Megatonnes by material — global ex-China, 2026–2035
$bn by material at spot prices — global ex-China, 2026–2035
Wind — vintage-dependent material intensities (t/MW installed)
Material intensity per MW varies with turbine class. The fleet reaching end-of-life in 2026–2035 was installed 2000–2015 — overwhelmingly sub-3 MW machines with higher per-MW steel and lower per-MW composite than modern 5+ MW turbines.
| Vintage | Class | Steel | Cast iron | Composite | Copper | Rare earths |
|---|---|---|---|---|---|---|
| Pre-2005 | Sub-1.5 MW | 128 t/MW | 42 t/MW | 8 t/MW | 5.0 t/MW | 0 |
| 2005–09 | 1.5–2.5 MW | 109 t/MW | 36 t/MW | 11 t/MW | 4.0 t/MW | 10 kg/MW |
| 2010–14 | 2–3 MW | 101 t/MW | 34 t/MW | 13 t/MW | 3.5 t/MW | 30 kg/MW |
| 2015+ | 3–4.5 MW | 90 t/MW | 30 t/MW | 15 t/MW | 3.0 t/MW | 50 kg/MW |
Sources: NREL REMPD (2023), Vestas LCA (V47–V136), IRENA (2017), USGS Wind Turbine Database. Steel = tower, nacelle frame, rebar. Cast iron = bedplate, gearbox housing, hub castings. Composite = blade GRP/CRP. Rare earth (NdPr) applies to PMSG drivetrains; negligible in DFIG turbines dominant pre-2012.
Solar PV — vintage-bucketed intensity (t/MW)
Silver paste intensity has fallen roughly 8× since 2005 as cell metallisation improved from BSF through PERC to TOPCon. Silicon wafer thickness has roughly halved over the same period. Bulk materials (glass, steel, aluminium, copper, polymer) are relatively stable across vintages and held flat. Three profiles are applied by deployment year:
| Deployment vintage | Technology | Silver | Silicon | Glass | Aluminium |
|---|---|---|---|---|---|
| Pre-2012 | BSF p-type, paste-heavy | 100 kg/MW | 4.5 t/MW | 45 t/MW | 20 t/MW |
| 2012–2019 | Multi-Si → PERC transition | 60 kg/MW | 3.8 t/MW | 45 t/MW | 20 t/MW |
| 2020+ | PERC mainstream / TOPCon | 25 kg/MW | 3.5 t/MW | 45 t/MW | 20 t/MW |
All vintages: Steel 30. Copper 4.5. Polymer 6 (EVA, backsheet). Silver intensities calibrated against Silver Institute World Silver Survey annual solar demand figures divided by IEA-PVPS annual installation data; pre-2012 value reflects the 2008–2011 scale-up period which dominates volumes in that bucket. At spot silver (~$2.4m/t), the difference between a pre-2012 and a post-2020 panel is approximately $180k per MW in silver recovery value alone. Sources: ITRPV 2024, Silver Institute World Silver Survey, IEA-PVPS, IRENA, Sander et al. 2019.
Battery storage — vintage-bucketed intensity (t/MW)
The fleet reaching EOL in 2026–2035 spans three distinct chemistry and duration vintages. Active materials (lithium, graphite, cathode metals) scale with energy content (duration × MW); structural and electrical materials are partly power-linked and scale less. Three profiles are applied by deployment year:
| Deployment vintage | Chemistry / duration | Li (LCE) | Graphite | Ni | Co | Steel |
|---|---|---|---|---|---|---|
| Pre-2018 | NMC-dominant, ~1.5hr avg | 0.45 t/MW | 1.5 t/MW | 0.60 t/MW | 0.20 t/MW | 14 t/MW |
| 2018–2021 | ~50/50 NMC:LFP blend, ~2hr avg | 0.60 t/MW | 2.0 t/MW | 0.40 t/MW | 0.13 t/MW | 17 t/MW |
| 2022+ | LFP-dominant, ~3hr avg | 0.90 t/MW | 3.0 t/MW | 0.20 t/MW | 0.05 t/MW | 19 t/MW |
Ni and Co intensities in the pre-2018 band reflect NMC622 cathode at pack level (~0.40 and ~0.13 kg/kWh respectively). The 2022+ LFP profile retains residual Ni/Co from BMS hardware and NMC tail-end installations. Sources: NREL ATB 2024, Argonne BatPaC v5, BNEF LCOES 2023, IEA GEO 2024.
Commodity prices (spot April 2026)
Steel HMS 1&2 $350/t (Argus). Cast iron scrap $245/t. Cu cathode LME $12,000/t. Al LME $3,500/t. Zn LME $3,300/t. Ag ~$75/oz ($2.4m/t). Si MG $2,500/t. Li₂CO₃ battery-grade $20,000/t. Graphite $600/t. NdPr oxide $65,000/t. Ni Class 1 LME $16,000/t. Co $28,000/t. Realised scrap values depend on form, contamination, lot size, and logistics.
Recovery rates
Steel 95% (shredder/EAF). Cu 90% (strip+smelt). Al 92% (remelt). Li 80% (hydromet). Graphite 50%. Ni/Co 85–90%. Rare earths 60% (magnet recycling nascent). Composite 0% (no commercial pathway at scale). Glass 90%.
Silver 65% — deliberately conservative. Silver in c-Si cells is screen-printed directly onto the silicon wafer and fired into the metallisation layer; recovering it requires acid leaching of the laminate after glass separation, a hydrometallurgical step that standard thermal-mechanical recycling does not perform. Most commercial solar recycling today recovers frames, copper wiring and bulk glass but leaves silver in crushed laminate cullet or low-grade smelter feed. Specialist operators (e.g. Veolia, FRELP) can achieve 85–90%+ at lab/pilot scale, but commercial-scale average recovery across the industry is estimated at 50–70% today. 65% is used as a mid-range practical rate. This assumption will improve materially as recycler infrastructure scales to meet the 2030s decommissioning wave and silver economics justify the capital outlay.
Scope & installation data
Global ex-China. Wind = onshore + offshore combined. Capacity reaching EOL = installed in year (Y − age threshold). Volumes are gross embedded. Concrete excluded (reused on-site or waste, not a commodity opportunity). Value chart and cumulative table show positive-value materials only for value metrics. Non-recyclable streams (composite, polymers) appear in the volume chart and mass column at zero commodity value. Installation data sources: GWEC, IRENA, IEA-PVPS, BNEF.
Important notice
This tool is provided for indicative planning purposes only. Volumes and commodity values are derived from publicly available installation data and material intensity benchmarks; they are not investment advice and do not constitute a formal commodity or recycling market forecast. Actual volumes, timing, and realisable values will vary depending on asset-specific characteristics, decommissioning rates, commodity prices, and processing pathway availability. Endenex accepts no liability for decisions made in reliance on this tool.