A close-up shot of racks of servers connected by various cables inside a data center

Digital Transformations, Environmental Degradation

Tommy Bui Nguyen

The next 30 years mark a unique period in humanity’s timeline. Actions we take between now and 2050 will decide the fate of billions of human beings. By 2050, we will know if we made the right decisions in the decades prior. We will have either set humanity on the correct trajectory to reduce greenhouse gas emissions, redistribute wealth, deliver justice, protect our ecosystems, and build prosperity for all lifeforms on our planet, or we will have caused ecological damage and immeasurable suffering for people and wildlife all around our world.

2020 was a warning sign from an undesirable future. While lockdowns caused global CO2 emissions to fall by 6.4 percent and granted a pyrrhic victory to the climate, the win was short lived. According to Jeff Tollefson in Nature, by the latter half of 2020, emissions bounced back, and as the world wakes up from the hazy dream of COVID-19, emissions are rising to levels seen in prior years.

Continuing to emit greenhouse gasses at this level is untenable. Massive droughts, wildfires, and record high temperatures are affecting all seven continents. Not so long ago, these same “natural” disasters would have been viewed as once-in-a-lifetime events, but as heat waves roast Western Europe and the United States, we are learning that global warming does not care about borders.

Every choice we make is an opportunity to begin a journey towards a different future, where the resources for humanity’s nourishment and the health of our planet are abundant. Getting there, however, will require an immense transformation of the world as we know it.

Dematerialistic Illusions

Starting in the 1970s, the Digital Revolution enabled the widespread use of computers and access to the internet. By the 1990s, the Commodore 64 and other personal computers had found their way into millions of offices and homes across industrialized nations.

That same period marked the rise of “dematerialization,” or the ability to produce more with less physical input. Computing and the digitalization of various parts of the economy made the effect seem larger than it actually was. Imagine this! Soon all media and most services will become bits on a computer, preached techno-utopians around the world. Everything will be cheap and available to anyone with a computer.

This seemingly naive viewpoint actually has roots in empirical data. In the U.S., the consumption of non-renewable resources like copper and steel had decreased as GDP continued to rise. As researchers like Andrew McAfee point out, burning coal and gas has gotten more efficient in the past decades, resulting in less emissions for the same amount of energy.

Nor is it a stretch of imagination to see how digital technologies would lead to dematerialization. Consider the advent of streaming movies and electronic books. Rather than use fossil fuels to supply the energy for the production and delivery of these services, we could now use renewable electricity to deliver them.

Must it not follow that digital services would reduce waste and use less energy?

The Jevons Paradox and the Lifecycle Analysis of Emissions

In 1865, economist William Stanley Jevons observed a phenomenon that would plague the 20th and 21st centuries. Shortly after the invention of the steam engine, a machine intended to make coal more efficient, coal consumption counterintuitively increased. Rather than doing the same with less, industry learned to do more with more.

Rather than doing the same with less, industry learned to do more with more.

The Digital Revolution suffered a similar fate. Despite optimism in digital technologies to reduce emissions and make the world more efficient, greenhouse gasses and material consumption continued to rise. Techno-utopians made two oversights when envisioning a material-free (or at least, material-minimal) future.

The first was the specter of the Jevons Paradox. Widespread computing did not create a world in which Netflix replaced AMC or Amazon replaced Barnes and Noble. Instead, it created a world where people stream more movies and go to theaters, own both personal libraries and Kindles. Echoing the same observation made by Jevons in the 1860s, increased efficiency did not lead to a decrease in net consumption — it only made more consumption possible by more people.

The second critical oversight was forgetting the many sources of emissions of digital technologies, which includes manufacturing and disposal. Computer manufacturing takes place all over the world. Oils for plastics are piped from the Middle East, lithium for batteries are mined in Chile, raw materials then are shipped to China to be processed and assembled, and the final product is flown back to the U.S. to be sold and placed in the hands of consumers.

Every step in this process generates emissions, and increasingly sophisticated manufacturing techniques for more powerful devices push emissions higher. In Patterns, Freitag et al. find that manufacturing, on average, accounts for 50 percent of the emissions generated for a personal computing device, depending on its lifespan before its disposal.

These observations conflict with the “green” vision that techno-utopians and modern economists had for digital services. Digital services can only significantly lower the emissions for a given activity if they replace the physical activity entirely. This means a world where Zoom calls replace most flights, and not a world where there are Zoom calls and flights.

Outrunning the Rebound Effect

Despite all the ways that digital devices produce emissions, their direct overall contribution to global warming is shockingly small. Freitag et al. find in the same Patterns study that in 2020, digital technologies accounted for an estimated 2.1 to 3.9 percent of global greenhouse gas emissions. This figure roughly captures the emissions generated by the manufacturing, use, and disposal of servers, personal computers, and smartphones.

This emissions figure is contrasted by digital technologies’ social and economic influence. For example, Apple, Microsoft, Google, and Amazon have a combined market cap of over $5 trillion. Big tech companies facilitate a vast amount of economic and social activity, yet surprisingly, their emissions remain relatively minor.

One explanation lies in the extraordinary energy efficiency gains seen in data centers over the last decade. According to the International Telecommunication Union, between the years 2010 to 2018, the number of individuals connected to the internet nearly doubled. This has led to data center workloads increasing sixfold, internet traffic increasing tenfold, and data center storage capacity increasing 25-fold. Yet, analysis published in Science by Masanet et al. showed that the electricity usage of data centers increased by only 6 percent.

At least for data centers, it seems that a true decoupling between material inputs and useful outputs has manifested. Research and engineering continue to improve compute, cooling, and energy efficiencies, with the ultimate effect of providing more with less. These efficiency gains can hold out for a few more years, as millions more become connected to the internet.

Another explanation for digital technologies’ small emissions footprint is the industry’s heavy investments in renewables. According to Google’s 2022 Environmental Report, since 2017, the company has purchased enough renewable electricity sources to match the amount of electricity it uses during its operations. Similarly, several big tech companies have banded together to invest $925 million in carbon capture technologies, the largest investment of its kind to date. Depending on how carbon accounting is done, these investments and purchases can push the emissions generated by tech companies downward.

Digital industries have managed to narrowly avoid the Jevons Paradox in the 21st century. By combining R&D, carbon offset purchases, and renewable investments, tech companies can continue to grow, increasing the amount of computing devices and services provided in the world, while reining in their emissions…

… Or not. What this analysis lacks is a hard look at how these technologies impact external industries. How can we think about the emissions that are enabled by digital technologies, but not directly caused by them?

Technology is Not (Carbon) Neutral

In May 2020, Greenpeace published a report documenting the ways in which cloud computing service providers such as Microsoft, Amazon, and Google have aided fossil fuel companies. The use cases vary from making the transportation of oil more efficient to building out machine learning models based on seismic data to find new oil sources. The consequences were all the same — oil was easier to pull out of the ground.

This is not the only example of how tech influences industry and consumer behavior in harmful ways for the environment. For example, booking flights became dramatically easier due to software improvements, causing a subsequent increase in emissions generated by the aviation industry.

We cannot entirely scapegoat software for the aviation industry’s emissions — airlines and airplane manufacturers must ultimately reduce flights or find ways to make flights more environmentally friendly. But it seems naive to ignore the role that digital technologies play in enabling these emissions.

Our understanding of how technology companies have shifted consumer and industry behavior is still quite small relative to its true scope and scale of transformation. Only recently has some of the data been made public to understand how technologies like social media and smartphones have impacted human behavior. Similarly, researchers and politicians have only recently begun holding technology companies accountable for the harms they have caused to our social and political systems.

Just as we attempt to hold tech companies accountable for social and political harm, we must also hold them accountable for the ecological degradation they enable. We cannot let companies like Google make promises to reach net-zero emissions by 2030 while still holding contracts with fossil fuel companies. We must reconcile the improvements that digital technologies bring with the equally real destruction that they can cause.

Beyond Digital Greenwashing

Our current economic and political systems incentivize the production and use of fossil fuels. Cheap, abundant, easily transportable energy has been the lifeblood of the 20th and 21st century. Governments dole out land permits to drill for oil, provide subsidies for natural gas, and are hesitant to buck the status quo. How can we guide industry leaders and politicians towards investing into renewables, research, and resilient infrastructure?

Part of the answer lies in digital technologies. While the power of computation can supercharge resource extraction, the reverse is also true. Computation can help preserve the Earth and reduce resource consumption. For example, satellite imagery combined with computer vision has helped identify illegal fishing operations harming ecosystems. Machine learning models can increase the efficiency of heating and cooling data centers. The key here is to align technology with missions and goals that seek conservation and sustainability, instead of narrow-minded growth.

We can also borrow from Silicon Valley’s playbook. Venture capital firms and Big Tech companies invest heavily in sustainability startups, many of which are attempting to scale experimental technology. As with all startups, many are expected to fail. But only a handful of them need to succeed to have a meaningful impact on emissions.

These investments signal that there is a growing appetite for combating climate change. In a story from The Atlantic, Nan Ransohoff, the Head of Climate at Stripe, says, “We are trying to buy ourselves time to get the right policy mechanisms in place to take this market where it needs to go.” Maybe, just maybe, tech companies and VC firms leading the way in these renewable investments and experimental technologies can give governments and politicians the push they need to follow suit.

The appetite for sustainability work is large and growing in the tech space. Employees and organizations see the value in rapidly reaching net-zero emissions and finding ways to accelerate net-zero efforts across other industries. The next step, with which tech is ever-so-familiar, is scaling up these projects to deal with the grand scope of emissions.

Beyond Techno Fixes

The emissions figure for the digital tech industry paints a picture that is both hopeful and grim. Despite the industry’s massive influence on commerce and society, it has managed to keep its direct emissions relatively small. Freitag et al. estimates annual emissions to be around 2.3 gigatons of CO2 and the IPCC WG3 report estimates that solar and wind energy alone could reduce emissions by two to four gigatons of CO2 by 2030, as the industry strives towards net-zero emissions. Given that tech companies are at the forefront of investing in renewable energy and experimental carbon capture technologies, I am hopeful for a net-zero world for our digital infrastructure.

At the same time, this means that even if the tech industry solves the issue of emissions internally, those changes will be nowhere near enough to prevent a climate catastrophe. The other 96 percent of greenhouse gas emissions comes directly from industries such as agriculture, construction, and manufacturing. And as much as digital technologies may make these industries more efficient, it is unlikely to be enough to completely draw emissions down to net-zero.

What we need is a Sustainability Revolution — a revolution that changes the core values in our political, economic, and cultural systems to protect the shared wealth of our natural commons.

The Exponential Roadmap Initiative, an organization supported by European politicians, CEOs and the United Nations, calls for using “exponential technologies [to] help dramatically reduce energy, material, and waste.” But we know that this cannot be the only change. The past few decades alone show us that efficiency gains typically lead to more net consumption, not less.

What we need is a Sustainability Revolution — a revolution that changes the core values in our political, economic, and cultural systems to protect the shared wealth of our natural commons.

Writers and activists like Naomi Klein and George Monbiot have spoken at length at what this new system may look like. It could be a world where our energy grid is solar-powered, distributed, and cooperatively owned, where open-sourced software and hardware help manage the system’s complexity without centralizing power and wealth. A world filled with public luxury, in the form of gorgeous parks, theaters, and infrastructure, and where everyone would be satisfied with private sufficiency.1 Whatever form this new system takes, at its core it will value sufficiency, longevity, and sustainability.

We Build the Future

Digital technologies have given the gifts of connection and education to humanity, while requiring relatively few material inputs. But just as a teenager must mature, so too must the tech industry grow wiser and more responsible for its actions. It must consider not only the impact it has now, but the impact it will have for generations to come. No longer can they avoid accountability under the guise of trying to create radical abundance in an always almost-here future.

Digital technologies have inherited the practices of previous industrial revolutions: exploit labor, concentrate power, and ignore externalities. Centuries of experience have taught us that this is not a sustainable mode of operation. Even in the digital realm, we cannot escape the fact that infinite growth on a finite planet is impossible.

Though the window to change our emissions trajectory is rapidly closing, it remains open. We must build a future that acknowledges planetary limits. It is time for us to seize this moment.

1 George Monbiot gave a talk in October 2020 where he coined the phrase “private sufficiency, public luxury.” He describes how unjust and unequal property rights has created a world that protects the ultra-wealthy and frames poverty as a personal failing. Monbiot goes on to argue that individuals embedded in Western cultures need to stop seeing themselves as “temporarily embarrassed millionaires,” and instead fight for systems that create a floor for quality of life that no one should fall below.

Tommy Bui Nguyen (he/him) is a software engineer based in Chicago, IL. When he's not at his desk writing code, you'll find him outside running, bouldering, or enjoying a cup of coffee and a book.