Head in the Clouds, Coal in the Ground: Who Pays When Big Tech Outsources Its Carbon?

Whilst the clock shows midnight, the light from your phone catches your face, prompting an AI to check messages and letting Netflix fill the background. Despite the clean impression, this familiar scene correlates to the infrastructure behind every stored file that accounts for up to 85% of the digital sector’s environmental impact (Schien et al., 2013). It rises still to date, yet the digital economy often remains excluded from carbon accounting. Thus, who bears the cost of the cloud’s rising weight, and why does the trade regime ensure they remain invisible?

Digital services, often comprising cloud-based infrastructure, require data centres that deplete more energy globally than some nation-states at 200 terawatt hours, accounting for 0.3% of overall carbon emissions (Burrington, 2015). Big Tech companies accelerate this infrastructure demand whilst escaping its consequences by claiming low-carbon status due to their ‘virtual-classified’ business, just as Amazon’s overclaim on reaching its climate goals earlier (Penn & Tan, 2024). Vrikki’s (2024) exposes their actions as ‘carbonwashing’, exhibiting that their own sustainability reports understate actual emissions by approximately 662%. The physical energy burden is built into supply chains unowned and uninhabited by Big Tech, as seen in the hidden environmental costs borne by the Congolese as a result of exploitative cobalt mining.

Nonetheless, the cloud is clean only for those who never have to see where it comes from. Emissions from Indonesian electricity generation rose from 199.91 million metric tonnes in 2020 to 239.26 million metric tonnes in 2023, a nearly 20% increase in just three years (Fernandéz, 2025). This number reflects the realities of Central Sulawesi and North Maluku, where the air is thick with sulphur dioxide from captive coal plants built to power nickel smelters servicing the same digital economy that markets itself as clean and carbon-free. CRI (2025) tells this story by presenting how indigenous communities live in ‘sacrifice zones’ as a downstream consequence of nickel extraction, bearing the health and livelihood costs of an industry they did not build, whilst the state, domestic firms, alongside Chinese and other multinational corporations remain largely unaccountable.

Somewhere in the haze between a server farm in Silicon Valley and a fishing village watching coal barges pass, Diprose et al.’s (2022) argument becomes irrefutable: the same tin that decarbonises electronics in the Global North recarbonises the communities that extract it. This is not accidental. Khan (2023) demonstrates that both the General Agreement on Trade in Services and the Information Technology Agreement lack the provisions necessary to regulate data-driven services, and that despite ongoing deliberations, bureaucratic limitations persist––a gap that the EU Carbon Border Adjustment Mechanism reveals as it generates WTO tension between developing nations that view it as a protectionist barrier (ZCA, 2024) and the European Parliament (2025), which asserts it as a carbon leakage prevention tool.

The WTO moratorium on customs duties on electronic transmissions, then, illustrates how the system enables uneven distribution of risks and gains from the digital economy by denying developing states $48 billion in tariff revenue and least developed countries a further $8 billion in fiscal space (Banga, 2022). This is reflected in objections raised at MC13, where India, South Africa, and Indonesia contested the erosion of their ‘fiscal autonomy’ and ‘tariff policy space’ (Rathi, 2026). Since e-commerce provisions across Preferential Trade Agreements vary in scope and enforceability, they produce incoherence rather than accountability and offer no reliable alternative (Khan, 2023). Unified multilateral frameworks that bind digital trade to carbon and fiscal accountability are the necessary long-term corrective (Khan, 2023), but they demand something the current order has not yet produced: the political will to rewrite regulations that work very well for those who wrote them.

Ultimately, international trade structures tax physical commodities but leave digital data flows largely untaxed and unmonitored. This creates a structural loophole where Northern tech firms claim low-carbon status, whilst the costs are quietly pushed to the least responsible for them. As long as 19th-century energy powers 21st-century illusions, there will be no green transition––only a displaced one.

References

Banga, R. (2022). WTO Moratorium on Customs Duties on Electronic Transmissions: How much tariff revenue have developing countries lost? (No. 157). Research Paper.

Burrington, I. (2015, December 16). The Environmental Toll of a Netflix Binge. The Atlantic. https://www.theatlantic.com/technology/archive/2015/12/there-are-no-clean-clouds/420744/

European Parliament. (2025, September 10). CBAM: Parliament adopts simplifications to the EU carbon leakage instrument. Europa.eu. https://www.europarl.europa.eu/news/en/press-room/20250905IPR30181/cbam-parliament-adopts-simplifications-to-the-eu-carbon-leakage-instrument

Climate Rights International. (2025, October 16). Indonesia: Widespread Environmental Rights Violations in Nickel Industry. Press Release. https://cri.org/indonesia-widespread-environmental-rights-violations-nickel-industry/

Diprose, R., Kurniawan, N., Macdonald, K., & Winanti, P. (2022). Regulating sustainable minerals in electronics supply chains: local power struggles and the ‘hidden costs’ of global tin supply chain governance. Review of International Political Economy, 29(3), 792–817. https://doi.org/10.1080/09692290.2020.1814844

Fernandéz, L. (2025, November 28). Emissions from electricity generation Indonesia 2014-2023. Statista. https://www.statista.com/statistics/1303565/indonesia-emissions-from-electricity-generation/

Khan, A. (2023). Rules on digital trade in the light of WTO agreements. PhD law dissertation, school of law, Zhengzhou University China.

Monserrate, S. G. (2022). The cloud is material: On the environmental impacts of computation and data storage.

Penn, I., & Tan, E. (2024, December 27). Amazon Says It Reached a Climate Goal Seven Years Early. The New York Times. https://www.nytimes.com/2024/07/10/business/energy-environment/amazon-clean-energy-climate-change

Rathi, R. (2026). The WTO E-Commerce Moratorium After MC13 (2024): Legal Status, Development Impacts, and India’s Strategy to 2026. Development Impacts, and India’s Strategy to.

Schien, D., Shabajee, P., Yearworth, M., & Preist, C. (2013). Modeling and assessing variability in energy consumption during the use stage of online multimedia services. Journal of Industrial Ecology 17, 6(2013): 800-813.

Vrikki, P. (2024). Measuring Up? The Illusion of Sustainability and the Limits of Big Tech Self-Regulation. Sustainability, 16(23), 10197. https://doi.org/10.3390/su162310197

Zero Carbon Analytics. (2024, November 7). Carbon Border Adjustment Mechanisms require coordinated global action. https://zerocarbon-analytics.org/finance/carbon-border-adjustment-mechanisms-require-coordinated-global-action/