What circularity changes and what it cannot overcome

The scenario results point to four key implications for the material strategy of the CRMs analysed. Together, they show that circularity is neither a single intervention nor a substitute for other CRM strategies, such as supply expansion, substitution, or sufficiency. The effectiveness of circular interventions depends on the material system, the stage of stock development, the pace of decarbonisation, and the resilience of the systems through which recovered material flows.

1. Circularity's effectiveness depends on the material system

The first implication concerns the different ways circularity operates across CRM systems. Narrow, slow, and close act on different parts of the system, and their contribution depends on where the main constraint lies. For mature materials such as copper and nickel, where substantial stocks already exist and recovery systems are relatively developed, circularity can reduce the material required to maintain service and provide substantial ability to adapt and recover from supply disruptions. For emerging materials such as lithium and cobalt, by contrast, the available stock of end-of-life material is still small relative to rapidly growing demand. Circularity can therefore contribute less to reducing primary supply requirements in the near term, but rather contribute to freeing up more demand.

Silicon demonstrates a different limit. Its physical characteristics and purity requirements restrict the extent to which conventional recovery can return material to equivalent applications, leaving material efficiency and substitution as the more effective strategies. The implication is that circularity strategies need to reflect the physical and timing characteristics of each material system and thus cannot be treated as a uniform package.

2. Transition speed and material availability are interdependent

The second implication concerns the relationship between decarbonisation speed and material strategy. Faster decarbonisation increases material demand while leaving less time for recovery systems to develop, linking the pace of the transition to the adequacy of material supply. This should not be interpreted as an argument for slower decarbonisation. The material requirements of the transition should be considered in context: fossil-based energy sources involve vastly greater annual material flows than the critical raw materials required to replace them,¹⁸⁴ and must continue to meet ongoing consumption, whereas critical raw materials are largely embedded in long-lived infrastructure and remain available for recovery at the end of their service life. The environmental costs of continued dependence on fossil-based energy sources, and the climate consequences of delay, also fall outside this model entirely. This shows that decarbonisation, circularity, and material requirements—particularly critical raw materials—are interdependent, with trade-offs and dependencies between them.

Part of this interdependence concerns timing. Secondary-material availability depends on in-use stocks accumulated in earlier decades. At the same time, more ambitious decarbonisation pathways front-load deployment into the 2030s and 2040s, meaning that the grid infrastructure, vehicle fleets, and storage systems deployed then will reach their end-of-life substantially later. The analysis therefore captures the consequences of faster decarbonisation on demand in full, but only the earliest associated recovery benefits: for lithium and cobalt the secondary material stock available is still growing when the modelled window stops. For precisely those materials where circularity appears weakest within the time frame of the analysis, its benefits are still building at the end of the projection period. The relationship is therefore, at least in part, a matter of timing and sequencing rather than evidence of incompatibility between decarbonisation and circularity—making clear the need to build recovery capacity early to help address this timing mismatch.

It is also worth noting that material demand here is not driven by decarbonisation alone, as the exploration of material streams in the baseline analysis shows. The scenarios are framed around decarbonisation ambition and account for its effects across sectors—decarbonisation in steel, for example, reduces silicon demand—but other significant sources of growth, including digital infrastructure and advanced manufacturing, sit largely outside of the decarbonisation framing. The interdependence described above therefore captures only part of the pressure on material systems.

3. Circularity progress needs to capture more than recycling

The third implication concerns how circularity progress is measured. The analysis shows that different levers act on different parts of the material system, yet circularity is often assessed primarily through recycling and recovery. This means that some of the strategies identified here as particularly effective can remain largely invisible: a system could substantially reduce its primary material requirements through material efficiency and lifetime extension while showing little or no improvement against conventional end-of-the-pipe measures. This is particularly important for materials like silicon, where end-of-life interventions have little impact, or copper and nickel, where end-of-life measures are already mature. A more complete assessment of circularity would therefore follow the sequencing of the circularity levers shown in the analysis: first, how the design of a product and the demand for it affect the quantity of the material that is needed; then the extent to which existing stock is kept in use through repair, refurbishment, and reuse; and finally the recovery of material at end of life. This distinction is reflected in the results. A system can improve while its Recycling Input Rate falls, because narrowing and slowing have reduced the volume of material entering the waste stream. Conversely, a higher collection rate does not necessarily translate into more functional secondary supply if additional material is diverted to lower-value uses.

There is also a qualitative dimension that volume-based indicators cannot capture. A tonne of recovered material is counted the same regardless of how it is used afterward, even though its value can differ substantially when materials are scarce. Measuring circularity therefore requires more than tracking recovery volumes. It also needs to capture how much primary extraction is avoided, how much material remains in use, and how much is recovered into high-value applications that preserve its function.

4. Circularity is necessary but insufficient on its own

The fourth implication concerns the scope of circularity as modelled here. The levers can make materials go further, last longer, and return to service more often, but they cannot reduce the underlying demand for the service itself. For example, narrowing lowers the material required per unit of service delivered, but the number of units demanded is external to the model and remains consistent throughout. Where demand growth exceeds what narrowing, slowing, and closing can offset—which is the case for lithium and silicon—the remaining gap requires some combination of additional primary supply, substitution, or moderation of demand itself.

Circularity needs to be understood as one component of a broader material strategy, rather than being seen as a cure-all solution. The model's demand trajectories are external to the impacts circularity can have, and suppressed demand in the model represents material use avoided through scarcity and price rather than through deliberate choice. What the critical raw material is used for matters as much as the quantity used. Demand-side strategies that reduce or prioritise material use can complement circularity, particularly during the early decades when demand growth is steepest, and recovery systems are least developed.