EQT is the largest producer of natural gas in the United States, drilling the Marcellus shale under Pennsylvania, West Virginia and Ohio. At the October 2 close it trades at 11.0 times earnings and 1.2 times book value, within 5% of its 52-week low and about 26% off its 52-week high — while the copper and uranium miners riding the same “AI needs raw materials” story trade at 19 to 148 times earnings. EQT has a signed 10-year contract, and further agreements, to fuel natural-gas power plants built specifically for AI data centers — including what one source is calling one of the largest single-site natural gas purchases in North American history, which remains an agreement in principle. Its CEO also sold roughly $15.0 million in stock this year. Both things are true, and neither one is the whole story on its own — this file checks both properly instead of picking whichever one is more convenient.
A cheap stock and a scary headline can both be checked. Neither should be taken on faith.A genuinely cheap, well-run gas producer with a real and verifiable AI-driven demand tailwind that mostly hasn’t reached its earnings yet — priced like the market hasn’t noticed the tailwind, while the market has very much noticed (and overpaid for) the copper and uranium version of the same story. Not a safe bargain. A commodity business, honestly priced, with a real catalyst arriving on a slower clock than the excitement about it.
EQT Corporation, headquartered in Canonsburg, Pennsylvania and tracing back to 1878, is an independent producer of natural gas. It holds roughly 2.0 million gross acres, of which 1.7 million sit in the Marcellus shale — the single largest natural gas field in the country, running under Appalachia. CEO Toby Z. Rice, whose family sold Rice Energy to EQT in 2017 before he returned to run the combined company in 2019, employs about 1,523 people.
In 2024, EQT completed an all-stock acquisition of Equitrans Midstream, the pipeline company that gathers and transports much of its gas out of the basin. This is not a footnote: Appalachian gas has historically sold at a discount to the rest of the country because there has been more gas in the ground than pipeline capacity to move it out. Owning the midstream pipes alongside the wells gives EQT more control over that bottleneck than a pure producer has — the balance sheet grew accordingly (total assets roughly $25 billion in 2023 to roughly $41 billion today), and so did the debt used to fund it.
Look at five years of reported earnings per share and the business looks unstable: a loss in 2021, then four years that bounce between less than fifty cents and nearly five dollars, with no visible trend.
Operating cash flow, a cleaner read on the actual business, was steadier — though still far from flat: $1.66 billion (2021), $3.47 billion (2022), $3.18 billion (2023), $2.83 billion (2024), $5.13 billion (2025). Free cash flow, after the roughly $2.0–2.3 billion a year EQT spends just to keep production from declining, moved similarly: $607 million, $2.07 billion, $1.16 billion, $573 million, $2.84 billion. Whichever line you read, the pattern is the same — a capital-intensive commodity business whose results move with the price of the thing it sells, not a steadily compounding one.
Natural gas is burned to heat homes, generate electricity and run industry across the entire country — and its price still moves 2 to 3 times in a single winter, because usage and price stability are different things. Three mechanics explain why:
Supply responds fast. A copper mine takes a decade or more to permit and build, so supply cannot chase a price spike. A shale gas well can be drilled and turned on in months. When the price rises, US producers open the taps within a quarter or two, and the new supply pushes the price back down — economists call this the shale treadmill, and it is the single biggest reason gas producers rarely earn a stock-market-style steady growth rate.
Demand swings with the weather. A cold snap versus a mild winter is the single largest swing factor in the domestic gas price, and no one can forecast a winter a year in advance.
EQT sits on the wrong side of a pipeline bottleneck. Appalachian gas has historically sold at a real discount to the national price because there is more gas in the ground than capacity to move it to where it is needed. CEO Toby Rice made the point himself, in a Fox Business interview published the same week as this file: natural gas near $4 per unit in Appalachia could fetch close to $20 in New England, purely because of limited pipeline capacity between the two.
One number that can mislead here: EQT’s beta — how much its stock moves with the broader market — is a low 0.58. That is not the same claim as “low volatility.” Beta measures correlation with the stock market as a whole; a commodity producer’s earnings can swing hard on their own schedule, driven by weather and gas prices, largely independent of what the S&P 500 is doing that week. Low beta and real earnings volatility are not a contradiction — they are two different measurements of two different things.
Figures below are from Financial Modeling Prep, queried directly for this file — not figures travisvaluation.ca carries; the app runs on SEC filings and vendor prices, not FMP, since the licensing question in September 2026 (see the standards page). Price-based ratios are at the October 2 close; earnings ratios are trailing twelve months through the second quarter of 2026; balance-sheet items are from the June 30, 2026 quarter, with fiscal-2025 figures shown beside them where they differ.
| Method | Reading | What it says |
|---|---|---|
| Price to earnings (TTM) | 11.0× | Well under Graham’s ceiling of 15 |
| Price to book (TTM) | 1.2× | Under Graham’s ceiling of 1.5 |
| Graham Number | $52.86 (fiscal 2025) $64.32 (trailing) | Price is about 5% below the lower, fiscal-year figure and 22% below the trailing one — a real, if modest, margin of safety on the conservative reading |
| Piotroski F-Score | 7 / 9 | Solid |
| Altman Z-Score | 2.25 | Grey zone (1.81–2.99) — not distress, not pristine either |
| Return on equity | 11.7% trailing 8.6% in fiscal 2025 | Modest to fair — up from 1.1% in 2024, a loss in 2021 |
| Net debt / EBITDA | 0.9× trailing 1.3× at fiscal 2025 year-end | Manageable leverage |
| Interest coverage | 10.3× | Comfortable |
| Current ratio | 0.67 (June 30, 2026) | Fails Graham’s 2.0 — typical for the industry, still a real fail |
| Dividend yield (TTM) | 1.32% | Payout ratio only 14% — safe, not an income play |
| Analyst consensus | 30 buy / 15 hold / 0 sell | Target $70.33, about 40% above the price |
This app’s discounted-cash-flow tool returned $178.75 against a $50.17 price — an implied 256% undervaluation. That is not printed above, and should not be trusted at face value: a gap that size is a sign the model’s growth and discount-rate assumptions are miscalibrated for a volatile commodity producer coming off a strong earnings year, not evidence of a real margin of safety. The same discipline that flagged OMAB’s Graham Number as the wrong instrument for a concession business applies here in reverse — a DCF is the wrong instrument to trust blindly on a business this cyclical, and repeating the number without saying so would be worse than leaving it out.
| Test | Result | Detail |
|---|---|---|
| Adequate size | Pass | About $31.4 billion market capitalization |
| Current ratio ≥ 2 | Fail | 0.67 |
| Long-term debt < net current assets | Fail | Working capital is negative (about −$572 million); long-term debt is $5.5 billion |
| Positive earnings, recent years | Fail | Net loss of $1.14 billion in fiscal 2021 |
| Twenty years of dividends | Fail | No dividend was paid in fiscal 2021, per the cash-flow statement |
| Earnings growth over the window | Fail | Not a meaningful test on a path this erratic — a loss to $4.79 to $0.45 to $3.33 is not growth in any steady sense |
| P/E ≤ 15 | Pass | 11.0 |
| P/B ≤ 1.5 | Pass | 1.2 |
Three of eight. That is a worse score than most issues featured here, OMAB’s included — and it is an honest one. This is not a case of Graham’s tests misreading an unusual business, the way a concession or a bank can trip his balance-sheet checks. EQT genuinely carries negative working capital, a real earnings gap, and a real dividend gap, because that is what a capital-intensive commodity producer looks like after a down cycle. The cheap price and the failing scorecard are both telling the truth.
Toby Rice, EQT’s President and CEO, sold stock on the open market four times in 2026: 10,511 shares at $55.17 and 1,731 shares at $53.46 on June 5 and 8, 86,472 shares at $54.17 also in early June, and 175,328 shares at $55.03 on August 14 — roughly $15.0 million in total, all at prices above where the stock trades today. On its own, a CEO selling stock into what turned out to be the top of a range, right before a slide toward a 52-week low, is exactly the kind of thing worth being suspicious of.
So this file checked it two ways. First, the question behind the question: is there a personal reason — a divorce, a legal problem — that would explain forced selling unrelated to his view of the company? A search turned up nothing. The one lawsuit connecting Rice to EQT is from 2019, a corporate dispute over EQT’s board following its acquisition of Rice Energy, which Rice dropped that same year once the board agreed not to manipulate the vote he had launched. Nothing personal, nothing current.
Second, and more directly: the actual SEC Form 4 filing was pulled and read. Its footnote states plainly: “The transactions reported in this Form 4 were effected pursuant to a Rule 10b5-1 trading plan adopted by the reporting person on March 6, 2026.” A 10b5-1 plan is set up in advance, on a fixed schedule, specifically so an executive’s later sales cannot be read as trading on private information — this one was adopted five months before the reporting period, structured to sell shares worth up to $15.0 million in gross proceeds between June 5 and September 15, 2026. The four sales, totalled from the share counts and weighted-average prices in the June and August filings, come to about $15.0 million — the ceiling the plan set. That is a scheduled, pre-committed sale, not a discretionary bet against the stock.
A 10b5-1 plan removes the “he saw trouble coming and dumped stock” reading — the mechanics of the plan make that implausible. It does not fully remove all signal: an executive still chooses when to adopt a plan and how large to make it, and academic research on 10b5-1 sales finds they carry a smaller, but not zero, information content compared with fully discretionary sales. The honest read is that this is much closer to routine diversification of a CEO’s net worth than to an opportunistic exit, and it should be reported that way — not omitted because it is inconvenient, and not treated as more damning than the primary-source filing supports.
Global data center electricity consumption is projected to more than double by 2026, and Goldman Sachs projects US data centers will grow from roughly 5% to roughly 14% of national electricity consumption by 2030. Copper and uranium miners have already been bid up on that story — Cameco trades at 148 times earnings and 7.4 times book value, Freeport-McMoRan at 35 times and 5.2 times, Southern Copper at 30 times and 13.5 times. The fuel actually running the gas turbines that power those data centers has not been bid up the same way.
EQT has real, dated deals on this exact demand, not a story about it:
| Counterparty / project | Terms | What makes it real |
|---|---|---|
| CPV Shay Energy Center, Doddridge County, WV | 10-year deal, 325,000 Dth/day | Priced off PJM power prices, not the discounted in-basin gas price — in service as early as 2031 |
| Homer City Redevelopment, Pennsylvania | Agreement in principle (announced July 15, 2025) for a 4.4 GW natural-gas plant, up to 665,000 MMBtu/day | Described as one of the largest single-site natural gas purchases in North American history — no definitive contract found since |
| Frontier Group of Companies, Shippingport Power Station | Roughly 800 MMcf/day | A second, separate power-generation counterparty in the same basin |
The mechanism matters as much as the headline: pricing gas off the downstream power price, rather than the depressed Appalachian wellhead price, is precisely how EQT captures the value of demand it could not reach before — the same structural fix its 2024 Equitrans Midstream acquisition was aimed at. This is the real version of the “AI needs minerals” story, verified rather than assumed.
The freight is arriving before the passengers are — and here, most of it hasn’t arrived yet either.Set September 29, 2026; reviewed at the October 2 date of value and unchanged.
1. Homer City stays “in principle.” If it does not convert to a definitive, signed agreement within a reasonable window, the single largest number in section 7 was a headline, not a contract. It was announced as an agreement in principle on July 15, 2025, and EQT’s own July 21, 2026 earnings release lists only the CPV Shay deal as signed; no definitive contract has turned up in the more than fourteen months since.
2. AI capital spending slows. The demand behind all three deals assumes the current pace of data-center buildout continues. Real strain is already visible elsewhere in the AI ecosystem this week — record credit-default-swap pricing on Oracle, layoffs exceeding 20,000 at the same company, and a force majeure notice on a New Mexico data-center project. If that spreads to the counterparties in section 7, the demand could arrive later, or smaller, than priced in.
3. The gas price falls without new contract volume offsetting it. Most of EQT’s production is still sold at the prevailing spot or hedged price, not the long-term contracts above. A soft winter or renewed oversupply would hit earnings before the AI contracts scale up.
4. Liquidity actually strains. Current ratio is already 0.67 and working capital negative. That is normal for the industry, but a further deterioration — rather than the usual pattern — would say the balance sheet is the real risk, not a rounding error.
| Considered | Rejected because |
|---|---|
| Cameco (CCO.TO) | The obvious Canadian uranium-for-AI name. Trades at 148× earnings and 7.4× book — the nuclear-renaissance story is already fully priced in. |
| Teck Resources (TECK-B.TO) | A genuine Canadian copper pure-play post-coal-divestiture, Piotroski 7/9, Altman 2.73. Modestly above Graham’s price ceilings (19× / 1.7×), and mid-merger with Anglo American — real complexity not fully diligenced here. |
| Freeport-McMoRan (FCX) and Southern Copper (SCCO) | Both quality copper operators, both priced for continued high copper prices (35× and 30× earnings respectively) — not cheap by any Graham measure. |
| Citigroup (C) | A genuinely cheap money-center bank near book value, with an independent Seeking Alpha piece confirming it trades at the lowest valuation to its Graham Number among the big four US banks. Set aside on Curtis’s explicit call to diversify away from financials after Stifel and Fairfax. |
| Nutrien (NTR.TO) | Passes Graham’s price tests cleanly, Piotroski 8/9, real institutional buying from BlackRock and BNY Mellon. A strong candidate on its own merits, set aside this week in favour of the AI-minerals theme once Curtis asked to pursue it. |
At the date of value the author owns no shares of EQT.
A cheap stock and a scary headline can both be checked. Neither should be taken on faith.