Back to News
Market Impact: 0.45

The Next Phase of the AI Boom May Not Come From Chipmakers

CEGMSFTGOOGLGOOGNVDATSMNEECCJNFLX
Artificial IntelligenceEnergy Markets & PricesRenewable Energy TransitionCommodities & Raw MaterialsTechnology & InnovationESG & Climate Policy
The Next Phase of the AI Boom May Not Come From Chipmakers

Rising AI-driven electricity demand—data centers already consume 1.5% of global power and have grown ~12% annually with IEA projecting up to 30% annual growth—has prompted major tech firms to contract nuclear capacity, boosting nuclear equities (Nuclear Energy Index +88% YoY). Notable commercial deals include Microsoft’s $3 billion, 20-year power purchase and DOE $1 billion loan to Constellation Energy to restart Three Mile Island (target online 2027–2028), Alphabet’s $1.6 billion, 25-year arrangement with NextEra to revive Duane Arnold by 2029, and sector supply plays such as Cameco (17% of 2024 uranium supply, 49% stake in Westinghouse), alongside a reported U.S. government commitment of $80 billion for AP1000 reactors—developments that materially reallocate long-duration baseload power exposure and favor nuclear-capex and uranium-exposure equities.

Analysis

Market structure: Nuclear generators (CEG, NEE) and uranium miners (CCJ) are direct beneficiaries as big-tech PPAs (Microsoft, Alphabet) lock in long-duration baseload demand; expect improved utilization and pricing power for baseload assets with multi-decade revenue visibility, shifting incremental capex away from intermittent renewables toward dispatchable capacity. Uranium supply concentration (Kazakhstan ~19% + majors like Cameco ~17%) suggests commodity tightness could materialize if demand accelerates at IEA’s projected ~30% p.a.; price spikes would flow straight to miner cash flows and nuclear EPC margins. Risk assessment: Tail risks include regulatory reversals (new safety moratoria) and multi-year project delays/cost overruns—Three Mile Island restart slips to >2028 would re-rate forward multiples negatively; geopolitical disruption to Kazakh supply is a high-impact tail that could push spot uranium >+50% within 12 months. Short-term catalysts (DOE loan tranches, PPA announcements) matter within 3–12 months; long-term execution (SMR licensing, grid upgrades) plays out over 3–7 years and determines ultimate ROI. Trade implications: Favor selective long exposure: CCJ for commodity + Westinghouse linkage; CEG/NEE for contracted cashflows — size tranches to 1–4% per name and hedge with options around catalysts. Use call spreads or LEAPS on CCJ ahead of DOE/PPA milestones (6–12 months) and consider long CEG vs short SO pair to express nuclear vs legacy thermal displacement; take profits at +30–50% or on missed milestones beyond stated deadlines. Contrarian angles: Market may be underpricing execution risk and timeline friction — nuclear equities up ~88% Y/Y could be partially priced for ideal timelines; historical parallel: post-2007 nuclear rallies collapsed after Fukushima-style shocks and regulatory tightening. Unintended consequences include accelerated secondary supply/resumption of idled enrichment if uranium spikes, capping long-run miner upside; therefore size positions and hedge event risk aggressively.