Before the Lock- In
Rethinking the Future of AI Infrastructure in UK and Ireland
This project was developed as part of Asmita’s Master’s thesis in the MDes Design Futures programme at the Royal College of Art, London (2024–2025). Spanning over six months, the work was situated within the RCA’s Independent Research Project (IRP) framework and supported by tutors, peers, and critical review sessions. The project investigates the design of equitable, sustainable, and adaptable AI infrastructures through a futures-oriented and systems-led design approach.
Externally, the project collaborated with a local authority based in Ireland, where a strategic backcasting workshop was conducted to co-develop long-term visions and planning approaches for AI infrastructure. Additional insight was gained through expert interviews with David Davies (Director, Data Centres at Arup) and Hanna Barakat (Research Manager, Computer Says Maybe), helping to ground the research in current industry and policy perspectives.
The project outcomes were showcased at the RCA School of Design Expo from July 17-19 2025, as part of the Royal College of Art’s annual graduate exhibitions.
AI Colonialism
By the mid-2030s, the UK entered a new era of AI-led growth, as national policy aggressively backs data infrastructure expansion.
Designated as Critical Infrastructure, data centers are fast-tracked through planning pipelines especially in economically deprived regions, framed as tech-led regeneration. Local councils, stretched for resources, sign long-term land and energy deals with hyperscale providers, attracted by promises of jobs and digital investment. Yet these developments operate with minimal transparency, locked behind NDAs and obscure
corporate structures. During repeated summer heatwaves, energy blackouts begin to affect domestic users, while data centers continue to hum, powered by private nuclear microgrids and hydrogen cells. Communities protest rising energy bills,
water shortages, and noise pollution, only to find their
concerns bypassed under emergency national policy. AI is now used to predict grid load, automate energy draws, and regulate entire utility networks in favour of centralised efficiency.
The race to dominate global AI capabilities reproduces colonial logics of extraction, justified by “innovation” and “security.”
Globally, this is part of a broader shift: infrastructure
giants exploit international overcapacity and weaker regulations to host overflow demand. Multinational cloud providers simply relocate to more permissive jurisdictions, or lobby governments to suppress local opposition. Austerity-weary citizens reject the framing of AI infrastructure as “the future,” questioning who benefits from automation and who absorbs the environmental cost. Worker resistance and climate lawsuits grow, but so too does state surveillance, powered by the very infrastructure now under critique.
“AI no longer lives in the cloud – it colonises the land, the grid, and the law.”
Platform Fragmentation
By the mid-2030s, the UK entered a new era of AI-led growth, as national policy aggressively backs data infrastructure expansion.
Designated as Critical Infrastructure, data centers are fast-tracked through planning pipelines especially in economically deprived regions, framed as tech-led regeneration. Local councils, stretched for resources, sign long-term land and energy deals with hyperscale providers, attracted by promises of jobs and digital investment. Yet these developments operate with minimal transparency, locked behind NDAs and obscure
corporate structures. During repeated summer heatwaves, energy blackouts begin to affect domestic users, while data centers continue to hum, powered by private nuclear microgrids and hydrogen cells. Communities protest rising energy bills,
water shortages, and noise pollution, only to find their
concerns bypassed under emergency national policy. AI is now used to predict grid load, automate energy draws, and regulate entire utility networks in favour of centralised efficiency.
The race to dominate global AI capabilities reproduces colonial logics of extraction, justified by “innovation” and “security.”
Globally, this is part of a broader shift: infrastructure
giants exploit international overcapacity and weaker regulations to host overflow demand. Multinational cloud providers simply relocate to more permissive jurisdictions, or lobby governments to suppress local opposition. Austerity-weary citizens reject the framing of AI infrastructure as “the future,” questioning who benefits from automation and who absorbs the environmental cost. Worker resistance and climate lawsuits grow, but so too does state surveillance, powered by the very infrastructure now under critique.
“AI no longer lives in the cloud – it colonises the land, the grid, and the law.”
