← Back to Blog
|9 min read|finance

The Copper Investment Thesis: AI Data Centers, Supply Deficits, and Why Copper May Be the Trade of the Decade

An AI-generated research report analyzing the copper supply-demand imbalance driven by AI data center buildouts, the energy transition, and structural mine supply constraints. Includes comparison to gold and silver, on-chain investment vehicles, fee modeling, and a go/no-go verdict.


Disclosure: This is an automated research report generated by Claude (Anthropic) on February 11, 2026. It was commissioned by Optimal as part of an investment thesis evaluation. Data was gathered by parallel AI research agents querying live sources. Nothing in this report constitutes financial advice. All claims are sourced — reference numbers link to the full source list at the bottom.


Executive Summary

Copper is entering what multiple major institutions are calling a structural supply crisis. AI data center buildouts, electrification, EV adoption, and defense spending are creating a multi-front demand surge — against a supply base that takes 17.9 years on average to expand [1].

Verdict: GO — but copper complements gold, it doesn’t replace it. Copper offers the best risk/reward of the three metals for growth-oriented allocation from today’s prices.

The Thesis in Three Numbers

ValueWhy It Matters
17.9 yearsAvg. mine discovery → production [1]No supply response until ~2044
475,000 mt2026 data center copper demand [2]4.3x increase from 2025
10M tonnesCumulative shortfall by 2035-2040 [3]25% of projected demand

Part 1: The Demand Case

Metal Prices — Current Snapshot (Feb 2026)

MetalPriceRun from 2022 LowPrice / Cost of Production
🟠 Copper$5.93/lb [4]+85%1.5-1.7x
🟡 Gold$5,060/oz [5]+212%~3x
Silver~$84/oz [6]+370%4-5x

Copper is the least extended of all three metals. Silver is the most overextended — it dropped 32% in days in late January [7].

AI Data Centers: The Headline Driver

A traditional data center uses 10-15 tonnes of copper per MW. An AI data center (GPU-dense) uses 25-33 tonnes per MW — roughly 2-3x more [8].

MetricNumber
Hyperscaler capex (2026)$600B+ (36% YoY increase) [9]
Data center copper demand (2025)~110,000 mt
Data center copper demand (2026)~475,000 mt (4.3x jump) [2]
Meta’s data center target10+ GW by end of 2026 [10]

Critically, data center copper demand is price-inelastic — developers buy regardless of price [10]. This is fundamentally different from construction demand.

Full Demand Breakdown (2026 → 2030+)

Demand DriverCopper Intensity2026 Est.2030+ Projection
AI Data Centers25-33 t/MW [8]475K mt [2]500K-1.1M mt [11]
Electric Vehicles53-83 kg/vehicle [10]~2M mt~2.5M mt
Solar PV4-5 t/MW [11]Growing 8.9% CAGR+4.22M mt by 2035
Grid Infrastructure25% of global demand [12]Expanding61% by 2040 [12]
Defense / MilitaryNATO 4% GDP target [13]+170K mtGrowing

S&P Global projects a 10 million tonne cumulative shortfall by 2035-2040 [3] — representing 25% of projected demand.


Part 2: The Supply Problem

Why There’s No Quick Fix

MetricTimelineSource
Global avg: discovery → production17.9 years[1]
US avg: discovery → production31.8 years[1]
US permitting alone7-10+ years[14]
Only US greenfield mine under constructionFlorence Copper[14]

Any copper mine started today would not produce until approximately 2044.

Active Supply Disruptions (2025-2026)

MineWhat HappenedImpact
Grasberg (Indonesia)Mudslide70% of forecast lost [15]
Quebrada Blanca (Chile)Cut from 310K → 235K tonnes [16]-75K mt
Kamoa-Kakula (DRC)Extended disruptions throughout 2025Ongoing
ICSG revision2025 mine growth: 2.3% → 1.4% [17]Structural

Deficit Trajectory

YearProjected DeficitSource
2026150K - 330K mtICSG / J.P. Morgan [18]
2027+350K mt additionalCiti [16]
2029~555K mtBloombergNEF
2030~766K mtBloombergNEF
2035-204010M mt cumulativeS&P Global [3]

Part 3: Copper vs Gold vs Silver

How Extended Is Each Metal?

CopperGoldSilver
2022 Low~$3.20/lb~$1,620/oz~$18/oz
Current$5.93/lb$5,060/oz~$84/oz
Run-up+85%+212%+370%
Production cost ratio1.5-1.7x~3x4-5x
Analyst consensusNear/above targetsAbove bull targetsExtreme risk

Gold’s Problem from Here

Gold at $5,060 has already exceeded both J.P. Morgan’s Q4 2026 target of $5,000 [19] and Goldman’s December 2026 target of $4,900 [20]. When the price runs ahead of even the bull-case targets, that’s a yellow flag.

Gold remains the best crisis hedge — it returned ~9% real annually during 1970s stagflation [21] and outperformed the S&P in 6 of 8 recessions since 1973 [21]. But from $5,060, you’re buying momentum, not value.

Silver’s Problem from Here

Silver at $84/oz and 4-5x production cost carries the highest risk. The gold-to-silver ratio compressed from 100:1 to ~60:1 — roughly fair value historically. The easy trade is over. Silver fell from $50 to $14 between 2011-2015 (72% decline) [22]. The recent $117→$80 crash shows this dynamic is alive.

Macro Scenario Scorecard

ScenarioProb.CopperGoldSilver
AI Boom / Growth30-35%BestFlatOK
Inflation + Growth20-25%GoodGoodBest
Recession20-30%❌ Worst (-30-50%)Best (+10-25%)Bad
Stagflation10-15%BadBest (+25-50%) [21]Mixed

Takeaway: If AI-driven growth is the dominant macro path, copper wins. If recession risk is high, gold wins. The answer is both, not either/or.


Part 4: On-Chain Copper — What Exists Today

The idea of a 1:1 physically-backed copper token with staking yield is compelling in theory. In practice, the space is 3-5 years behind gold tokenization.

Available Products

ProductTypeLiquidityVerdict
TCu29Physical token (1.2B lbs reserves)~$18K pools [23]Real but unusable at scale
Ondo COPXonCOPX ETF tokenized~$800/day [24]Mining equity, not commodity
CPER ETFCopper futures~$200M AUMBest pure copper exposure
COPX ETF41 mining stocks$7.37B AUM [25]2.18% dividend, leveraged play

Why Copper Tokenization Is Hard

$10,000 of copper weighs approximately one ton. Compare that to gold, where $10,000 weighs ~0.2 ounces. Storage costs run 2.5-4% annually for copper vs 0.1-0.5% for gold [26] — a 10-40x disadvantage.

There are no DeFi lending markets, no options vaults, and no meaningful liquidity pools for copper tokens. The CFTC’s digital assets pilot [27] and emerging two-lane regulatory framework [28] clear the regulatory path, but infrastructure isn’t there yet.

Best Implementation Today

COPX (copper miners ETF) + covered call overlay. With implied volatility at the 98th percentile (~56% IV) [29], covered call premiums can generate 5-10% annualized income on top of price appreciation. The newly launched Global X CPCC ETF (Dec 2025, Canada) is the world’s first copper covered-call ETF.


Part 5: Investment Bank Price Targets

Institution2026 Avg Target2027Long-term
Goldman Sachs [20]~$5.17/lb~$4.88/lb$6.80/lb by 2035
J.P. Morgan [2]~$5.48/lb (Q2 peak ~$5.67)
Bank of America [16]~$5.13/lb~$6.13/lb
World Bank [30]~$4.45/lb~$4.54/lb

Copper at $5.93/lb is currently above most 2026 average targets. Either analysts are behind the curve, or the market has front-run the deficit.


Part 6: The Tariff Wildcard

In July 2025, the Trump administration announced a 50% tariff on copper imports under Section 232 [31]. Refined copper was exempted — the tariff applies to semi-finished products (pipes, wires, rods).

The impact was immediate: ~500,000 tonnes flooded US ports in one month. The COMEX premium surged to 30%+ over LME prices [11].

By June 30, 2026, the Commerce Department must report on domestic copper markets. Goldman Sachs estimates a ~55% probability of phased refined copper tariffs (15% Jan 2027, rising to 30% Jan 2028) [32].


Part 7: What Could Kill This Thesis

RiskSeverityHistorical Precedent
Global recession☠️ CriticalCopper fell 69% in 2008, 45% in 2020 [33]
China slowdown☠️ CriticalChina = 50-58% of global demand [2]
Goldman surplus scenario⚠️ MediumIf tariffs weigh on growth, deficit may not materialize [32]
Substitution⚠️ Low-MediumAluminum and fiber optics in some applications [34]
Price already extended⚠️ Medium+85% from 2022 low — AI narrative partly priced in

Conclusion: Recommended Allocation

Portfolio Split

MetalWeightRole
🟡 Gold50-60%Core strategic hedge. Central bank buying [35] provides structural floor. Best for recession/stagflation.
🟠 Copper30-40%Growth thesis. AI/EV/grid demand real, supply constrained for a decade. Best forward risk/reward.
Silver0-10%Avoid at current levels. 4-5x production cost [36], extreme volatility.

Time Horizon

PhaseTimelineCatalystCopper Target
Current2026Grasberg shutdown, 150-330K mt deficit$5.93 (today)
AccelerationH2 2026-2027AI buildout, tariff clarity$6.50-$8.00/lb
Structural2028-2030No new mines, deficit widens$8.00-$12.00/lb
Peak stress2030+Energy transition demand triples [12]$10.00+/lb

The structural thesis is sound. If you have a 2-3 year horizon and can stomach a recession-driven drawdown, copper offers asymmetric upside that gold and silver cannot match from current levels.


Sources & References

<a id="ref-1"></a>[1] Mining Visuals — Copper Mines Average Time from Discovery to Production is 17.9 Years

<a id="ref-2"></a>[2] J.P. Morgan — 2026 Copper Outlook: Supply Deficits and Price Forecasts

<a id="ref-3"></a>[3] S&P Global / PR Newswire — Substantial Shortfall in Copper Supply Widens

<a id="ref-4"></a>[4] Trading Economics — Copper Price

<a id="ref-5"></a>[5] Trading Economics — Gold Price

<a id="ref-6"></a>[6] Fortune — Current Price of Silver (Feb 2026)

<a id="ref-7"></a>[7] Yahoo Finance — Copper Surges in Unsustainable Rally, Joining Silver and Gold

<a id="ref-8"></a>[8] Tom’s Hardware — Why Copper Markets Are Feeling the Pinch

<a id="ref-9"></a>[9] IEEE ComSoc Tech Blog — Hyperscaler Capex $600B in 2026

<a id="ref-10"></a>[10] BHP Insights — Why AI Tools and Data Centres Are Driving Copper Demand

<a id="ref-11"></a>[11] Sprott — Copper: Wired for the Future

<a id="ref-12"></a>[12] Mining.com — AI to Boost Copper Demand 50% by 2040: S&P

<a id="ref-13"></a>[13] Modern War Institute — As America’s Military Rearms, It Needs Minerals

<a id="ref-14"></a>[14] Copper.org — Can US Copper Meet Surging Demand from AI Data Centers?

<a id="ref-15"></a>[15] AG Metal Miner — Copper Market Deficit: Grasberg

<a id="ref-16"></a>[16] Mining.com — Copper’s Tight Supply and Tariff Risks Set for a Volatile 2026

<a id="ref-17"></a>[17] Mining Weekly — Copper Market Expected to Be in Surplus in 2025 Followed by Deficit in 2026

<a id="ref-18"></a>[18] The Oregon Group — ICSG Warns of 150,000t Deficit in 2026

<a id="ref-19"></a>[19] J.P. Morgan — Gold Prices Outlook

<a id="ref-20"></a>[20] Goldman Sachs — Copper Prices Forecast to Decline from Record Highs in 2026

<a id="ref-21"></a>[21] CME Group — How Does Gold Perform with Inflation, Stagflation, and Recession

<a id="ref-22"></a>[22] Investing News — Copper Price History

<a id="ref-23"></a>[23] Phemex — What is TCu29?

<a id="ref-24"></a>[24] CoinGecko — Ondo Tokenized COPX ETF

<a id="ref-25"></a>[25] ETF Database — COPX Overview

<a id="ref-26"></a>[26] LME — Warehouse Charges

<a id="ref-27"></a>[27] CFTC — Digital Assets Pilot Program

<a id="ref-28"></a>[28] Regulatory & Compliance — Two-Lane Highway for US Crypto Regulation

<a id="ref-29"></a>[29] Barchart — Copper Futures Volatility

<a id="ref-30"></a>[30] World Bank — Metal Prices Poised to Strengthen Further

<a id="ref-31"></a>[31] CNN — Trump Copper Tariff

<a id="ref-32"></a>[32] Goldman Sachs — Why Record High Copper Prices Aren’t Forecast to Last

<a id="ref-33"></a>[33] JE Gross — Copper Bull and Bear Markets

<a id="ref-34"></a>[34] Fortune — Copper Shortage Outlook: Systemic Risk

<a id="ref-35"></a>[35] World Gold Council — Gold Outlook 2026

<a id="ref-36"></a>[36] Sprott — Gold & Silver Outlook 2026


This report was generated on February 11, 2026 using parallel AI research agents (Claude, Anthropic). Sources include Goldman Sachs, J.P. Morgan, Bank of America, World Bank, S&P Global, BloombergNEF, CME Group, and ICSG. This is not financial advice.

Automated report produced for Optimal | Finance Category