Popular narratives would have you believe that globalisation is merely the consequence of technological innovation that facilitates increased economic and political integration. This story intentionally obscures how upscaling from local contexts to global arenas is a vital strategy for capital to continue accumulating. It has also been a successful strategy to thwart labour organising by pitting communities in one locality against communities elsewhere. 

Globalisation is an active policy choice. Yet, as with any policy imposed from the top down, there is scope for grassroots resistance to contest it. The systems analysis that underpins labour environmentalism highlights how localised struggles are intertwined: they fight against the same overarching structure of global capitalism that constitutes a broad array of intersecting oppressions. 

The rapidly growing AI tech sector is emblematic of this. By following AI supply chains, we can trace the environmental devastation and exploitation of human labour that is usually hidden, exposing interconnected injustices and identifying the struggles that resist them. 

The labour behind the idea

In 2024, the Kenyan national assembly proposed the Business Laws (Amendment) Act, aiming  to enhance the country’s appeal for foreign investment by cutting down on capital regulation. Prioritising the financial interests of multinational tech companies, Information and Communication Technology Parks are being developed as Special Economic Zones (SEZ) in which national labour regulations and environmental protection protocols are suspended in favour of cheaper labour and resources. 

Of course, multinational tech companies are not just self-proclaimed leaders in innovation, but also in labour and resource exploitation. The economic productivity the Kenyan SEZ authority boasts of is rarely captured for local benefit but instead funnelled back to foreign investment funds. 

What remains are jobs defined by their precarity: content moderators and data labellers employed to sift through AI data and review hours of potentially traumatic content, in unsafe working conditions and for little pay. Mental health challenges due to exposure to graphic content are commonplace. In 2023, the African Content Moderators’ Union (ACMU) filed a case against tech giant Meta, citing low pay, union-busting practices and human trafficking of workers. 

Facing increasing pressure, Meta has since moved many of its content moderation facilities to Ghana, with workers there now facing the same precarious working conditions labour unions were organising against in Kenya. Capital’s ability to shift scale and location is leveraged against labour’s local organising power. In the digital economy, labour remains just as invisibilised as the environmental impact of digital infrastructure. 

The cost of the cloud

The environmental impact of training and maintaining AI systems is an environmental burden which companies like Meta and OpenAI try their best to conceal. Data centres have become the newest frontier for exploiting the natural world. 

Hyperscale AI data centres generate large amounts of e-waste due to high innovational turnover, noise pollution displacing people and wildlife, and ever-increasing pressure on energy demand. Their strain on freshwater resources is enormous. A single medium-length question posed in Chat GPT can consume around half a litre of water for processing and cooling. Just this year, the United Nations Institute for Water, Environment, and Health presented its report warning of ‘water bankruptcy’ for freshwater sources around the globe. 

In the second half of 2025, activists managed to stop or delay 20 data centre projects across the United States, affecting investments of $64 billion

In Chile, 16 new data centres have been built since the onset of chronic drought conditions in 2010, but not without opposition. Activist groups like the Socio Environmental Community Movement for Land and Water (Mosacat) have had success in resisting the construction of new Google data centres in Santiago. Due to continuous community action, one project had its license suspended and is pending a renewed environmental impact assessment. 

Organizing efforts against data centres are also happening closer to home for the responsible US tech-giants. In the second half of 2025, activists managed to stop or delay 20 data centre projects across the United States, affecting investments of $98 billion.   

The new old extractivism

Elsewhere, AI computing’s new extractivist frontier remains closely linked with deeply entrenched legacies of colonial mining extractivism. 

The Democratic Republic of the Congo (DRC) is home to some of the world’s largest amounts of copper, cobalt, lithium and coltan reserves. These minerals are extracted to be used for data centre hardware that powers AI. 

Cobalt mining in the DRC is linked to food shortages, drinkable water scarcity, economic insecurity, and severe health risks. Mine workers are buried alive in collapses, brutalized, and suffer from respiratory disease. 

China and the United States own most of the mines in the DRC. The Congolese people have no say over what is mined, how it’s mined, or if it even should be mined. Congolese miners and lawyers are taking legal action, but political instability has left their actions unacknowledged. 

The exploitation of mine workers, data labellers, and content moderators is intertwined and fuelled by environmental ravaging. Their day-to-day struggles are different, but through transnational collaboration they could strengthen their efforts to protect people and land from globalised tech corporations. 

Mechanisms of transnational organising

Retracing the AI tech supply chain shows that there is no singular global solution that will meet the diverse needs of all local communities organising to resist. Centring the localised communities and their needs can enable transnational cooperation that holds space for the multitude of tactics required to push for transformative change. 

Legal activism denouncing environmental destruction and labour exploitation can help to name and shame companies and apply institutional pressure, while shared mutual aid funds make solidarity tangible by meeting material needs. Transnational organizing that can connect these different struggles harbours the potential to bridge divides between North and South and red and green. 

Beyond global unions, organising in collaborative platforms and developing shared agendas also means connecting to other social justice movements. Linking and sharing different experiences of exploitation and resistance plays an important part in uniting workers from different contexts. 

The tech companies that exploit land and people alike rely on neoliberal globalisation for their profits. Yet, they face opposition wherever they go. Supported by their strong local base, those that resist are globally connected too.