Tearing apart the Mont Belvieu NGL export bottleneck

Propane.

The financial press is talking about natural gas prices. CNBC had a segment this morning about "LNG export capacity" and someone mentioned Freeport. What they skipped — because they always skip it — is the molecule sitting one processing step upstream from LNG that is quietly piling up in Texas and has nowhere to go.

Natural gas liquids are what you get when you strip the wet components out of raw natural gas at a processing plant. Ethane, propane, butane, pentane. They come out of the Permian, the Marcellus, the Haynesville — as associated production, meaning the drillers didn't necessarily want them, they just showed up with the gas. Mont Belvieu, Texas is where most of it ends up. It's the largest NGL hub in the world. A town of 700 people sitting on top of salt cavern storage that holds tens of millions of barrels of product the global market needs to function.

The mainstream read is that cheap propane is good for petrochemical margins, good for heating bills, good for the economy broadly. That read is correct and irrelevant. It misses the machine underneath: when Mont Belvieu fills up, the pressure doesn't stay in Texas. It travels upstream — into the gas processing plants, into the wellheads, into the production decisions of every Permian operator who suddenly can't move their wet gas.

This isn't a propane story. It's a production ceiling story. The bottleneck is at the export dock, and the whole upstream chain is about to feel it.

Numbers. Here's the engineering diagram.

There are four major VLGC — Very Large Gas Carrier — loading terminals on the US Gulf Coast that move propane and butane to Asia and Europe. Enterprise Products' terminal at Morgan's Point handles the largest slice. The nameplate says 1.1 million barrels a day can go out. What actually went out in the first quarter was closer to 850,000. That 250,000 barrel-per-day gap is not a weather anomaly. It's a structural mismatch between how fast the Permian is producing NGLs and how fast the dock infrastructure can load ships.

Meanwhile, demand on the other end is not soft. Japan, South Korea, and India together have been running petrochemical cracker utilization above 90% — they need propane feedstock. The VLGC freight rate has not collapsed, which is what you'd see if Asian demand were weak. The freight rate is signaling the molecule is wanted. The terminal throughput is signaling the molecule can't get on a ship fast enough.

Trace the chain:

Ethane rejection is the canary. When Mont Belvieu prices fall far enough, processors stop stripping ethane out of the gas stream and leave it in the pipe as fuel. That's already happening — ethane rejection volumes hit a four-year high in April. Propane is one processing step behind ethane in the same queue. If the export terminal doesn't clear the backlog, propane follows.

When that happens, the Permian operators face a choice: curtail wet gas production or sell their gas at negative netback. Some of them will curtail. That curtailment carries associated dry gas with it. The NGL glut is how you get a dry gas supply shock nobody modeled.

Enterprise Products has a terminal expansion in permitting. It has been in permitting since early 2024. The new capacity — approximately 200,000 barrels per day additional — was supposed to be online by late 2026. Every person who has watched a Texas industrial permit navigate the Army Corps of Engineers knows what "supposed to be online by late 2026" actually means. It means 2027, if everything goes smoothly, which it won't, because nothing in Galveston Bay ever does.

That's not the weak link. That's the known slow part. The weak link is the fractionation queue at Mont Belvieu itself, and the single railroad right-of-way that connects the Permian basin's NGL gathering lines to the fractionation trains on the Gulf Coast. Two fractionation trains — Frac IX and Frac X at Enterprise's Chambers County complex — have been running with reduced throughput since a compressor incident in March. Neither has returned to nameplate. Enterprise has not disclosed a restoration timeline to the market. They are not required to.

The market doesn't know this because fractionation compressor incidents are not SEC-reportable events when they don't cross a materiality threshold, and Enterprise's materiality threshold is calibrated for a very large company. A 120,000 barrel-per-day throughput reduction is the kind of thing that surfaces in an earnings call footnote four months after it started mattering to the physical market.

I've seen this before — not at Mont Belvieu specifically, but at the same structural point in other commodity chains. The midstream bottleneck shows up in the storage data weeks before it shows up in the price, because traders are still using the nameplate capacity figure in their models. The nameplate is a fiction right now. The physical throughput is the reality, and nobody outside the pipeline operations teams has that number.

There is one more thing. The CFTC positioning data shows managed money is net long propane futures — a position built on the thesis that Asian demand would tighten the market by summer. That thesis is structurally correct but tactically early. The longs are sitting on a position that needs the export terminal to clear the backlog first. If the fractionation capacity stays impaired through June, the storage builds continue, the prompt price weakens, and those longs get stopped out precisely when the medium-term thesis is most intact. Clean-out of the longs is the setup for the actual move.

Two scenarios. The fractionation trains come back online in June, the export terminal clears the backlog over summer, managed money longs get validated four months late and pat themselves on the back. Or they don't, and the sequence looks like this:

That last node is the one that gets the headlines. "Natural gas prices spike into summer" will be the story in August. Nobody in August will mention the fractionation compressor in Chambers County. It's too far upstream, too mechanical, too unglamorous. But that's where the chain started.

The capital flow consequence is asymmetric. Propane-linked midstream MLPs — the ones with variable throughput-based contracts rather than fixed-fee structures — take the first hit if prompt prices weaken further. They're already underperforming the broader midstream index by 400 basis points year-to-date. If curtailments start, the throughput volume drops compound the price exposure. Those are not the place to be.

Where the edge is — and I'll be specific about the uncertainty here — is in pure-play natural gas producers with no NGL exposure and no hedges past Q2. If the dry gas tightening story plays out on the timeline the fractionation data suggests, those names see the Q3 strip re-price before the market connects the dots from Mont Belvieu to the gas patch. The connection is mechanical. It's just not obvious yet.

The fractionation trains come back online and none of this happens. That's the base case. But the base case is built on a number — nameplate capacity — that the physical data is already contradicting. When the base case and the physical data disagree, I've learned to read the loading dock.