Big Tech’s data centres and AI growth are driving emissions and environmental pressures, raising questions about real sustainability
As the IndiaAI Summit, scheduled to be held in mid-February 2025 in New Delhi, brings global attention to the future of AI-led growth, the environmental externalities associated with the rapid expansion of AI-driven and AI-enabled technology sectors can no longer remain a secondary concern.
Big Tech’s expanding data centre footprint, together with the rapid growth of AI-driven applications, has made technology-related environmental challenges more significant than ever. Under-reported emissions, concerns around greenwashing, creative accounting, and sharply rising electricity consumption have raised questions about whether AI and Big Tech-related emissions may ultimately outweigh the sustainability benefits often attributed to these technologies. In this context, the relationship between technology, economic growth and climate outcomes becomes increasingly important in the broader sustainability debate and warrants closer attention.
Big Tech’s expanding data centre footprint, together with the rapid growth of AI-driven applications, has made technology-related environmental challenges more significant than ever.
Big Tech refers to large, dominant multinational technology companies that are key drivers of the global digital economy. The technology sector has expanded rapidly in recent years and is expected to grow strongly throughout the decade, with electricity demand from ig Tech and AI projected to increase by more than 60 percent by 2030.
Table 1: Some Major Reports and Research on Big Tech and the Environment Nexus
| Organisation | Key Findings and Insights |
| Greenpeace, 2019 | Gave a ranking to major Big Tech companies on their energy policies (procurement and transparency) |
| University of Massachusetts, 2019 | Large AI models can generate emissions equivalent to the lifetime emissions of cars |
| OECD, 2022 | The lifecycle of AI and the Environment |
| Carbon Disclosure Project | Calls out under-reporting and creative accounting for emissions |
| Uptime Institute, 2023 | Water intensity in data centres |
| Influence Map, 2023 | Found a wedge between sustainability policy wedges and lobbying against climate regulation |
Source: Author’s Own Compilation.
The tech (or digital) sector generates emissions and environmental pressures through energy-intensive data centres, continued dependence on fossil fuels for electricity, growing volumes of electronic waste, high water consumption for cooling, and resource-intensive activities such as mining and manufacturing that contribute to pollution and ecosystem stress. These impacts can also be framed under the UN SDG framework, as shown in 1. On the emissions front, the environmental footprint of Big Tech can be grouped into two main categories. Scope 1 emissions emerge from sources directly owned or controlled by firms, such as on-site fuel use and backup generators. Scope 2 emissions arise from purchased energy, most notably the electricity consumed by data centres and related digital infrastructure. As AI deployment expands, data centres are expected to become a more significant driver of emissions. Currently, emissions from data centres come from electricity use (Scope 2). Direct emissions from data centres themselves (Scope 1) account for roughly one percent of total emissions worldwide.
Figure 1: Environmental Impact of Big Tech: Recipient Climate-Related SDGs
Source: Generated using AI with the author’s own inputs
Although emissions accounting and the SDG framework provide useful tools to measure and frame Big Tech’s environmental footprint, they offer only a partial view. The environmental impact of digitalisation is not static; it evolves with the scale, intensity, and trajectory of technological change, creating effects that unfold over time.
Much like the historical relationship between industrialisation and emissions, the link between digitalisation and environmental outcomes is non-linear and dynamic. Evidence suggests that in the early stages of digital adoption, emissions often rise as energy demand increases and efficiency gains lag behind scale effects. As digitalisation deepens, this relationship can reverse. This is because, at more advanced stages, emissions and environmental footprints tend to decline, with declines driven by industrial upgrading and the diffusion of green technologies. This produces a classic inverted U-relationship between technological adoption and emissions over time, a pattern also observed in China.
Much like the historical relationship between industrialisation and emissions, the link between digitalisation and environmental outcomes is non-linear and dynamic.
This dynamic aligns with the logic of creative destruction, in which technological progress renders older and more resource-intensive technologies obsolete and replaces them with more efficient alternatives. While major innovation shocks such as the internet, smartphones, and more recently AI are typically growth-enhancing in standard macroeconomic frameworks, they can also intensify environmental pressures during periods of rapid expansion. From an environmental perspective, a key question concerns how such technological innovations affect emissions and ecological outcomes. Evidence from BRICS economies suggests that when technological progress is paired withclean technologies and efficiency gains, carbon emissions tend to decline.
Technology-driven economic expansion, including the rapid establishment of data centres and the growing deployment of AI, has intensified Big Tech’s carbon footprint. At the same time, technological progress itself offers the means to mitigate these emissions, provided there is a robust and binding environmental policy. This underscores the need for a form of creative destruction that is environmentally constructive, where high-emission technologies are gradually replaced by cleaner and more efficient alternatives. To achieve this, the key lies in identifying and encouraging positive technological shocks that align innovation with cleaner growth, rather than allowing innovation to translate into environmentally harmful outcomes. Evidence suggests that technological shocks oriented towards clean energy deployment and efficiency improvements can significantly enhance sustainable development outcomes.
Against the backdrop of the technology–environment nexus and the growing role of technology shocks and clean-technology progress, a critical question arises around how large technology firms manage their expanding environmental footprint and the effectiveness of these strategies. Major technology companies have adopted a wide range of measures, including public net-zero commitments by firms such as Siemens, Nvidia, IBM, Amazon, and Microsoft; large-scale renewable energy procurement and climate-innovation initiatives, including investments in renewable capacity, reforestation, and decarbonisation research by Microsoft, Amazon, and Apple; and efforts to improve data-centre efficiency through advances in cooling technologies and hardware optimisation. Select operators, such as Start Campus, have further moved towards 100 percent renewable-powered data centres using seawater-based cooling systems, significantly reducing water and electricity intensity. In parallel, firms have increasingly relied on market-based mechanisms such as carbon credits, exemplified by Amazon’s recent launch of a carbon-credit service, alongside the deployment of AI-enabled climate solutions for monitoring ecosystems, optimising electricity consumption, and managing power grids, including platforms such as Geoverse that analyse millions of locations globally to assess nature and biodiversity.
In effect, decarbonisation risks becoming an exercise in signalling and reputation management rather than a process of substantive, ground-level environmental transformation.
However, despite the breadth of these initiatives, central and increasingly contentious concern remains the role of greenwashing and positive environmental signalling, which is rampant. Sustainability strategies are often accompanied by carefully curated disclosures, self-regulated measurement and reporting systems, and headline claims that rely on practices such as the extensive use of Renewable Energy Certificates (RECs), the outsourcing of emissions across scopes, the framing of operational efficiency gains as decarbonisation pathways, and heavy dependence on ESG branding, carbon offsets, and flexible emissions accounting. Many large technology firms, including Amazon and Microsoft, make use of market-based mechanisms and environmental disclosures as part of their broader sustainability strategies. Collectively, these approaches function less as instruments of genuine emissions reduction and more as mechanisms for managing perceptions around environmental externalities, distorting capital and resource allocation, weakening climate accountability, and delaying meaningful decarbonisation. In effect, decarbonisation risks becoming an exercise in signalling and reputation management rather than a process of substantive, ground-level environmental transformation.
Looking ahead, the rapid expansion of artificial intelligence and data-centre infrastructure, alongside rising electricity demand, makes it essential to curb environmental signalling. This underscores the need for pragmatic emissions and environmental management frameworks that firms cannot simply optimise on paper, supported by stronger regulatory oversight. Three caveats emerge in this technology–growth–environment nexus: (a) environmental signalling must be replaced with verifiable emissions-reduction actions; (b) data-centre operators must prioritise genuine operational measures such as clean and reliable renewable power over efficiency metrics alone; and (c) mitigation efforts should extend beyond core operations to include complementary approaches, such as nature-based solutions, to address residual emissions. A shift from perception management to substantive clean-technology adoption is critical for aligning technological progress with environmental outcomes.
Manish Vaidya is a Research Assistant at the Observer Research Foundation.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Manish Vaidya is a Research Assistant with ORF’s Centre for New Economic Diplomacy. His work centres on research and active engagement in applied economics, with a ...
Read More +
PREV
