Staring at that red dial while your Netflix feed refreshes, do you get a kind of queasy feeling? As Spotify floods your audioscape all the waking day, does that algorithmic vapor make you wheeze? Thrilled to access illegal video documents of the 2012 Egyptian uprising at https://858.ma, do you imagine the route they traveled, via satellites, undersea cables, and data hubs, to reach you? Immersed in a Minecraft livestream with your new besties in Seoul and Johannesburg, do spectral fingers tickle the nape of your neck?
Those moments of unease are the reckonings of a fact nobody wants to think about: streaming media has a massive carbon footprint. I'm here to tell you that the energy consumption and CO2 emissions that result from streaming media constitute a growing threat to the planet. Streaming video is now responsible for over 1 percent of greenhouse gas emissions, and that number is increasing rapidly. For example, the 1.7 billion streams of “Gangnam Style” in its first year consumed 298 gigawatt hours of energy, more than the annual electrical consumption of the entire population of Burundi.1 In a worst-case scenario, by 2030 communications technology could consume 51 percent of global electricity, and produce 23 percent of greenhouse gas emissions.2 Another report predicts that by 2024, video will comprise 74 percent of data traffic; according to another, 89 percent in 2030.3 Those of us in the professions of making and studying media have a responsibility to confront this issue, understand its complex parameters, raise awareness, and lead a movement of conscientious, small-footprint media practices.
Scholars are examining environmental effects at every stage of production, exhibition, distribution, consumption, archiving, and disposal—and arguing for ecological true-cost accounting.4 They shred the ideology that digital media are immaterial.5 They ground media ecology in non-anthropocentric worldviews.6 Thanks to these scholars of media ecology, materiality, big data, and infrastructure, as well as to activist organizations like Greenpeace and the Shift Project, we are gaining a solid understanding of the energy consumption, water usage, toxicity, and waste of media that belie corporate myths that digital media are immaterial and the “cloud” is made of fluffy vapor. However, thus far scholars have not addressed the environmental impacts of streaming media in detail, even though streaming has transformed the economic landscape of media consumption in the past few years. Let's deal with it!
I think part of why cinema and media scholars have paid little attention to this issue is that a finding from 2014, based on 2011 data suggesting that streaming a movie produces fewer CO2 emissions than watching a rented DVD,7 continues to circulate. In that finding, often cited in engineering journals, what accounted for the disparity between DVDs and streaming was almost entirely the emissions resulting from driving to the video store and the energy used by DVD players and monitors; the energy required to produce and ship DVDs counted for less than 1 percent of total energy use. (Work is needed to compare the environmental impact of plastic disposal with that of producing more new electronic devices.) Crucially, the 2014 study also assumed that consumers watch about five movies per month. This assumption has become laughably inaccurate, since now many media consumers stream for more than five hours per day.8
In some cases, such as videoconferencing instead of flying to a conference, it is true that streaming consumes less energy than a few of the activities it has come to subsume. But media streaming has also given rise to new consumption habits, new addictions that media corporations are happy to feed, and new flavors to the bait of “immersive,” high-definition moving images that require a lot of bandwidth. The 2014 study comparing DVD use and streaming observed that CO2 emissions would rise with higher data rates. As mentioned above, mobile video traffic currently comprises more than half of mobile data traffic and is rising at an alarming rate. Meanwhile, “demand” for higher image quality has increased the energy consumption of streaming video.
To get facts about the energy use of streaming video you need to know some engineering and statistics, to imaginatively extrapolate between best- and worst-case scenarios, and to interpret the findings of reports in light of the vested interests they represent. In preliminary research I have focused on reports aimed at nonspecialists, scanning the articles by engineers to which they refer, and interpreted them critically according to the interests they represent. Some of these reports are produced by academic researchers, some by researchers for media corporations, and one by an energy agency. The most pessimistic conclusions I've found are Greenpeace's 2017 Clicking Clean report, the Shift Project's 2019 report, and a widely cited 2015 report by Anders Andrae and Tomas Edler of Huawei Sweden, authors of the above-cited worst-case scenarios.9 A more moderate report, optimistic about increased efficiency and corporate self-regulation, is “Sustainable Energy Development” (2017) from the International Energy Agency (IEA), an agency with power to influence government and corporate decision making. This report argues that IT energy consumption can stay at the current rate if we switch to renewable energy, increase network efficiency, and improve energy efficiency for the billions of connected devices. That's three big ifs. The IEA was founded during the oil crisis of 1974 to guarantee reliable access to energy for its twenty-nine wealthy member countries. Its objectives now include “sustainable energy policies that spur economic growth and environmental protection in a global context”10—a pair of goals that do not easily coexist, as reflected in the membership of oil and coal corporations among the agency's board. Oil Change International and Greenpeace UK, in a warning to energy investors, state that IEA's report is aligned with 2 percent global warming and does not meet the standards of the Paris accords.11 There are three huge hurdles to minimizing this footprint: renewable energy, greater efficiency, and reducing demand.
Convincing our governments to abandon the exploitation of fossil fuels, develop renewable energy, and regulate corporate and industry energy usage are the most important goals to slow or prevent catastrophic climate change. It's important to distinguish between energy consumption and CO2emission, a distinction surrounded by a haze of question marks. The energy footprint depends in part on the number of points in the network, whether the streaming corporation, like Netflix, economizes by building local caches,12 and the source of electricity used at each one. Streaming media can have a low impact if the energy source at all points is renewable. Does the local electricity grid use hydro or coal? What about the electricity grid at the hub? A lot of data hubs for both North American and overseas destinations are based in Virginia,13 where coal is still the major energy source. China and India rely on coal. South Korea is one of the most wired countries in the world, and its largest energy source is coal.14
And I'm sorry to say that even if the majority of the planet's energy sources do eventually become renewable, it's questionable that these would supply sufficient electricity to sustain current levels of consumption. Thus it is important to increase efficiency of streaming media and to reduce demand for it.
In general, small devices are more energy efficient than large ones, though again, this is partly because the energy consumption has been outsourced to the server. At the same time that efficiency increases, media corporations are issuing more energy-intensive products using very high resolution, including 4K and 8K (4,000- and 8,000-pixel) video and multiuser platforms for games, all of which demand extremely high bandwidth.
The optimistic IEA report says that in the best-case scenario, increased efficiencies will cancel out increased energy use: streamlined computing, better cooling, switching to renewables, using smartphones rather than larger devices. But given all the variables, this scenario looks unlikely. One variable is where in the world we are talking about. Most studies focus on wealthy parts of the world where infrastructure and consumer platforms are updated fairly regularly, for more efficiency (good) as well as more bandwidth (bad). But in poorly infrastructured parts of the world, governments may not be able to afford to convert to renewable energy (bad); people may be using older, energy-inefficient devices for longer (good and bad); and most people will not have access to energy-intensive broadband (bad for media democracy, possibly, but good for the environment).
Currently most streaming media is third-generation, or 3G, technology. Mobile media consume more energy than fixed-line networks, but their efficiency has been increasing with each generation so far: 2G uses one hundred times as much electricity as fixed-line networks, 3G ten times, and 4G four times as much.15 However, as we know, their faster speeds mean people are doing a lot more media streaming. Controversy surrounds the emergent 5G or fifth-generation technology. 5G is thought to be more energy efficient, but this efficiency depends on a great many variables and looks unlikely to outweigh increased demand.16 Media corporations support the move toward energy efficiency, but these efficiencies are outweighed by corporations' determination to grow their markets. With the exception of Netherlands-based Fairphone, no media corporation is interested in making long-lasting, non-obsolescent devices.
Speaking of 5G, this controversial new platform, which requires the installation of many small towers, on top of the extant large cell-phone towers, will flood the ether with unprecedented levels of electromagnetic frequencies (EMF). The adverse health and environmental effects of higher levels of EMF are well documented, but telecommunications companies and governments continue to ignore them.17
STREAMLINING DEMAND—OR DEMANDING STREAMLINING
Therefore, we need to focus on the third hurdle: reducing demand. This one might be the most obstreperous, but it is also the one that we scholars and makers of media can best influence. The only person I've seen promulgating this view is UK engineer and data-center consultant Ian Bitterlin, whose provocative observation that banning high-definition color cameras on phones could reduce data traffic in Europe by 40 percent has made a mark in popular media.18 I want to emphasize that it is the new energy-intensive media products—swathed in myths of media immateriality and progress toward verisimilitude—that give rise to demand, not the other way around. This is a classic observation about marketing and the creation of needs. It's important to argue that consumers don't “naturally” demand higher resolution if the solution I propose is to be feasible.
First, I advocate slow media. Walking to the library to borrow a movie or game on USB, for playback on an energy-efficient device, would likely be the most energy-efficient way to consume media. Elegant, slim, and slow, this is a consumption mode I would love to make fashionable. But it would not meet the desires for immediacy, intimacy, and interactivity that streaming satisfies.
The idea that we must limit energy-intensive streaming media goes against the notion that unlimited internet access is a right. Streaming media, a commodity that didn't exist twenty years ago, is now deeply imbricated in the habits of everyone who has a smart phone—which is a fast-increasing majority of the world's population. The number of devices in use (as opposed to in landfills) in 2030 is anticipated to reach 1.4 per person, or a total of more than ten billion devices, for the projected population of 7.61 billion.19 We know the majority of what people stream is light entertainment—porn, games, viral videos. I expect it's the same for people in wealthy and poor parts of the world. As Lawrence Liang defends media piracy that is frivolous as well as oppositional,20 similarly I don't want to say people should only stream serious content. It would be great if consumers could just describe viral videos to their friends, instead of forwarding them! But if people are not going to stop streaming media to their wireless devices, we need to make low-impact media attractive. Low-definition, small-file media. I'm thinking ASCII porn.
It is time to sing the praises of small-file aesthetics. In contrast to corporate ideologies of high resolution, lossless compression, and “immersion,” most media images actually circulating are low resolution, glitchy, and intermittent.21 They have an audiovisual “aesthetics of poverty” that emphasizes the physical affordances of digital media and the ingenuity of their makers and receivers. Low-tech aesthetics are embraced in memes, cell phone text works, and the retro Vaporwave genre. Computer animation can be both beautiful and energy efficient. Compression and datamoshing have their own piquant aesthetics. A lot of the performance media people love, like Snapchat and TikTok, don't suffer from being low resolution. Also, given that video playback through game consoles with internet access accounts for as much as 50 percent of all streaming video,22 it's a good time to cultivate game nostalgia too. Strip Pong, anyone?
Much of this creativity comes from the economic fringes. Artists, makers, hackers, pirates, and consumers in regions passed over by 4G and 5G are well versed in tinkering, retrofitting, and making do with supposedly outdated media technologies.23
In short, we have abundant aesthetic models for media works that celebrate low technology, slow speed, and small files. I am particularly excited about ASCII movies. You can build them from scratch, or you can use a simple program by the angelic Akhilesh Sharma to convert video frames to frames of ASCII text.24 Low-res, small-file media are sexy, literally, because unlike high-resolution media that, in Marshall McLuhan's term, “push” out at the viewer, these crunchy and approximate movies can draw you in.
In August 2020, with friends in Vancouver, I will launch the Small File Media Festival that will stream works of extremely limited file size, to show that being green can be elegant, gritty, and wildly attractive.25 In addition, colleagues Joseph Clark, Adel Iskandar, and I did a teach-in on the topic for Global Climate Strike week in September 2019, and in April 2020, with Clark, Lucas Hilderbrand, Jason Livingston, and Denise Oleksijczuk, I held the online roundtable “Let's Deal with the Carbon Footprint of Streaming Media,” scheduled to be presented at this year's Society for Cinema and Media Studies conference. And with colleagues in media studies and engineering at my university I have embarked on a year-long project to translate IT engineering research on this topic into terms accessible to the public. You are welcome to find those materials on my web site and adapt them for your own use.26
Mission Critical Power Team, “Let them eat bandwidth: Ian Bitterlin on optimisation or bust,” Mission Critical UK, November 20, 2019, https://missioncriticalpower.uk/the-drivers-of-power.
Anders S. G. Andrae and Tomas Edler, “On Global Electricity Usage of Communication Technology: Trends to 2030,” Challenges 6, no. 1 (2015): 117–57.
Mohammed H. Alsharif, Anabi Hilary Kelechi, Jeong Kim, and Jin Hong Kim, “Energy Efficiency and Coverage Trade-Off in 5G for Eco-Friendly and Sustainable Cellular Networks,” Symmetry 11, no. 3 (2019): 408; and Andrae and Edler, “On Global Electricity Usage”: 142.
For example, Sean Cubitt, Finite Media: Environmental Implications of Digital Technologies (Durham, NC: Duke University Press, 2017); and Hunter Vaughan, Hollywood's Dirtiest Secret: The Hidden Environmental Costs of the Movies (New York: Columbia University Press, 2019).
For example, Lisa Parks and Nicole Starosielski, eds., Signal Traffic: Critical Studies of Media Infrastructures (Champlain, IL: University of Illinois Press, 2015).
For example, Adrian Ivakhiv, Shadowing the Anthropocene: Eco-Realism for Turbulent Times (Goleta, CA: Punctum Books, 2018); and Jennifer Fay, Inhospitable World: Cinema in the Time of the Anthropocene (Oxford, UK: Oxford University Press, 2018).
Arman Shehabi, Ben Walker, and Eric Masanet, “The energy and greenhouse-gas implications of internet video streaming in the United States,” Environmental Research Letters 9, no. 5 (2014).
Cisco, Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2017–2022 White Paper (February 2019), https://s3.amazonaws.com/media.mediapost.com/uploads/CiscoForecast.pdf; Jens Malmoden and Dag Lunden, “The Energy and Carbon Footprint of the Global ICT and E&M Sectors 2010–2015,” Sustainability 10 (2018): 3027–58; and Andrae and Edler, “On Global Electricity Usage.”
Greenpeace, “Clicking Clean: Who Is Winning the Race to Build a Green Internet?” (2017), www.greenpeace.org/international/publication/6826/clicking-clean-2017; and the Shift Project, “Climate Crisis: The Unsustainable Use of Online Video” (2019), https://theshiftproject.org/en/article/unsustainable-use-online-video; Andrae and Edler, “On Global Electricity Usage.”
Oil Change International and Greenpeace UK, “The International Energy Agency and the Paris Goals: Q&A for Investors,” January 2019, priceofoil.org/content/uploads/2019/02/IEA-and-the-Paris-Goals-Q-A.pdf.
Ramon Lobato, Netflix Nations: The Geography of Digital Distribution (New York: New York University Press, 2019).
Lobato, Netflix Nations.
Greenpeace, “Clicking Clean.”
International Energy Agency, “Sustainable Energy Development.” Calculation based on Andrae and Edler, “On Global Electricity Usage”; Cisco, Cisco Visual Networking Index; and other sources.
For example, Alsharif, Kelechi, Kim, and Kim, “Energy Efficiency and Coverage Trade-Off in 5G”; and Hanna Pihkola, Mikko Hongisto, Olli Apilo and Mika Lasanen, “Evaluating the Energy Consumption of Mobile Data Transfer—From Technology Development to Consumer Behaviour and Life Cycle Thinking,” Sustainability 2018, 10, 2494–2510. The IEA report is silent on 5G.
Agostino Di Ciaula, “Towards 5G communication systems: Are there health implications?,” International Journal of Hygiene and Environmental Health 221, no. 3 (2018): 367–75; and Igor Yakymenko, et al. “Oxidative mechanisms of biological activity of low-intensity radiofrequency radiation,” Electromagnetic Biology and Medicine 35/2 (2015): 1–16. I thank Denise Oleksijczuk for these references.
Nicola Jones, “How to stop data centers from gobbling up the world's energy,” Nature 561 (September 12, 2018), www.nature.com/articles/d41586-018-06610-y. See also Bob Found, “Alarmed over climate change? Then get off the internet,” Halifax Chronicle-Herald, October 12, 2019, www.thechronicleherald.ca/opinion/local-perspectives/bob-found-want-to-curb-climate-change-then-get-off-the-internet-362077; and Mission Critical Power Team, “Let them eat bandwidth.”
Andrae and Edler, “On Global Electricity Usage.”
Lawrence Liang, “Beyond representation: The figure of the pirate,” in Making and Unmaking Intellectual Property: Creative Production in Legal and Cultural Perspective, ed. Mario Biagioli, Peter Jaszi, and Martha Woodmansee (Chicago: University of Chicago Press, 2011), 353–75.
Olga Goriunova, Art Platforms and Cultural Production on the Internet (New York: Routledge, 2012); Hito Steyerl, “In Defense of the Poor Image,” The Wretched of the Screen (Berlin: Sternberg Press, 2012), 31–45; Paula Cardoso Pereira and Joaquín Zerené Harcha, “Revolutions of Resolution: About the Fluxes of Poor Images in Visual Capitalism,” TripleC 12, no. 1 (2014): 315–27; and Laura U. Marks, “Arab Glitch,” in Uncommon Grounds: New Media and Critical Practices in North Africa and the Middle East, ed. Anthony Downey (New York: I.B. Tauris, 2014).
Shehabi, Walker, and Masanet, “The energy and greenhouse-gas implications.”