Amazon digs deep for AI gold as copper mine partnership fuels massive data center expansion

AWS has signed an agreement with Rio Tinto’s Nuton venture linked to the Johnson Camp copper mine in Arizona, securing approximately 14,000 metric tons of copper cathode over four years (via a two-year contractual arrangement) to support its expanding AI data center buildout. Amazon will direct the copper to component manufacturers for AWS data centers and provide cloud analytics to help optimize Nuton bioleaching production, emphasizing lower-carbon and lower-water-intensity sourcing; average data centers use roughly 5,000–15,000 tons of copper and benchmark prices recently neared $6 per pound. The deal highlights tightening copper demand from AI infrastructure and Amazon’s supply-chain focus on low-carbon inputs, with modest but relevant implications for copper markets and suppliers.

Analysis

Market structure: Amazon’s deal materially benefits Rio Tinto (RIO) and vertically integrated, low-carbon copper producers while increasing pricing leverage for high‑quality miners with stable output. AWS and large hyperscalers become de‑facto demand anchors reducing short‑term price volatility for contracted volumes, but incremental spot demand from accelerated AI data‑center builds will keep upward pressure on copper prices (sustained upside if prices move above $5.50/lb for >3 months).

Risk assessment: Tail risks include a tech capex pause (AI chill) that collapses copper demand, a bioleaching operational failure at Johnson Camp, or US regulatory constraints on mining/water that delay production; probability low–medium but impact high. Immediate (days) moves will be modest; 3–12 month horizon is key for copper price discovery; structural tightness plays out over 1–3 years as data‑center rollouts continue.

Trade implications: Favor quality miner exposure (RIO) and selective data‑center equities (EQIX/DLR) vs broad copper‑junior exposure; implement options to cap downside while keeping upside (9–18 month call spreads on copper/miners). Pair trades (long RIO, short COPX or FCX) capture Rio’s cleaner‑production premium and balance‑sheet advantage.

Contrarian angle: The headline volumes (14k t over 4 years) are modest relative to a single hyperscale project, so the market may overstate near‑term demand impact—risk of mean reversion if AI capex normalizes. Conversely, investors underprice regulatory/ESG value for low‑carbon copper which could re‑rate RIO and low‑carbon producers over 12–36 months.

Key Decisions for Investors

• Establish a 2–3% long position in RIO (Rio Tinto) with a 6–12 month horizon to capture low‑carbon copper premium; set a stop‑loss at 12% downside or if LME copper closes below $3.50/lb on a 30‑day SMA.
• Add 1–2% long in AMZN to play AWS capex (or buy a 9–12 month call spread ~10–15% OTM to limit capital) and exit/reevaluate if AWS capex guidance is cut by >15% on next quarterly report (~next 90 days).
• Trade copper via a 9–18 month call spread on copper futures: buy Dec 2026 $4.75 calls, sell $6.25 calls sized to equal 1–2% portfolio risk; trim if copper falls and closes below $4.00/lb weekly.
• Pair trade: long RIO 1.5% vs short COPX (Global X Copper Miners ETF) 1.5% to capture quality premium; unwind after 6–12 months or if RIO underperforms COPX by >10% (reassess on production/cost disclosures).