AI demand collided with physical limits. The next moat is control of land, power, water, and grid access. Firms with locked-in megawatts hold a stronger position than companies shipping GPUs. Cheap, firm power sets the pace of the entire buildout. The players who already secured those electrons now sit on assets that matter more than their core businesses. A surprising mix of legacy industries hold this advantage. Aluminum smelters with subsidized hydropower. Crypto miners with long-term PPAs in Texas. Hydrogen producers with public power allocations. Old industrial campuses with behind-the-meter generation. These firms run weak or volatile businesses, but they hold premium power rights. For many, the returns from selling or repurposing those megawatts for AI data centers beat the economics of continuing the legacy product lines. More candidates sit in the shadows. Pulp and paper mills with large biomass or gas co-gen units. Fertilizer and chemical plants with long-term Gulf Coast power contracts. Silicon metal and ferroalloy producers with dedicated hydro or gas supply. Older steel operations with captive power. These sites consume huge amounts of energy and often struggle to earn their cost of capital. Their power access is the real prize. As AI buildout pressure grows, these overlooked operators may find their grid connections valued far above their underlying businesses. When do we start buying #IP, #AA, #CF, #LYB, #DOW, etc. as AI plays because its more profitable to sell your power to an AI Hyperscaler than to make your core product?