Midstream: Accumulation and in-use stocks

Silicon's in-use stock roughly doubles across the projection period, from approximately 190 million tonnes in 2020 to 380 million tonnes in 2050, but its composition reveals a striking disconnect between what dominates by mass and what drives criticality. The in-use stocks chart below shows that the stock is overwhelmingly dominated by ferroalloy-grade material—silicon embedded in steel and cast iron—which accounts for the large majority of total in-use stock throughout the period and remains dominant at the end of the projection, accounting for 80% of silicon stocks by 2050. Chemical and alloying applications make up most of the remainder (around 15%), feeding primarily into glass and ceramic applications.

The two applications that underpin almost all of silicon's strategic importance—solar photovoltaics and semiconductors—are, by mass, almost invisible within the total stock. Solar-grade polysilicon grows from 3.7 million tonnes in 2020 to 17.8 million tonnes by 2050, but remains only 4% of total silicon in-use stocks by 2050. Semiconductor-grade polysilicon is smaller still, remaining below 1% of total stock throughout the entire projection period, despite being associated with the most strategically sensitive silicon supply chains examined in this study.

This divergence between mass and strategic value has important implications for how silicon circularity should be approached. A circularity strategy focused on volume would naturally prioritise ferroalloy applications, such as steel and cast iron, since they contain the overwhelming majority of silicon in use. However, as the downstream section shows, silicon incorporated into steel and cast iron is essentially unrecoverable. A strategy focused on strategic value would instead target semiconductor and solar streams, which represent only a small fraction of total tonnage but carry far greater importance for supply chain resilience. These material streams require fundamentally different interventions, and treating them as equivalent risks either pursuing large-volume streams that cannot realistically be recovered or overlooking smaller streams with much higher strategic value. Even within the strategic-value stream, however, recovery does not automatically translate into displacement of primary production: high purity requirements of, for instance, semiconductor-grade silicon mean that when this is recovered it often re-enters the supply chain at a lower value than before.¹⁵⁶

The solar stock in particular warrants attention. It is growing rapidly, is comparatively well characterised, and much of the first generation of large-scale photovoltaic deployment is now approaching end-of-life—creating one of the few meaningful near-term opportunities for silicon recovery.¹⁵⁷ This emerging opportunity is already reflected in growing research and investment into recovering solar-grade silicon from retired photovoltaic panels.¹⁵⁸