The Energy, Water, and Waste Behind the World’s Fastest-Growing Technology
By Canon Otto
Convener, Global Sustainability Summit
Founder, CleanCyclers
Contributor, SustainabilityUnscripted
Artificial Intelligence is being celebrated as the defining technology of our time.
It promises efficiency, optimisation, automation, and innovation across nearly every sector—from healthcare and finance to agriculture and climate modelling. Governments are investing in it. Corporations are racing to deploy it. Societies are rapidly integrating it.
But beneath this acceleration lies a question we are only beginning to confront:
What is the environmental cost of intelligence at scale?
At SustainabilityUnscripted, this question is no longer theoretical. It is becoming central to how we evaluate the sustainability of the digital economy itself.
The Illusion of the “Invisible” Industry
Digital technologies often carry an assumption of lightness.
No smoke.
No visible waste.
No obvious extraction.
AI, in particular, is frequently perceived as an intangible system—algorithms running quietly in the background. But this perception is misleading.
Artificial Intelligence is not weightless.
It is built on physical infrastructure.
Behind every AI model lies a vast network of:
- Data centres
- Servers and processors
- Cooling systems
- Transmission networks
- Energy grids
These systems consume significant amounts of electricity, water, and raw materials.
The environmental cost is not absent.
It is simply hidden.
The Energy Demand of Intelligence
Training and running large AI systems requires enormous computational power.
Data centres—the backbone of AI infrastructure—operate continuously, processing vast volumes of information. As AI adoption accelerates, so does the demand for these facilities.
The result is a sharp increase in:
- Electricity consumption
- Grid pressure
- Dependence on energy-intensive computing
In regions where energy systems still rely heavily on fossil fuels, the carbon footprint of AI becomes substantial.
At the Global Sustainability Summit, a recurring concern is emerging: as industries decarbonise, digital infrastructure may become one of the fastest-growing sources of emissions if left unaddressed.
Efficiency gains from AI cannot be evaluated in isolation. They must be weighed against the energy required to produce them.
The Water Behind the Cloud
Less visible, but equally significant, is the role of water.
Data centres generate heat. To maintain operational stability, they require cooling systems—many of which rely heavily on water. In some cases, millions of litres of water are used to cool servers and maintain optimal performance.
This creates a new and often overlooked pressure point:
- Increased water demand in already water-stressed regions
- Competition between digital infrastructure and local communities
- Long-term strain on freshwater systems
The “cloud” is not abstract.
It is deeply connected to physical ecosystems.
Through the lens of Canon Otto, sustainability must account for these hidden resource flows. If innovation depends on invisible consumption, then sustainability must make that consumption visible.
The E-Waste Problem of Rapid Obsolescence
AI does not only consume energy and water. It also drives hardware turnover.
Advanced AI systems require specialised processors—GPUs and other high-performance chips—that evolve rapidly. As new generations of hardware are released, older systems are replaced, often before the end of their functional life.
This creates a growing stream of electronic waste:
- Servers decommissioned at scale
- Obsolete processors and components
- Complex materials that are difficult to recycle
At CleanCyclers, this issue is particularly urgent.
E-waste is one of the fastest-growing waste streams globally. Without strong recovery and recycling systems, valuable materials are lost, and hazardous substances risk contaminating environments.
This is where creativity turns waste into opportunity—but only if circular systems are intentionally designed into the lifecycle of digital infrastructure.
Efficiency vs Expansion: The Rebound Effect
One of the most complex challenges associated with AI is the rebound effect.
AI can improve efficiency:
- Optimising logistics
- Reducing energy waste
- Enhancing resource management
But increased efficiency often leads to increased usage.
As AI becomes cheaper and more accessible, demand expands. More models are trained. More data is processed. More infrastructure is built.
The result is a paradox:
Efficiency gains can drive total consumption upward.
At SustainabilityUnscripted, this dynamic is increasingly recognised as a defining challenge of the digital sustainability era. Technology alone does not guarantee environmental improvement. Without constraints, it can accelerate resource use.
Rethinking Digital Sustainability
The environmental cost of AI does not mean the technology should be rejected.
It means it must be governed differently.
A sustainable digital future requires:
- Renewable energy integration for data centres
- Water-efficient cooling technologies
- Longer hardware lifecycles and repairability
- Circular systems for e-waste recovery
- Transparent reporting of digital carbon and resource footprints
These are not optional considerations. They are essential to aligning technological progress with environmental responsibility.
At the Global Sustainability Summit, conversations are shifting toward this reality: digital transformation must now be matched with sustainability transformation.
The Role of Systems Thinking
The challenge of AI’s environmental impact is not isolated. It reflects a broader pattern.
Technologies often solve visible problems while creating invisible ones.
This is why systems thinking is central to the work of CleanCyclers. Waste, energy, water, and materials are interconnected. Addressing one without considering the others leads to unintended consequences.
AI is no different.
Its benefits are real.
But so are its externalities.
A Final Reflection
Artificial Intelligence represents one of humanity’s most powerful tools.
But power without accountability creates imbalance.
If the digital economy continues to expand without integrating sustainability into its core design, it risks becoming a major environmental burden—hidden behind the illusion of innovation.
Through CleanCyclers, and SustainabilityUnscripted, one principle remains clear:
There is no such thing as a sustainable future built on invisible costs.
The intelligence we are building must now learn to account for the systems it depends on.
