Damage Math

Damage Math

The drawdown erased roughly $9.0B of market value peak-to-current (−21.5%), and $11.8B (−28.1%) at the July trough. A conservative, probability-weighted DCF-lite puts the cash value actually destroyed at about $3–4.4B. The remaining ~$5–6B gap is real, but it rests entirely on the trial's 0.71 probability that the hit is temporary; at the 29% permanent tail — a 32% fall in reserves per share from dilution — the gap closes.

The whole tab is one piece of arithmetic: a numerator (how much earning power fell), a denominator (how much price fell), and the difference between them under two transparent scenarios. The diagnosis that picks between the scenarios is not mine — it is the trial's, ruled at a 0.71 probability the impairment is temporary.

The near-term hit — the numerator

The trigger was a commodity input, not a company event: Appalachian spot gas fell to $2.89/MMBtu through the Q2-2026 shoulder season. That flowed straight into the print. Second-quarter net income fell to $211 million, $0.34 per diluted share, from $784 million, $1.30 per diluted share a year earlier — the filing attributes the drop to "lower derivative gains and lower average realized natural gas prices" [1].

Consensus followed the print down — but only at the front. Over the 90 days into the trough, current-year (FY2026) normalized EPS was cut 14.4%, from $4.8543 to $4.1542; next-year (FY2027) fell 12.4%, from $5.0013 to $4.383 [2]. On the CapIQ vintage the FY2027 cut is deeper still, from $4.77 (as of 2026-02-02, near the price peak) to $3.98 now, −16.5%. FY2028, by contrast, barely moved — $5.19 to $5.16.

Source: consensus EPS-trend vintages, current vs 90 days ago [3]; FY2028 on CapIQ momentum series [4].

Two facts complicate the numerator, and both belong here. First, the cut is a front-end shave, not a level reset: even after the 14.4% cut, FY2026 EPS still sits 36% above the FY2025 actual of $3.05 [5]. Second, consensus free cash flow did not fall at all — it keeps rising across the forecast, $3.12B (2026) to $3.83B (2028), a 9.5%-to-11.7% yield on today's market cap [6]. And the company moved the other way from a company in distress: mid-drawdown it raised 2026 production guidance "by roughly 90 Bcfe at the midpoint" and lowered full-year capital [7].

Source: consensus forward FCF, converted to yield on the current $32.8B market cap [8]; yield derivation per fit_features.consensus_forward_yield.

So the numerator is genuinely two-sided. A single trough quarter and a front-loaded EPS cut are the marks of a short-term earnings effect; a flat-to-rising forward FCF strip and raised volume guidance say the earning power itself was not marked down. That tension is what the trial exists to resolve.

The price change — the denominator

The market did not shave the front end; it repriced the whole claim. From the 2026-03-25 peak of $67.93 to the 2026-07-31 close of $53.29 (prices per the daily feed; fit_features.capitulation_gauge), market capitalization fell $9.0B on a 615.7M diluted-share base [9] — −21.5%. At the 2026-07-10 trough of $48.85 the loss was $11.8B, −28.1%.

Enterprise value fell about as much, and not because debt grew — because it shrank. Net debt came down from $7.69B at December 31, 2025 to $5.54B at June 30, 2026, so EV compressed roughly in step with equity: from ~$47.5B at the March peak to ~$38.4B now, about −$9.1B [10]. A business deleveraging into its own drawdown is being valued as if its cash-generating capacity had permanently stepped down.

Sources: share count [11]; net-debt reconciliation [12]; consensus FCF [13].

Set the two moves side by side: consensus FY2026 EPS fell 14.4% and FY2027 fell 12–16%, while forward FCF rose and the market cap fell 21.5% (28.1% at the trough). Price fell more than the front-year earnings estimate, and in the opposite direction to forward cash flow. Whether that is an overshoot or a correct anticipation of a permanent step-down is the question the NPV arithmetic frames.

The NPV arithmetic — two scenarios

The method is deliberately transparent: value EQT as a perpetuity of normalized free cash flow, then ask how much each diagnosis subtracts from it. Assumptions, stated:

• Discount rate 10%. This is Ruchir's own yield reference bar. It is conservative for a commodity E&P (12%+ is defensible), and a lower rate makes the permanent-damage perpetuity larger, so it works against finding a mispricing.
• Normalized (mid-cycle) FCF $3.3B — the average of consensus FCF for 2026–2029 ($3.12B, $3.14B, $3.83B, $3.37B) [14]. No terminal growth — the LNG and data-center demand upside is set aside.
• Shares 615.7M [15].

This calibrates almost exactly to the tape: $3.3B ÷ 0.10 = $33.0B, against a current market cap of $32.8B. In plain terms, the market is already capitalizing EQT at roughly $3.3B of perpetual cash flow with zero growth credit. At the peak ($41.8B) it was capitalizing about $4.18B — so the drawdown is the market lowering its estimate of perpetual cash flow by about $0.90B per year.

Scenario T — temporary. The low-price window depresses FCF for a few years, then reverts. Take three years at a punitive $1.0B/year shortfall (Q2-2026 annualized to ~$1.3B of FCF against ~$3.3B normalized; full-year 2026 consensus is barely below normalized because H1 was strong, so $1.0B for three years overstates the dip). PV of damage = 1.0/1.10 + 1.0/1.21 + 1.0/1.331 = $2.5B.

Scenario P — permanent. A level step-down in steady-state FCF, capitalized as a perpetuity. Two anchors bracket it. Low: the FY2027 EPS cut of $0.62/sh (90-day) on 615.7M shares is −$0.38B/year → $3.8B. High: capitalize the full $0.90B/year reduction the tape now implies → $9.0B. Range $3.8B–$9.0B, midpoint ~$6.4B.

Probability-weighted. At the trial's p_temporary = 0.71: expected damage = 0.71 × $2.5B + 0.29 × $6.4B = $3.6B (range $2.9–4.4B across the permanent anchors).

Source: DCF-lite derived from consensus FCF and the trial's p_temporary; discount rate 10%, no terminal growth. Permanent range $3.8B–$9.0B; midpoint shown [16].

The gap. Price damage of $9.0B (peak-to-current) against probability-weighted NPV damage of ~$3.6B leaves a gap of about $5.4B — roughly 60% of the drawdown, and about 16% of the current $32.8B market cap. The gap is real and material.

Its size, though, is entirely a function of the diagnosis. If the hit is purely temporary, NPV damage is ~$2.5B and the price overshot by ~$6.5B — it fell about 3.6× the value actually lost. If it is purely permanent at the high anchor, NPV damage is ~$9.0B and the gap closes to zero: the price move would simply equal the value move. A "no-mispricing" outcome is a live reading, not a rhetorical concession — it is what the 29% permanent tail buys.

The trial

The temporary-versus-permanent question was argued by two opposing corpus-cited briefs and ruled on by three independent judges. The ruling: the probability the impairment is temporary is 0.71, with a per-judge range of 0.56 to 0.72 and a mean of 0.66; the tally records the result as not contested, with an order-stability gap of 0.07 between reading sequences. The diagnosis this report carries is that number — it is not mine to override, and my arithmetic only frames what it is worth.

Source: trial tally, per-judge decisions (p_temporary 0.71; mean 0.66; spread 0.16; not contested).

The temporary case at its strongest. The adverse input is a mean-reverting commodity price, not an asset write-down. EQT generated "$330 million of free cash flow attributable to EQT in Q2 despite natural gas prices averaging just $2.89 per MMBtu," which it frames as proof of "our advantaged position at the low end of the cost curve" [17]. Through the drawdown the balance sheet strengthened, not weakened: "We exited the quarter with net debt of just under $5.7 billion… with Fitch upgrading EQT to BBB during the quarter" — a credit upgrade into a price trough, the opposite of a franchise impairment [18]. And forward earning power rose rather than fell — consensus FCF climbs to $3.83B by 2028 while management raised volume guidance [19][20].

The permanent case at its strongest. The shareholder's unit of ownership shrank permanently, independent of gas price. Diluted shares rose from 260.6M in FY2020 to 615.7M in FY2025 — a 136% increase, a 5-year CAGR of 18.8% — issued for the Tug Hill/XcL, Equitrans, and Olympus acquisitions, with no repurchases in 2024 or 2025 [21][22]. The dilution did not buy proportional reserves: total proved reserves rose only from 27,597 Bcfe (FY2023) to 28,046 Bcfe (FY2025), so proved reserves per diluted share fell from about 66.8 to 45.6 Mcfe/share — a 32% decline in the resource each share owns [23][24]. Even if gas recovers, each share owns less proved resource and must fund more midstream infrastructure — MVP firm capacity runs "through June 30, 2044," and Q2-2026 added LNG vessel leases of about $590M undiscounted.

Source: proved reserves 27,597 Bcfe (FY2023) and 28,046 Bcfe (FY2025) over diluted shares of 413.2M and 615.7M [25][26][27].

The two cases do not actually contradict on the facts — they disagree on which fact governs. Corporate free cash flow can be healthy (temporary) while per-share resource ownership is permanently lower (permanent); the market cap already embeds the 615.7M-share base, so the dilution explains why the level is lower, not, on its own, why the drawdown is permanent. The judges weighed both and landed at 0.71 temporary — closer to the cyclical reading, but with a live 0.29 tail that the arithmetic above cashes as the difference between a $5.4B mispricing and none.

Which line broke, and whether it self-corrects

For a single-segment Appalachian producer there is essentially one revenue driver, and it is the one that moved: the realized natural-gas price. Q2-2026's $2.89/MMBtu is the proximate cause of both the earnings dip and the estimate cuts [28][29]. The repricing mechanism on the temporary side is the demand calendar management points to — power, LNG, and data-center pull tightening the market into 2028–2029 — which, if it converts to firm contracts, lifts realized price and basis without EQT changing its cost structure.

The mechanism on the permanent side is different in kind and does not self-correct with price: share count and fixed infrastructure obligations. Dilution is not reversed by higher gas; it is reversed only by retiring shares, and there were none in 2024–2025 [30]. The clean test between the two, from the trial's flip conditions, is observable and near: whether the FY2026–27 10-K books a property impairment or negative proved-reserve/PUD revision (a real asset-base cut) rather than just a current-year price mark, and whether FY2026–27 realized FCF lands near the ~$3.1B consensus or keeps getting cut. The first would move the diagnosis toward permanent and close the gap; the second, toward temporary and widen it.