Apollo 1o
Based upon extrapolation of historical demand trends and linear econometric relationships:
In 2024, total U.S. demand was 4,347 TWh¹ [1] U.S. Energy Information Administration.
For 2025, demand is estimated to be 4,419 TWh² 1.647% increase from '24 [2] EIA
For 2026, demand is predicted to be 4,492 TWh³ 1.645% increase from '25 [3] EIA
Legacy forecasts do not account for personal and commercial EVs...
GM — zero-emissions lineup by 2035 + 50% EV sales by 2030.
Hyundai Genesis — 100% electrification by 2040.
Volkswagen — 55% sales electric by 2030.
Ford — 2 million EVs annually by 2030.
Amazon — 100K Rivian delivery vans.
FedEx — zero-emissions by 2040.
UPS — partnership with Arrival.
Uber — electric fleet by 2030.
Lyft — fully electric fleet by 2030.
Tesla — autonomous robotaxis, Q3 2025.
Nor do they account for AI computational infrastructure...
Microsoft — $50+ billion 2025 CapEx for AI infrastructure.
Alphabet — $12.4 billion Q1 2025 alone, $50+ billion annually.
Amazon — $148 billion through 2040 for data centers, majority US.
Meta — 600,000 H100 GPU equivalent capacity, 350K H100s in 2024.
Let's see what utility providers & grid operators are saying about the numbers.
MISO, PJM, CAISO fast-track processes for data center interconnection requests.
Utilities restructuring planning process around data center load multi-billion $ expansions.
Growth expectations doubled between 2022-2024 reports of 15%+ CAGR at substation level.
2,500+ GW of new generation & storage awaiting interconnection backlogs & processing capacity.
How should you feel about all of this?
EIA's forecasted growth of 50.4% through 2050 did they cut their research budget?
Best case scenario demand 2x by 2050 (all good, 2x capacity with storage alone)
More realistic case demand 3x to 4x by 2050 (don't worry, solar+BESS > BESS)