Top Data Center Stocks: BofA Highlights Key Materials Plays

Bank of America identified 67 stocks with a combined market cap of about $5.5 trillion that could benefit from AI data center build-out demand. The highlighted names are mainly materials and infrastructure suppliers such as copper, aluminum, uranium, lithium, and steel producers, including Jiangxi Copper, Norsk Hydro, Kazatomprom, Ganfeng Lithium, and Antofagasta. The piece is broadly constructive for these suppliers and related data-center infrastructure names, but it is mainly a thematic screen rather than a company-specific catalyst.

Analysis

The market is still underappreciating that the AI build-out is becoming a utilities-and-metals trade, not just a semicap capex trade. The first-order winners here are copper, aluminum, and uranium-linked names, but the second-order beneficiaries are grid equipment makers, EPC contractors, and water-treatment suppliers that sit one layer deeper in the capex stack and should see order books re-rate before headline data-center starts do. That matters because the bottleneck is increasingly not GPUs, but power delivery, interconnects, and permitting, which tend to extend the revenue tail from months into multiple years.

For NVDA, this is indirectly positive: every incremental data-center node requires more non-chip infrastructure spend, which signals that hyperscaler capex remains durable rather than peaking. The risk is that investors confuse sustained build-out with immediate semiconductor upside; if power constraints delay commissioning, near-term GPU demand can lag the rhetoric even as long-cycle demand improves. The cleaner beneficiaries may therefore be the industrials and materials names with visible backlog conversion, while semis benefit later and more selectively.

The contrarian point is that the trade is probably crowded in the most obvious resource names, while the real upside may sit in under-owned proxies with operating leverage to electrification and grid capex. However, these stocks are also highly sensitive to China growth, FX, and commodity price retracement, so any deceleration in AI capex or a copper pullback could hit them faster than the broader market expects. The best setup is to use pullbacks after a strong AI-capex headline cycle, not chase extended moves intraday.

Near term, the catalyst window is 1-6 months as hyperscaler budgets, transmission orders, and utility interconnect filings filter through; over 1-3 years, this theme remains structurally bullish as power becomes the scarce input. The main reversal risk is a cooling in AI spending or regulatory delays that push out data-center starts, which would compress the forward multiple on the supply chain before end-demand is visibly impaired.

Key Decisions for Investors

• Long LUN.TO vs short a basket of over-owned AI semis for 3-6 months: copper exposure offers cleaner leverage to the infrastructure bottleneck, while semis face execution and timing risk if power interconnects delay deployments.
• Add STLD on weakness for a 6-12 month trade: steel demand tied to data-center construction is less crowded than copper, with better downside protection if commodity prices mean-revert because volumes can offset price volatility.
• Hold/accumulate CCJ on a 6-18 month horizon: uranium is a later-cycle winner if AI load growth forces more firm power, but expect higher volatility and policy-driven gaps; use dips rather than momentum entries.
• Use BAC as a relative-value long only if you want the theme without commodity beta: the bank’s exposure is cleaner through financing/advisory and is less dependent on spot metal prices, but upside is lower than the miners.
• Do not chase NVDA here; instead buy NVDA on weakness only if you see simultaneous strength in power-infrastructure names, which would confirm the capex cycle is still broadening rather than simply rotating.