Every time we ask artificial intelligence a question, generate an image or request help with an assignment, something remarkable happens. Within seconds, a machine produces what might once have taken hours of human effort.
It feels almost weightless. A few words on a screen, an instant response, a little digital magic.
But behind that magic are physical machines, enormous data centres, electricity networks and cooling systems. The digital world may live on our screens, but its environmental footprint exists in the real world.
The question is no longer just what artificial intelligence can do for humanity. It is what building this future will demand from our planet.
According to the International Energy Agency's 2026 report, Key Questions on Energy and AI, global data-centre electricity consumption reached an estimated 485 terawatt-hours in 2025. The agency projects that figure could approach 950 terawatt-hours by 2030. Artificial intelligence is a major driver of this growth, although the figures include data centres supporting other digital services.
That demand has consequences beyond electricity. Depending on their design and location, data centres can require substantial cooling, water and additional power infrastructure. Their hardware also depends on resource-intensive manufacturing and global supply chains.
Yet the story is not as simple as declaring AI an environmental villain. AI can help forecast extreme weather, improve agricultural decisions and optimise renewable-energy systems. Its environmental impact depends on how it is developed, powered and used. The challenge is ensuring that its benefits do not obscure its costs.
For Africa, this distinction matters enormously.
The continent needs digital infrastructure, investment and technological innovation. Kenya, in particular, has ambitions to strengthen its position as a regional technology hub. But a proposed $1.5 billion AI data-centre project in Mombasa, reported by TechTrendsKE in September 2026, illustrates why the infrastructure behind that ambition deserves scrutiny. The proposed development would combine AI computing facilities with dedicated power generation fuelled by liquefied natural gas.
The project is a proposal, not evidence of an established environmental impact. Nevertheless, it raises an important question: as Kenya attracts more energy-intensive digital infrastructure, how will the country balance investment opportunities with its climate commitments?
Where will the electricity come from? What cooling systems will facilities use? How will their environmental impacts be assessed, and will communities have access to relevant information before major decisions are made?
These questions are not arguments against development. They are questions about the kind of development we choose.
With COP31 approaching in Antalya, Türkiye, from 9 to 20 November, the timing could hardly be more relevant. On 21 September, the COP31 presidency announced the Antalya Pledge on AI, an initiative intended to encourage climate-responsible AI development. The presidency has called for greater transparency around AI's energy consumption and for infrastructure to align with global climate goals.
This creates an opportunity to move beyond celebrating AI's potential and ask how its environmental footprint will be measured and managed. The proposed pledge is an important starting point, but its practical impact will depend on the commitments that emerge and how they are implemented.
Governments should strengthen environmental oversight. Technology companies should disclose meaningful information about their electricity use, water consumption and emissions. Investors should assess environmental risks alongside financial returns. Communities should have opportunities to participate in decisions that affect their resources.
For Kenya and the wider African continent, there is an opportunity to shape responsible digital development rather than simply follow global trends. That means welcoming innovation while insisting on transparency, efficient resource use and credible environmental safeguards.
Artificial intelligence may help us imagine solutions to some of the world's most difficult problems. But innovation alone does not guarantee sustainability.
As delegates prepare to gather at COP31, we should ask not only how AI can help protect our planet, but also how we can protect our planet while building AI.
If the future is becoming more intelligent, shouldn't the way we build it become more responsible?
