NG/LNG - This Week's Main Drivers and the Look Ahead | 9.20.26
This Week's Main Drivers:
EIA storage increased +44 BCF well below last year's +87 BCF and the five-year average of +74 BCF—another materially bullish storage signal. Yet natural gas remains unable to break and hold above $3/MMBtu.
LNG operations remain a key variable. Cameron LNG lost one train mid-week, while Golden Pass Train 1 continues its slow ramp, now around 62%. Meanwhile, the Strait of Hormuz remains severely restricted, with LNG flows still disrupted. Asian LNG demand is responding to high prices; China's 2026 demand is projected to fall by more than 6 MTPA.
ERCOT is undergoing a structural shift: solar capacity has surpassed 40 GW and now exceeds wind capacity, while the system's combined gas, coal and nuclear capacity is being increasingly challenged by the combined renewable-plus-storage fleet. Late-summer power-price volatility across ERCOT, PJM, MISO and SPP highlights the growing challenge of matching intermittent supply with 24/7 demand.
The Look Ahead:
Qatar remains a major uncertainty. Wood Mackenzie expects 54 MTPA of uncontracted Qatari LNG volumes by 2035, potentially adding substantial future supply—but today's disruption is the immediate issue. QatarEnergy has extended force-majeure cancellations into November.
Meanwhile, data centers are becoming a political "hot wire" as midterms approach. Is nuclear the answer to 24/7/365 AI demand?
The U.S. is moving that direction: DOE closed a $1.9 billion loan supporting the restart of NextEra's 615-MW Duane Arnold nuclear plant, with Google contracted for its output. Holtec has also loaded fuel at Palisades nuclear power plant, with contractual power obligations beginning March 2027.
The bigger issue: America's aging grid was not designed for today's non-seasonal, rapidly growing data-center load. Transformer shortages, long lead times and grid disturbances are becoming strategic constraints.
Energy security increasingly depends on firm power, flexible gas and a grid capable of handling AI-scale demand.
Sources:
Comments