OilPriceAPI Research
A Hedge Fund Says the U.S. Runs Short of Natural Gas by 2028. We Rebuilt Their Math.
TL;DR: We took the viral Chronometer Partners letter (“Got Gas”, June 2026) and tried to replicate its arithmetic from public EIA data. The storage math checks out — almost exactly. Storage days-to-cover has fallen every single year since 2013, from 72.7 in 2010 to 49.1 in 2025. One number in the letter is actually conservative: EIA-measured 2025 LNG exports averaged 15.1 Bcf/d, above the 12.0 in their table. What we could NOT verify from public data: their claim that production physically maxes out at 132 Bcf/d. That’s where the debate lives.
The claim
In June 2026, Chronometer Partners LP published a letter forecasting “a shortage of natural gas for which there is no precedent” starting in the second half of 2028, with U.S. working gas storage effectively exhausted by 2030. Their arithmetic:
- Supply: ~112 Bcf/d today, rising to a maximum of 132 Bcf/d by Q4 2030 — capped by well inventory and midstream infrastructure that takes 5–8 years to build
- New demand: +20 Bcf/d of already-approved LNG export capacity (15 → 35 Bcf/d nameplate) plus +5 Bcf/d of approved gas-fired power generation, much of it for AI data centers
- The buffer: storage capacity that grew only ~7% since 2010 while demand grew 54%
Extraordinary claims deserve independent checking. Everything below is computed from the U.S. Energy Information Administration’s public API. Every series ID is listed at the bottom so you can rerun it yourself.
What we could replicate
| Metric | Letter (2010) | EIA (2010) | Letter (2025) | EIA (2025) | Verdict |
|---|---|---|---|---|---|
| Electric power demand (Bcf/d) | 20.2 | 20.2 | 36 | 35.8 | ✅ replicated |
| Residential+Commercial+Industrial (Bcf/d) | 40.4 | 40.3 | 45.5 | 46.8 | ✅ replicated |
| LNG exports (Bcf/d) | 0.1 | 0.2 | 12 | 15.1 | ⚠️ EIA higher than letter |
| Key demand total (Bcf/d) | 60.7 | 60.7 | 93.5 | 97.7 | ✅ replicated (tighter) |
| Working storage capacity (Bcf) | 4,364 | 4,410 | 4,683 | 4,793 | ✅ replicated (+8.7% vs their +7%) |
| Storage days-to-cover | 71.9 | 72.6 | 50.1 | 49.1 | ✅ replicated |
* EIA working gas capacity, latest annual figure (2024). “Days-to-cover” = working storage capacity ÷ key demand (their metric: how many days total storage capacity could supply power burn + res/com/ind + LNG exports). 2025 EIA dry production averaged 107.7 Bcf/d, consistent with the letter’s “~112 as of June 2026.”
Three observations from rebuilding this:
- The core storage arithmetic is real, not cherry-picked. Days-to-cover has declined every year since 2013 — a 16-year monotonic squeeze, not an artifact of endpoint selection.
- The letter may actually understate the demand side. EIA-measured 2025 LNG exports (15.1 Bcf/d annual average) already exceed the 12.0 Bcf/d in their exhibit. Using EIA’s numbers, key demand grew to 97.7 Bcf/d, not 93.5.
- The cushion is shrinking by design, not by accident. Storage capacity has been near-flat for 15 years because the market stopped paying for seasonal storage — while both ends of the system (production and exports/power burn) scaled dramatically.
Solid line: computed from EIA data (working storage capacity ÷ key demand). Dashed segment and 2030 point: Chronometer forecast — not EIA data.
Latest reading: 3,024 Bcf (Jul 10, 2026). Source: EIA weekly working gas in underground storage, Lower 48.
What we could NOT verify
Honesty requires drawing the line clearly. The letter’s most important claims rest on data nobody can check publicly:
- “Production maxes out at 132 Bcf/d.” This comes from well-level future-inventory economics built on proprietary data (Enverus acreage positions, well cost curves). Public data can tell you what wells have produced — we maintain 60M+ well-level production records from 19 state agencies ourselves — but not how much economic inventory remains at a given price. This is the single biggest assumption in the letter.
- “Midstream can’t be built in time.” Directionally supported by history (one major greenfield interstate pipeline completed in the last decade), but “5–8 years minimum” is a judgment, not a dataset.
- Data-center gas demand. Their load book comes from a commercial power-markets vendor. Public proxies exist — ISO interconnection queues, EIA-860M planned generators — but they’re noisy and full of speculative projects.
So the honest framing is: the runway math is verifiable and correct; the wall at the end of the runway is a proprietary-model claim. If their production ceiling is wrong, the thesis softens. If it’s right, the storage arithmetic above says the system has no buffer for error.
Worth noting: Chronometer is a fund with positions in gas producers, midstream, and uranium. That doesn’t make the letter wrong — but it’s a reason to check the math yourself, which is exactly what this page is for.
What to watch (the falsifiable version)
This thesis makes near-term testable predictions. The leading indicators, all trackable via our API:
- Weekly storage vs. the 5-year band — the letter predicts injection seasons start underperforming from 2027–2028
- The 2028+ Henry Hub futures strip — currently ~$3.50–3.70; the letter calls this complacent. If the market starts agreeing, the back of the curve moves first. Henry Hub spot and futures are available through our natural gas price API
- Waha and regional basis — Permian associated-gas bottlenecks show up in basis blowouts before they show up in national balances (we track Waha hub prices directly)
- Monthly power burn — the letter needs +5 Bcf/d by 2030; EIA data will confirm or refute the ramp quarter by quarter
- LNG feedgas/exports — already running above the letter’s table; each new terminal startup (2028–2029) is a step-change
We’ll update this page as the data comes in. Maybe the letter is right, maybe it’s early, maybe it’s wrong — but the referee is public data, and now you have it.
Methodology & sources
- Chronometer Partners letter: link to PDF
- All data: EIA API v2 (api.eia.gov). Series: dry production
N9070US2; power burnN3045US2; residentialN3010US2; commercialN3020US2; industrialN3035US2; LNG exportsN9133US2; weekly Lower-48 working gasNW2_EPG0_SWO_R48_BCF; working storage capacityNGA_EPG0_SACW0_NUS_MMCF - Annual Bcf/d figures are calendar-year averages of monthly EIA data; the letter’s table uses Q1 2010 / Q4 2025 snapshots, which accounts for small differences
- 2030 figures shown anywhere on this page are Chronometer’s forecasts, not EIA data, and are labeled as such
- This page is data journalism, not investment advice