Opinion 2 — The Efficiency Illusion: Why Record U.S. Gas Production Masks a Looming Supply Crunch — Weekend Edition

# The Efficiency Illusion: Why Record U.S. Gas Production Masks a Looming Supply Crunch

The Efficiency Illusion: Why Record U.S. Gas Production Masks a Looming Supply Crunch The EIA announced this week that U.S. industrial natural gas consumption hit a record 23.6 billion cubic feet per day in 2025, up 1% from 2023's previous high. Markets yawned. Henry Hub futures closed Friday at $2.96, and speculators are holding a net short position of 119,870 contracts—the most bearish stance in months. The consensus view: we're drowning in gas. Total production hit 39 trillion cubic feet last year, dry gas output climbed 4%, and the U.S. remains the world's largest producer by a mile. But there's a number buried in the EIA's Short-Term Energy Outlook that tells a different story entirely. The agency forecasts industrial consumption will grow 1.2% in 2026 and 1.7% in 2027, then adds an almost throwaway line: "growth remains gradual because increases in industrial activity are partially offset by efficiency gains" that "reduce the amount of natural gas needed per unit of output." Do the math. If consumption grows 1.2% while efficiency improvements reduce gas intensity by 2-3% annually—the historical norm for industrial processes—then actual industrial *capacity* is expanding somewhere between 3.2% and 4.2% per year. That's not gradual. That's a buildout boom hiding in plain sight. The market is pricing natural gas as if we have structural oversupply. In reality, we're producing just enough to mask accelerating underlying demand. The "glut" is an accounting illusion, and when efficiency gains plateau or reverse, that hidden capacity growth will suddenly become very visible demand growth. Nobody is positioned for it. The Chemicals Subsector Is Building an Empire Industrial gas demand isn't some monolithic blob. The EIA notes that "the chemicals subsector is the largest" consumer within the industrial category. Translation: petrochemical crackers, ammonia plants, methanol facilities, and hydrogen production units. These are multi-billion-dollar capital projects with 20-30 year lifespans. Once they're built, they run. And they're being built right now at a pace the consumption numbers don't reflect. Consider what 3-4% annual capacity growth actually means on the ground. A world-scale ethane cracker consumes roughly 0.25-0.35 Bcf/d of natural gas feedstock and fuel. Just 10-12 new crackers over three years—entirely plausible given the Gulf Coast buildout—adds 3-4 Bcf/d of structural demand. But if those facilities are 15-20% more efficient than the ones they're replacing or supplementing, they might only show up as 2.4-3.2 Bcf/d in consumption statistics. The EIA reports the net increase. The market prices the net increase. But the physical infrastructure is locked in at the gross level. This matters because efficiency gains aren't linear or infinite. A new combined-cycle power plant might be 60% thermally efficient versus 33% for an old coal boiler it replaces—a one-time step change. But squeezing another 2-3% out of an already-optimized cracker? That gets harder every year. Meanwhile, the capacity keeps coming. The American Chemistry Council reported $240 billion in announced chemical industry projects as of 2024, overwhelmingly concentrated in natural gas-heavy processes. January Showed Us What Happens When Efficiency Can't Help Industrial demand hit 26.1 Bcf/d in January 2026 during Winter Storm Fern, according to preliminary EIA data. The full-year 2026 forecast is 23.9 Bcf/d, with summer troughs around 22.6 Bcf/d. That's a 3.5 Bcf/d swing—15% from peak to trough—within a single year. Here's the problem: those efficiency gains that smooth out the annual average don't help during a cold snap. An ammonia plant running at nameplate capacity in January doesn't suddenly become more efficient because it's winter. It just runs at maximum physical throughput. The seasonal variance in industrial demand is much larger than the year-over-year growth, but futures curves price the average, not the peak exposure. Total U.S. gas demand hit 165.6 Bcf/d as a seven-day average during Fern. Production averaged 110 Bcf/d in 2025, or about 106.8 Bcf/d after accounting for seasonal patterns (39 Tcf annual divided by 365 days). Even with record output, we were pulling from storage and cutting interruptible industrial loads to meet that January peak. The market treated it as a weather event. It was actually a stress test of the efficiency-masked capacity buildout. If another severe winter hits in 2026-27, that "record production" suddenly looks tight when industrial alone can spike to 26-27 Bcf/d, residential and commercial add another 50-60 Bcf/d, and power generation needs 40+ Bcf/d. We're producing 107 Bcf/d on a good day. The buffer is much thinner than the annual averages suggest. The Counter-Argument Doesn't Hold Up The bear case goes like this: shale can always produce more. Look at 2025—rig counts dropped 7% (from 415 to 386), yet we still set production records. Technology and efficiency trump everything. True, but incomplete. Those production records came from drilled-but-uncompleted wells (DUCs) and productivity gains on existing acreage. The Dallas Fed survey noted that rig cuts happened "during Thanksgiving 2025 despite oil above $90 per barrel," suggesting producer confidence is broken even at prices that should trigger expansion. More importantly, associated gas from oil drilling has been masking the Haynesville and Marcellus production slowdown. If oil drilling stays subdued—and WTI at $101 with Brent at $109 suggests it might, given the volatility—the gas market loses its invisible subsidy from associated gas. The efficiency illusion works both ways. Just as efficiency gains hide capacity growth on the demand side, they hide depletion and reduced drilling on the supply side. Permian gas production grew 4% in 2025, but how much of that was from new wells versus better completions on existing wells? The latter is a one-time gain. The former requires sustained drilling. What Happens When the Mask Comes Off The most likely catalyst is a plateau in efficiency improvements. The low-hanging fruit—replacing 1970s-era equipment, optimizing heat recovery, switching from steam crackers to catalytic processes—is mostly picked. The Gulf Coast buildout is installing state-of-the-art facilities from day one. There's no 20% efficiency gain waiting in the wings. When that happens, consumption growth and capacity growth converge. That 3-4% annual capacity expansion suddenly shows up as 3-4% demand growth, not 1-1.7%. On a base of 23.9 Bcf/d, that's an additional 0.5-0.6 Bcf/d per year that the market isn't pricing in. Over three years, that's 1.5-1.8 Bcf/d of unexpected demand—equivalent to a major LNG export terminal. The second catalyst is a cold winter coinciding with this transition. January 2026 was a preview. If industrial demand spikes to 26+ Bcf/d while the efficiency gains are slowing, and production is only growing at 2-3% annually (the EIA's current forecast), the market will reprice violently. Speculators are net short 119,870 contracts. Henry Hub is at $2.96. The market is positioned for oversupply. But the physical reality is a chemicals buildout that's expanding faster than the consumption statistics reveal, masked by efficiency gains that are quietly running out of room to improve. When that mask comes off—and it will, probably within 18-24 months—the repricing won't be gradual. It will be abrupt, painful, and entirely predictable to anyone who looked past the headline numbers. The efficiency illusion is the most dangerous kind: it's technically true, statistically valid, and fundamentally misleading. Markets that rely on it are about to learn an expensive lesson.