Silver: The Connective Tissue of the Energy Transition

TL;DR

Silver has the highest electrical conductivity of any element (63.0 MS/m, ~5% better than copper) and the highest thermal conductivity of any metal (429 W/m·K).¹ It is irreplaceable in solar paste, high-current contacts, REBCO tape stabilization, and high-density power electronics. Roughly 72% of global silver production comes as a byproduct of copper, lead, and zinc mining² — which means silver supply does not respond to silver prices the way copper supply responds to copper prices. The market has run a structural deficit for four consecutive years.³ In November 2025 the USGS added silver to the official US critical minerals list.⁴ If solar PV growth continues on its current trajectory, the industry alone could require 14,000 tonnes/year of silver by 2030 against a global supply of roughly 34,000 tonnes — about 40% of the world’s silver going into one application.⁵

This is a materials-supply problem with no clean substitution path. It shapes every assumption a hardware company makes about cost-down curves on the products it sells.

§1 — Where the silver goes

Application	Silver use per unit	2024 industrial share
Solar PV cell (crystalline Si)	15–25 g per panel	~29% (197 Moz)5
Battery EV	25–50 g per vehicle (~67–79% more than ICE)	rising fast
AI datacenter (silver contacts, switchgear, thermal interfaces)	~8–12 Moz/yr aggregate	rising fast
REBCO superconducting tape	~1–3 µm Ag cap layer per tape km	small absolute, strategic
Conventional grid (relays, breakers, contactors)	Ag-Cu, Ag-Ni, Ag-W alloys	mature, large baseline

Substitution research has been ongoing for two decades. Aluminum paste in solar cells loses 1–2% absolute efficiency relative to silver paste — a margin that compounds badly across a 25-year panel lifetime. Copper plating on cells works in the lab and has been demonstrated at pilot scale, but the industrial line that runs at silver-paste throughput with copper-plating yield does not yet exist at gigawatt scale. The specific solar-paste reduction roadmap is real (silver content per cell has fallen from ~150 mg in 2010 to ~80 mg in 2024 and continues to drop), but the absolute volumes still rise because cell production rises faster than per-cell loadings fall.⁵

§2 — Why supply doesn’t respond

Roughly 28% of silver mine production comes from primary silver mines. The remaining 72% is byproduct from base-metal mining.² When silver prices double, primary silver miners can expand — but expanding a silver mine takes 7–10 years from discovery to production, and the global pipeline of advanced primary-silver projects is small. Byproduct silver responds to base-metal prices, not silver prices. Copper and zinc demand drives the bulk of silver supply, and that supply expands when base-metal economics justify the capex, not when silver does.

Recycling closes part of the gap. Industrial silver recovery from electronics, photographic chemistry, and catalyst spent material runs at roughly 180 Moz/yr globally — meaningful but not enough to close a 200+ Moz annual deficit by itself.³ PV-cell silver recovery from end-of-life panels is technologically feasible; the recycling infrastructure is not at the scale required.

§3 — Why a materials company cares

Three consequences flow directly into product strategy:

1. Cost-down assumptions on hardware that uses silver are weaker than typical learning-curve models suggest. A solar panel manufacturer that assumes silver tracks copper-style elasticity is going to mis-budget its 2027–2030 BOM. The same is true for any HTS magnet manufacturer assuming the silver cap layer is a rounding error.

2. Substitution research has tangible payback. Aluminum-paste cell architectures, copper-plated cells, and silver-reduction roadmaps for power electronics are not academic exercises. They are direct cost-of-goods levers in a market where the input metal is structurally constrained.

3. Recycling supply chains are an untapped second source. PV recycling, REBCO recycling (the silver cap is recoverable), and electronics waste streams contain orders-of-magnitude more silver than the niche recovery industry currently captures. The bottleneck is process — efficient leaching, separation, and refining at scale — and processes are materials-science problems.

Armadillo Labs is not a silver miner and not a recycler. We are a materials-science shop whose work in superconductors, thermoelectrics, and battery materials runs through the same supply-chain reality as everyone else’s. When we screen a candidate REBCO replacement, the silver-stabilization question is part of the engineering economics. When we look at battery chemistries and grid interconnect, silver consumption shows up in the BOM. The metal is the connective tissue. Designing around its constraints is part of the work.

---

Footnotes

1. Standard reference values; CRC Handbook of Chemistry and Physics, 103rd ed. (2022), Section 12. ↩

2. Silver Institute, World Silver Survey 2024 (Metals Focus, April 2024), Section 2 (Mine Production). ↩ ↩²

3. Silver Institute, World Silver Survey 2024, Section 1 (Market Balance). Cumulative 2020–2024 deficit ≈ 678 Moz. ↩ ↩²

4. U.S. Geological Survey, “2025 List of Critical Minerals,” final designation published November 2025. ↩

5. Hallam, B. et al., “The silver learning curve for photovoltaics and projected silver demand for net-zero scenarios” Progress in Photovoltaics: Research and Applications 30, 1071–1086 (2022). Updated industry projections from Ghent University / Engie Laborelec, Resources, Conservation and Recycling, August 2025. ↩ ↩² ↩³