NVIDIA CorporationHigher natural gas prices could increase AI compute energy costs from 10% to 20-30%, threatening hyperscaler profit margins and potentially reducing demand for NVIDIA's chips.

The U.S. could face a structural natural gas shortage beginning in 2028 as surging AI data center demand collides with growing LNG exports, according to Chronometer Partners CIO Matthew Smith. U.S. natural gas production is projected to rise about 20 billion cubic feet per day to roughly 132 Bcf/day by 2030, but LNG export commitments alone are expected to consume much of that increase, leaving little room for new AI-driven electricity needs. Smith warns that energy costs, currently about 10% of AI compute expenses, could climb to 20% or even 30% if natural gas prices double or triple, threatening hyperscaler profit margins. The analysis suggests natural gas producers, nuclear generators, and electricity infrastructure companies may become as strategically important as chipmakers in the next AI investment cycle.
NVIDIA CorporationHigher natural gas prices could increase AI compute energy costs from 10% to 20-30%, threatening hyperscaler profit margins and potentially reducing demand for NVIDIA's chips.
Advanced Micro Devices Inc