As computing moves from millions of distributed devices into centrally managed assets and sites, the opportunities for circularity are shifting
27 July 2026
The latest updates and expert perspectives in Technology, delivered straight to your inbox when they matter most.
AI is not only changing what computing can do, but also where computing happens, who controls the physical assets behind it, and how those assets should be managed.
For decades, technological progress largely meant putting more computing power into the devices we carried and used every day. Every new generation of smartphone, laptop, or workstation became faster because the hardware itself became more capable.
Artificial intelligence and cloud computing are beginning to change that equation.
Devices are increasingly acting as interfaces to computing power that exists elsewhere: in hyperscale data centres, cloud platforms, and progressively specialised AI infrastructure. We are moving towards a future in which the intelligence behind digital services does not reside on the device itself, but in the infrastructure behind it.
This architectural shift is already reshaping investment decisions, supply chains, and digital infrastructure, with significant implications for circularity.
As computing moves from millions of distributed devices into centrally managed assets and sites, the centre of gravity for circularity moves with it.
The rapid growth of AI is driving unprecedented investment in data centres and cloud infrastructure, with estimates suggesting over $400 billion was invested by leading technology companies in 2025 alone. Hyperscale operators are expanding capacity globally to support more compute-intensive services, while organisations across almost every sector are becoming more dependent on cloud-based applications and AI models.
The result is a shift in where physical assets sit within the technology ecosystem: rather than computing capacity being embedded entirely within devices owned by individuals or organisations, an increasing proportion is concentrated within infrastructure owned and managed by cloud providers and hyperscalers.
From a sustainability perspective, this shift in ownership fundamentally changes what is possible through circularity.
AI infrastructure is creating something that has traditionally been difficult to achieve in technology: predictable, large-scale asset flows. In centrally managed fleets, operators know where servers are, how they are configured, how intensively they have been used, and when they will need to be upgraded or replaced.
That visibility and scale allow circularity to be built into planned asset cycles in several practical ways:
We have already seen this dynamic play out in the EV sector, where growing fleets of batteries have created the predictable asset volumes needed to justify investment in collection, remanufacturing, and recycling infrastructure. AI infrastructure could create similar conditions for servers, accelerators, cooling equipment, racks, and other high-value computing equipment.
All in all, AI is accelerating demand for processors, memory, storage, and critical minerals. The new sustainability opportunity of AI is not that computing becomes less resource intensive, but rather that the resources it requires increasingly sit within systems where ownership, visibility, and lifecycle management already exist and can be deliberately strengthened.
While this shift towards centralised computing creates significant opportunities for resource optimisation and circularity, it also concentrates environmental impacts in ways the technology sector is only beginning to fully appreciate.
Energy demand is perhaps the most visible example: electricity demand from data centres soared by 17% last year, with AI-focused sites climbing even faster. AI workloads consume substantially more electricity than traditional computing, driving rapid expansion of data centre capacity across multiple regions. Water demand for cooling is increasing alongside it (projected to equal the basic annual domestic needs of 1.3 billion people by 2030), while growing demand for semiconductors is placing additional pressure on global supply chains for critical materials.
This concentration of energy and resource demand means local communities increasingly experience the environmental consequences of digital infrastructure directly.
Historically, sustainability in tech often focused on improving the efficiency of individual devices. While that remains important, it is now only one part of a much larger system.
Many technology companies have committed to increasing recycled content within future products and infrastructure, but the pace of AI investment means demand for critical materials is growing far faster than end-of-life assets are becoming available. Even with highly effective collection and recycling systems, secondary material supply alone is unlikely to satisfy future demand.
In this tech ecosystem, circularity is essential not because it can replace primary extraction entirely, but because it reduces dependence on it while helping retain scarce materials within the economy for longer.
As computing becomes increasingly centralised, sustainability becomes a question of infrastructure design: where facilities are built, how they are powered, how equipment is managed throughout its lifecycle, and how materials are recovered when assets reach the end of service.
This transition is also reshaping where investment flows. Countries with growing digital economies, expanding renewable energy capacity, and supportive industrial policy are increasingly positioning themselves as locations for cloud and AI infrastructure.
India provides one of the clearest examples. Major investments announced by Amazon Web Services, Microsoft, and Google reflect not only the country’s growing demand for digital services, but its strategic importance within the next generation of global computing infrastructure.
Across Africa, investment is also beginning to accelerate. Earlier this year, the International Finance Corporation committed US$100 million to support the expansion of data centre operator Raxio across several African countries, recognising digital infrastructure as an increasingly important foundation for economic development.
These investments represent genuine opportunities. New digital infrastructure can support economic growth, improve digital capability, attract skilled employment and strengthen local technology ecosystems. Countries that have historically participated in technology value chains primarily through resource extraction or end-of-life processing can now capture more value through infrastructure ownership, digital services, and advanced technical capability.
However, there is also a risk that the geography of environmental impact shifts alongside these opportunities. Over recent decades, environmental regulation surrounding mining has become increasingly stringent across many developed economies, while extraction has continued to expand in jurisdictions where regulatory oversight, enforcement capacity, or environmental protections are often less mature. The result has been that many of the environmental and social impacts associated with producing critical materials have become geographically separated from the markets that ultimately benefit from them.
As AI infrastructure expands globally, there is a similar risk that environmental burdens simply relocate rather than reduce. Data centres built without access to low-carbon electricity, increasing pressure on local water resources, or weak requirements for asset recovery and responsible end-of-life management could repeat that pattern. The opportunity for emerging economies is significant, but so too is the opportunity to establish high expectations for how that infrastructure is designed, operated, and eventually decommissioned.
The question now is whether sustainability standards will evolve with this new geography of computing, and rapidly enough.
The sustainability debate around AI has largely focused on operational impacts such as electricity consumption and water use. Those discussions are important, but they are only part of a broader transition taking place across the technology sector.
The more fundamental shift is architectural: as computing moves away from the device and towards centrally managed infrastructure, circularity moves from product-level efficiency to system-level design.
For technology companies, this means thinking beyond the efficiency of individual products. It means considering how infrastructure is designed for modular upgrades, how hardware remains in productive use for longer, how components are recovered, and how critical materials are retained within managed systems rather than lost at end of life.
Realising this opportunity will also require greater attention to interoperability. Many hyperscale operators continue to optimise hardware stacks around proprietary architectures, custom server designs, and organisation-specific components. While these approaches can deliver performance gains within individual environments, they also risk reducing compatibility across the wider market, limiting opportunities for reuse, refurbishment, and remanufacture once equipment leaves its original owner. Assets that retain significant functional value may therefore struggle to find productive second or third lives simply because there are too few pathways for them to be redeployed.
This is not a new challenge for the technology sector. The cloud itself was built on shared technical standards that allowed digital systems to interoperate and scale globally. As AI turns cloud infrastructure into one of the world’s largest concentrations of physical assets, there is an opportunity to apply the same thinking to circularity. Greater standardisation around areas such as component interfaces, modularity, and lifecycle information could help strengthen secondary markets, extend asset lifetimes, and improve material recovery across the industry.
No single organisation can solve this in isolation. Establishing those standards, and the systems needed to support them, will require collaboration across manufacturers, hyperscalers, recyclers, standards bodies, and policymakers. If done well, that collaboration could help ensure that circularity scales alongside AI infrastructure, rather than becoming constrained by proprietary ecosystems.
For organisations investing in AI and cloud infrastructure, the challenge is no longer simply to deploy more compute. It is to make infrastructure decisions that optimise performance while shaping future resource flows, operational resilience, and long-term environmental outcomes. Circularity is no longer something to consider once hardware reaches the end of its life; it is becoming a design principle for the infrastructure itself.
This is why the transition to AI-enabled computing should be seen as more than a software revolution. It represents a fundamental shift in the architecture of computing: one that changes where physical assets are concentrated, who controls them, and where the greatest opportunities exist to improve resource efficiency, recover critical materials, and reduce environmental impacts.
The next phase of digital growth will be shaped not only by those who can deploy AI fastest, but by those who can build the infrastructure behind it responsibly. Circularity should therefore be treated as a strategic design requirement, not a downstream waste-management solution. The organisations that recognise this early, and design their infrastructure for circularity, resilience, and responsible resource management, will be better positioned not only to manage the impacts of this transition, but also to shape the next generation of digital value creation.
Anthesis helps technology companies translate circularity into tangible business value by reimagining products, materials, and revenue models with circularity in mind. We combine deep tech sector expertise with proven strategies to drive resilience, unlock new revenue streams, and reduce lifecycle impacts.
Transforming production & consumption with real-world expertise to deliver a resilient circular economy, and a healthy, sustainable future for all.
We are the world’s leading purpose driven, digitally enabled, science-based activator. And always welcome inquiries and partnerships to drive positive change together.
