Cognition is becoming cheap. The next binding constraint on AGI value-capture is the physical world: power, machines, fabs, and the atoms inside them. Mining is the slowest layer of that stack — permitting cycles run a decade or more, China dominates downstream processing for nearly every "critical" mineral, and capex deployed in the 2010s downcycle has left the pipeline structurally thin. When an AGI economy demands 2–10x more copper, neodymium, uranium, silver, and helium, supply curves cannot bend in time.
The trade is asymmetric: where demand-multiplier > supply-elasticity, prices have to clear higher. Equities and royalty/streaming names compound that move via operating leverage and reserve revaluation. The risks are cyclical (mining stocks are violently volatile), political (resource nationalism, export controls), and substitution (rare earths can sometimes be designed around). Reasoning from given AGI is happening, the structural call is: own the materials China can't print, the miners that already permit-cleared, and the royalty companies that own price upside without operating risk.
| Material | Spot (May 2026) | AGI-era demand multiplier (10y) | Supply elasticity (10y) | Key constraint |
|---|---|---|---|---|
| Copper | $6.43/lb ($14,176/t) | 2.0–2.5x (data centers, electrification, robots, grid) | ~1.2–1.3x | 15–25y permit cycles; Chile/Peru declining grades; Grasberg delayed to 2028; Kamoa recovery |
| Uranium | $85/lb U₃O₈ | 2.5–3x (SMRs, data-center PPAs, reactor restarts) | ~1.3x near term | Kazakh sulfuric acid; Cigar Lake/McArthur River concentration; conversion + enrichment bottleneck |
| NdPr / heavy REE | $135–140/kg NdPr ox; HREE international ~4x ex-China | 3–5x (humanoid motors, EV traction, wind, MRI) | ~1.2x ex-China | China = 85% mine, 90% magnet; MOFCOM Apr-2025 export licensing on Sm/Gd/Tb/Dy/Lu/Sc/Y |
| Silver | $77.59/oz | 2x (solar, EV, AI server contacts) + monetary tail | ~1.1x (mostly by-product) | ~70% by-product of Cu/Pb/Zn — silver supply moves with base-metal capex, not silver price |
| Tin | ~$36–37k/t (BMI 2026e $35k) | 1.8–2.2x (solder for every chip, every PCB, every robot) | ~1.0–1.1x | Wa State (Myanmar) suspension; Indonesia permit lag; thin project pipeline; smelter feed shortage |
| Helium | Crude ~$300–500/Mcf (private market) | 1.8–2.5x (advanced-node fabs, MRI, fusion R&D, leak detection) | Inelastic — by-product of LNG/natgas | Cannot be synthesized; lost to space when vented; US BLM reserve depleted; Russia Amur ramp slow |
| Gallium / Germanium | Gallium ~$580/kg; Ge ~$3,600/kg (post-controls) | 2–3x (GaN power, SiC, IR optics, satellite) | By-product (Al/Zn) — extremely inelastic | China = ~98% gallium, 60% germanium; Aug-2023 export controls; suspension only through Nov-2026 |
| Lithium | China carbonate ~$13–15k/t (rebounding from cycle low) | 3–4x (grid storage dominant, robots small) | ~2.5x (brine + DLE expandable) | Less binding than feared — DLE + brine resource is large. Mean-reversion trade, not scarcity. |
| Platinum / Iridium | Pt ~$1,400/oz; Ir ~$5,000/oz | 1.5–2.5x (hydrogen electrolyzers, fuel cells, semi catalysis) | Very low (SA grid + depleting reefs) | South African grid crisis; depleting Bushveld grades; PGM credits to nickel |
| Graphite (anode) | Spherical ~$3,500–4,500/t | 2x (batteries) — softer if humanoids run smaller packs | Synthetic graphite is elastic with petcoke feed | China = 95% spherical processing; synthetic substitution caps natural-graphite upside |
| High-purity quartz | Effectively gated by Sibelco/Quartz Corp | 2–3x (wafer crucibles for every leading-edge fab) | ~1.0x — Spruce Pine NC is essentially the world's only deposit | One deposit. One. The bottleneck of the chip bottleneck. No public pure-play. |
| Ticker | Mkt cap | Valuation note | Leverage to Cu | Key risks | Priced-in |
|---|---|---|---|---|---|
| FCX Freeport-McMoRan | $89.1B | ~14x fwd EBITDA at $4.50 Cu, single digits at $6+ spot | High — every $0.10/lb ≈ $400M EBITDA | Grasberg Indonesia gov't disputes; Section 232 cuts both ways | Partially |
| IVN.TO Ivanhoe Mines | ~CA$16B | NAV implied by Kamoa 550kt run-rate + Platreef + Western Foreland | Very high — single-asset Cu pure-play | DRC political risk; Kakula dewatering execution | No |
| SCCO Southern Copper | ~$95B est | Premium multiple; lowest-cost producer | High | Peru/Mexico permitting; Tia Maria stuck 15 yrs | Partially |
| TECK Teck Resources | ~$25B | QB2 ramp finally clearing | High | QB2 cost overrun history; Chile water | No |
| ERO Ero Copper | ~$2B | Tucumã ramp option | Very high small-cap | Single-country (Brazil); execution | No |
| ANTO.L Antofagasta | ~£24B | Centinela 2nd concentrator coming | High | Chile water; royalty regime | Partially |
| FM.TO First Quantum | ~CA$22B | Cobre Panama restart optionality (huge if it returns) | Very high | Panama remains shut; balance sheet | No |
| Ticker | Mkt cap | Valuation note | Leverage | Key risks | Priced-in |
|---|---|---|---|---|---|
| CCJ Cameco | $47.1B | $108.17 close 5/26. Westinghouse stake adds reactor optionality | High — long-term contract book at $80+/lb price points | Cigar Lake operating; Kazatomprom JV exposure | Partially |
| LEU Centrus Energy | $3.6B | $183 close, 52w range $116–464. P/E 64. Only US HALEU producer. | Extreme — winner of US enrichment story | Cap structure; gov't funding lumpiness; competition from Urenco/Orano if they re-enter US | Partially |
| NXE NexGen | ~$5B | Arrow project pre-production; permits cleared 2025 | Very high (binary on financing & build) | Construction cost; ~2028 first production | No |
| DNN Denison | ~$2.5B | Wheeler River ISR project | High | Pre-production risk | No |
| UEC Uranium Energy | ~$3B | Texas + Wyoming ISR; physical uranium inventory | High | Persistent share issuance dilution | Partially |
| SPUT Sprott Physical Uranium | ~$8B | Trades at NAV discount/premium ± 3% | 1:1 to spot | None operationally; structural short squeeze potential | No (purest beta) |
| Ticker | Mkt cap | Valuation note | Leverage | Key risks | Priced-in |
|---|---|---|---|---|---|
| MP MP Materials | $11.93B | $66.99 close 5/26 (+232% YoY). EV/EBITDA elevated; pricing for execution. | Very high — vertical integration ramp | NdFeB scaleup execution; HREE separation; pricing pressure if China lifts controls | Mostly |
| LYC.AX Lynas | ~A$8–9B | Only meaningful non-China NdPr producer at scale; Kalgoorlie cracking/leaching operational; US Texas plant in build | High | Malaysia operating consent; US plant timing | Partially |
| UUUU Energy Fuels | ~$1.5B | White Mesa REE separation + uranium | High small-cap | Sub-scale; thin margins | No |
| USAR USA Rare Earth | ~$1B | Round Top heavy REE + magnet plant build | Very high (binary) | Pre-revenue; financing | No |
| NEO.TO Neo Performance | ~CA$700M | Estonia separation + Asia magnet ops | Medium | Asian feedstock dependence | No |
| ILU.AX Iluka | ~A$5B | Eneabba refinery (Australian gov't backed) coming 2027 | Medium | Refinery commissioning | No |
| Ticker | Mkt cap (est) | Valuation note | Leverage | Key risks | Priced-in |
|---|---|---|---|---|---|
| PAAS Pan American Silver | ~$15B | Diversified Americas; benefits from $77 silver | High | Mexico royalty/mining reform | Partially |
| WPM Wheaton Precious Metals | ~$45B est | Streaming; ~60% silver, 40% gold; royalty model | Medium (no opex) | Counterparty mine performance | Partially |
| FNV Franco-Nevada | ~$45B est | Pure royalty; diversified across Ag/Au/PGM/oil | Lower beta, lower risk | Major royalty counterparty issues (Cobre Panama legacy) | Yes |
| SBSW Sibanye Stillwater | ~$8B | Pt + Pd + Au + battery metals; deep value if PGM cycle turns | Very high (operating + financial) | SA grid; balance sheet | No |
| SLV / SIVR | n/a | Physical silver ETFs | 1:1 to spot | None operational; tracking small | No (purest beta) |
| Ticker | Mkt cap (est) | Valuation note | Key constraint |
|---|---|---|---|
| AFM.V / AFMJF Alphamin Resources | ~CA$2B | Bisie tin mine DRC — top-quartile grade, ~5% Sn | DRC security (M23 conflict); single-asset |
| MLX.AX Metals X | ~A$700M | Renison tin (Tasmania) | JV partner; thin margins |
| RGLD Royal Gold | ~$13B est | Royalty; gold-tilted but Cu/Ag exposure rising | Royalty buyout pace |
| SAND Sandstorm Gold | ~$2B est | Pure royalty; small-cap leverage | Portfolio quality |
| RNGZF Renergen | ~$200M | Only listed pure-play helium | Balance sheet fragility; capex execution; single asset |
| VNP.TO 5N Plus | ~CA$1B | Germanium, tellurium specialty metals refinery | Quiet quality compounder; under-followed |
| Instrument | Exposure | Note |
|---|---|---|
| SPUT.TO Sprott Physical Uranium | U₃O₈ spot | Cleanest uranium beta. Trades NAV ± few %. Buys physical, can squeeze the spot market structurally. |
| U.UN / URNM / URA / URNJ | Uranium equities baskets | URNJ = junior tilted; URA = blue-chip tilted. |
| PSLV.TO / SLV / SIVR | Physical silver | PSLV redeemable for bars (Sprott structure); SLV is the volume leader, SIVR is lower-fee. |
| PHYS.TO / GLD / IAU | Physical gold | Hedge / monetary leg for the basket. |
| COPX | Global copper miners ETF | ~40 names; mid-cap tilted; cleaner than CPER (which trades futures). |
| REMX | Rare earths & strategic metals ETF | Heavily China-weighted; suboptimal for the "ex-China" thesis. Prefer single names. |
| LIT | Lithium & battery tech | Half tech half miners; use for cycle bottom, not stock-picking. |
| SIL / SILJ | Silver miners ETFs | SILJ = junior, more leverage. |
| GDX / GDXJ | Gold miners | Royalty companies (FNV, WPM, RGLD) sit inside these. |
| LME copper / CME HG futures | Direct copper | Levered; for sized macro positioning only. |
| Royalty / streaming (FNV, WPM, RGLD, SAND, TFPM) | Diversified commodity inflation | Lower-risk way to play the cycle. They own price upside without operating downside. |
Permitting talent in the US, Canada, and Australia is the bottleneck-of-the-bottleneck. There are perhaps a few hundred people globally who can shepherd a Tier-1 mine through 10-year permitting; their employer's projects are worth more accordingly.
The only credible Western vertical in NdPr → NdFeB magnets, and the magnet is the most AGI-leveraged single component (every motor in every humanoid, every server fan, every EV drive). MP is mining + separation + (now) commercial magnets, with a $1.74B cash cushion, DoD/Apple offtake, and HREE separation imminent. At $11.9B mkt cap it's already priced for execution — so the asymmetry is narrower than 12 months ago. But the geopolitical moat (MOFCOM controls + 30–60% domestic premium) is widening, not narrowing. Risk: a Chinese policy reversal would compress prices; an execution stumble at Fort Worth would punish multiple.
Two instruments, one trade. CCJ for the contract-book operating leverage (long-term contracts re-pricing at $80+) plus the Westinghouse stake (reactor-build optionality). SPUT for clean physical-uranium beta with structural squeeze potential as utilities re-contract. AGI-era electricity demand is monotonic up; uranium is the only baseload molecule that scales with permits already filed (and SMRs filling the gaps). 60/40 CCJ/SPUT.
Highest-grade major copper asset in the world (Kamoa-Kakula 3.5–4% Cu vs industry ~0.5%), 380–420 kt 2026 guide → 550 kt run-rate, plus Platreef (PGM/Ni) and Western Foreland exploration. DRC risk is real but priced in. At CA$16B, you're paying ~10x EBITDA at $4 Cu, ~5x at $6 spot. Friedland discount/premium washes out across cycle. The cleanest single-name way to own structural copper undersupply.
Royalty/streaming model is the lowest-risk way to compound through commodity cycles. WPM is ~60% silver — directly leveraged to solar + AI server contact growth — and has no opex. FNV adds gold, oil, and broader diversification. Both already priced rich-ish, but in a basket they earn 8–12% real CAGR through cycles with much lower vol than miners.
Only US HALEU producer at a moment when every SMR vendor (Oklo, X-Energy, TerraPower) needs HALEU and the US gov't has made it strategic. Stock has run hard (52w $116–$464, now $183), P/E 64. Truthful read: the equity already prices in success, the multiple is fragile, and Urenco/Orano could re-enter US enrichment and compress moat. But if the US enrichment buildout is real, LEU is the only public listed beneficiary at scale. Position-size, don't concentrate.
Compiled 2026-05-26. Prices and market caps captured May 22–27, 2026.
Track 4 of agi-investment-tracks · materials-mining analyst · for Ravi