While precious metals have captured headlines, the most consequential constraint on artificial intelligence expansion is reliable, affordable electricity at industrial scale.
Gold has surged above $4,100 per ounce, silver has pushed past $70, and palladium has climbed back to $1,350, drawing capital from investors seeking refuge from dollar weakness.
Yet institutional money is increasingly rotating beyond traditional safe havens and toward companies that generate steady, growing cash flows by powering the data center boom.
Canadian billionaire investor Kevin O’Leary has identified this shift early, backing BitZero (NASDAQ: AIBZ) as a strategic play on the intersection of cheap energy and compute infrastructure.
Securing land and low-cost power contracts is the single most critical prerequisite for data center developers, hyperscalers, and crypto miners operating in today’s capacity-constrained environment.
Tania Tsoneva, head of infrastructure research at CBRE Investment Management, one of the world’s largest real-estate investment firms, has made the urgency plain: “The need for new capacity is very urgent — it needs to be procured now.”
Founded in 2021, BitZero has assembled a clean-energy portfolio spanning more than one gigawatt of growth capacity across four strategic sites in Norway, Finland, and North Dakota.
Its flagship hydro-powered facility in Namsskogan, Norway, delivers 40 MW of self-mining capacity at power costs below $0.05 per kWh, placing it among the lowest-cost operators globally.
According to CEO Mohammed Bakhashwain, each million dollars of capital deployed into BitZero’s grid and mining equipment generates roughly $700,000 in annual net profit.
That efficiency stems from vertical integration, with the company operating as a licensed grid operator at the 132 kV level and eliminating the middle-layer grid fees most competitors still absorb.
BitZero’s all-in energy cost of approximately 4.3 cents per kWh is less than half that of major U.S. peers, pushing its cost per Bitcoin toward $50,000 today and below $40,000 once new hardware is fully deployed.
Ultra-lean operations allow five staff to run a 40 MW facility using fully automated monitoring and fault-response systems, compressing overhead and protecting margins at scale.
The company’s newly announced 110 MW Norway project represents a potential inflection point, with a binding letter of interest projecting roughly $176 million in annual recurring revenue through long-term contracted compute capacity.
Pricing under that agreement escalates by 3% annually, with the customer covering energy costs separately, a structure that dramatically improves margin visibility and reduces exposure to power-price volatility.
BitZero believes the facility could be delivered as early as the third quarter of next year, a timeline made possible by already-secured power access, existing infrastructure, and partnerships with established EPC contractors and cooling-system providers.
That speed-to-market advantage stands in sharp contrast to the 3-to-5-year build timelines facing many competing AI data center developments due to grid bottlenecks and permitting delays.
Broader market comparisons underscore how significantly energy-advantaged digital infrastructure companies can be rewarded, with investors in TeraWulf and BitMine Immersion seeing one-year gains of more than 554% and 269%, respectively.
Investors seeking exposure to the power side of the AI boom are also looking at Vistra Corp. (NYSE: VST), whose fleet of natural gas, nuclear, and renewable assets positions it to benefit from surging data center demand.
Eaton Corporation (NYSE: ETN) supplies the electrical equipment, power management systems, and grid infrastructure required to connect and operate energy-intensive AI facilities at the scale hyperscalers now require.
GE Vernova (NYSE: GEV) has emerged as a key beneficiary of accelerating investment in power generation and grid modernization as utilities and hyperscalers compete to add new capacity at pace.
Together, these companies reflect a fundamental reorientation in how markets are pricing the AI buildout — away from semiconductors alone and toward the electricity and infrastructure that make computation physically possible.
