supports the use of AI applications—is expected to reach $422.55 billion by 2029, growing at a compound annual rate of 44% over the next six years".
Cloud leaders such as Google and Microsoft have touted studies that the cloud enables computing possibilities with much smaller ecological footprint than if the same compute were happening on local devices, however, an understanding of Jevons Paradox unravels this suggestion. Technological efficiencies and capacity gains achieved through cloud computing do not necessarily translate to reduced environmental impact. Instead, these efficiencies tend to fuel increased usage and expand possibilities, thus perpetuating resource consumption in a vicious cycle. The rapid growth of genAI places additional strain on the cloud and on data centers – requiring additional energy, water, and resources.
Like all complex challenges, addressing the environmental and social challenges of genAI and the cloud will require understanding, collaboration, and commitment by an ecosystem of stakeholders – each possessing strengths, areas of growth, and responsibility.
The private sector, while (on the one hand) at the root of the issue, is (on the other) also innovative, impactful, and influential. Several tech giants have pledged to achieve carbon-neutral data centers in the near future. However, self-regulation by the private sector has rarely gone far enough in ensuring the safety of citizens or the environment – recent climate pledge shortcomings perhaps the most recent example.
The private sector needs to be realistic and honest about its impact, be bold in its commitments, and look beyond the quarter - to - quarter profit prioritization as a means of adding value to its constituents and communities.
The public sector, while (on the one hand) riddled with lobbyists, re-election cycles, and a lack of understanding on modern technology, is (on the other hand), cautious, process-focused, and a necessary counterbalance to the motives that drive the private sector.
The government needs to educate itself on the breadth, complexity, and impact of big tech, be firm and consistent with climate regulation, and evolve its systems, functions, and analysis. Without these actions, it will remain impossible for the government to move at a rate that keeps pace with the private sector and includes stakeholders from across industry to ideate solutions. The White Houses’ recent reinvestment in the Inflation Reduction Act, which includes climate legislation that improves the accessibility of clean energy is a step in the right direction.
As citizens, we can (at times) under-appreciate our ability to drive change, be busy with everyday life, and be guilty of ignoring the inputs needed to create luxuries as mundane as Netflix. However, we have the power of the wallet, are creative and nimble, and unfettered by complex organizational structures. We must also educate ourselves and challenge our peers to engage, we must hold our representatives and companies accountable with our wallets, votes, and voices, and we must do the small, consistent things that add up the change – even deleting your old emails and photos can have an impact.
Like all complex challenges, addressing the environmental and social challenges of genAI and the cloud will require understanding, collaboration, and commitment by an ecosystem of stakeholders – each possessing strengths, areas of growth, and responsibility.
