Autoresearch: Appalachian/Haynesville gas producer exposure to an AI-datacenter price floor

Generated by /autoresearch on 2026-07-22. Synthesized across 3 rounds from web search + fetch (two primary industry sources — naturalgasintel.com, rbnenergy.com — returned HTTP 403 and are noted in Provenance; findings rest on S&P Global, EIA-derived reporting, company filings/decks, and analyst notes). Treat as raw material — review before promoting. Context: vault/projects/stock-market

Summary

This pass closes the ⚠ unverified step 4 of ai-gas-demand-to-appalachian-producer-price-floor — the producer-exposure leg. The finding: the hypothesis has real differentiated exposure, but which producer captures a price floor depends heavily on hedge posture (an over-hedged producer gives the floor away) and egress (a gas-rich producer can be price-trapped by basis). On both axes the ranking is legible. EQT is the most unhedged of the majors (only ~25% of 2026 hedged) and its CFO argues Appalachian basis "structurally tightens through the end of the decade." Expand Energy (EXE) — North America's largest gas producer at ~7.5 Bcfe/d — runs a "Hedge to Wedge" program that explicitly retains upside, sits in the Haynesville (short-haul to Gulf LNG + Southeast data centers), and markets itself as "powering U.S. data centers." Antero (AR) is the most hedged (60%+ of 2026), so it gives away the most of any bare-gas-price floor, though it has the best NGL and premium firm-transport mix. Demand-lock is materializing: ~101 GW of announced behind-the-meter (BTM) gas generation and 35+ data-center campuses in dedicated, take-or-pay supply agreements.

Findings

Hedge posture — the single biggest differentiator of who captures the floor

An AI-driven price floor only reaches the equity of a producer that hasn't already sold its upside forward. The 2026 hedge books diverge sharply:

• EQT — ~25% hedged for 2026, the most exposed to upside. EQT raised its 2026 hedge percentage only from 7% to 25%, adding collars with weighted-average floor/ceiling of $3.94 / $5.70 per MMBtu (S&P Global Commodity Insights). As the largest Appalachian producer, EQT offers the most direct beta to a sustained Henry Hub lift.
• Antero (AR) — 60%+ hedged for 2026, ~1/3 hedged for 2027. The 2026 book is 43% swaps at a $3.91 floor plus 18% collars ($3.25 floor / $5.66 ceiling) (S&P Global; Antero 1Q2026 deck). AR gives away the most of a 2026 price floor — but its heavy NGL cut and premium firm transport to the Gulf are a partly-independent value driver.
• Range Resources (RRC) — hedges "portions"; moderate. RRC guides 2.35–2.40 Bcfe/d in 2026 growing to 2.6 Bcfe/d in 2027 and hedges to stabilize cash flow without a disclosed fixed percentage in the reporting reviewed (Range Q4'25/2026 guidance; Motley Fool transcript). Range's ~30-year inventory life is the long-runway angle.
• Expand Energy (EXE) — "Hedge to Wedge," designed to retain upside. EXE, North America's largest gas producer (~7.5 Bcfe/d), runs a "proactive hedging program that reduces volatility... while retaining upside exposure," booked ~$200M of realized hedge gains, and generated record Q1'26 FCF of $1.7B while cutting gross debt $1.3B (Expand 8-K/AInvest; Investing.com Q1 slides). One analyst frames EXE as "the best way to play AI in 2026" (Barchart).

Ranking for capturing a bare price floor (most upside retained → least): EQT ≈ EXE > RRC > AR.

Egress / basis — a gas-rich producer can still be price-trapped, but Appalachia is de-constraining

• Appalachian takeaway is constrained but improving. Production is expected to grow ~2% in 2026 to average ~37.6 Bcf/d, aided by Mountain Valley Pipeline and Williams' Regional Energy Access (+0.8 Bcf/d firm, online late 2024); the MVP Southgate Extension (+0.55 Bcf/d) is permitted through June 2026 (Natural Gas Intelligence, EIA-based).
• Basis is structurally tightening. Henry Hub forwards average $3.813 (2026) → $3.867 (2027); Appalachian Tetco M-2 forwards average $2.940 (2026) → $3.017 (2027), and forward strips for Winter '26/'27 through Winter '27/'28 have "strengthened to all-time highs." EQT's CFO: "Appalachian basis should structurally tighten through the end of the decade" (RBN Energy basis tag; AEGIS Hedging Dominion South). A tightening basis means more of any Henry Hub floor reaches the Appalachian wellhead — directly rebutting the hypothesis's egress concern.
• In-basin demand is the egress bypass. Behind-the-meter data-center load sited on top of Marcellus/Utica gas turns an egress problem into a local off-take — the producer sells into demand without needing long-haul pipe. Haynesville (EXE) has the structural edge of short-haul to both Gulf LNG and the Southeast data-center build. EIA sees Haynesville overtaking Permian in 2027 gas growth as Henry Hub tops $4 (NGI/EIA).

Direct demand-lock — the datable catalyst the hypothesis asked for

• Scale of announced BTM gas: ~101 GW of behind-the-meter natural-gas generation announced by data-center developers in the US as of 2026, with >57 GW already having disclosed equipment orders; >19.2 GW of named BTM deals in Jan–mid-Apr 2026 alone (Marcellus Drilling News; Avanza Energy).
• Direct producer↔demand contracts exist. Analysts estimate 35+ US data-center campuses have executed or are in advanced negotiation for dedicated gas supply agreements, ~18 GW of BTM capacity; typical structure is a 10+ year take-or-pay (80–90% of volumes), Henry-Hub-indexed with a floor and ceiling (Avanza Energy). Floors written into off-take contracts are literally the price floor the hypothesis posits — and midstream (Williams) explicitly cites Haynesville + Marcellus fueling data-center growth (NGI/Williams).
• SoftBank's Stargate is developing a hyperscale gas corridor on Appalachian Basin gas (Marcellus/Utica), leveraging Ohio HB15's 90-day approval and $4.2B of AEP Ohio transmission ([search-surfaced, corroborate before citing as a chain step]).

Contradictions and open questions

• Precise RRC and EXE 2026 hedge percentages weren't pinned to a single figure (Range discloses a schedule; EXE describes the program qualitatively). Confirm from the latest 10-Q hedge tables before sizing.
• Antero's NGL/premium-transport value may offset its heavy gas hedging — AR is not purely a "gave-away-the-upside" story; a fuller model would weigh its C3+ realizations and Gulf firm transport.
• The floor could be capped from the supply side (the hypothesis's own falsifier): Haynesville overtaking Permian growth in 2027 and faster associated-gas growth could re-loosen the market — a rig-add / production-cadence watch item, not resolved here.
• Two core industry primaries (RBN Energy, Natural Gas Intelligence) 403'd on direct fetch — the egress/basis specifics lean on secondary summaries; verify the "structurally tightening basis" claim against a fetched RBN/EQT primary before graduating the mechanism.

Provenance

Rounds run: 3 (full).

Sub-questions by round:

• R1: (1) 2026 hedge books / % hedged for EQT/AR/RRC/EXE; (2) Appalachian takeaway/egress + basis vs Henry Hub; (3) hyperscaler↔producer direct gas supply agreements.
• R2: (1) RRC + EXE hedge/egress specifics — targeting the two producers R1 left thin; (2) Appalachian producer positioning playbook (RBN).
• R3: (1) Expand Energy hedging posture + data-center leverage — targeting "who retains upside at scale."

Anchor source (Grokipedia): not attempted — a hedge-book/egress question has no encyclopedic anchor of load-bearing value.

URLs fetched (successful + failed):

• S&P Global — Antero/EQT/CNX diverge on hedging — industry — the hedge-percent divergence (load-bearing).
• [Failed: naturalgasintel.com Appalachian-producers-brace-for-demand-surge] — HTTP 403.
• [Failed: rbnenergy.com new-Appalachian-gas-producer-playbook] — HTTP 403.
• Search-snippet-level (not deep-fetched): EIA/NGI Haynesville-overtakes-Permian; Marcellus Drilling News BTM; Avanza Energy "$416B gas grab"; Barchart EXE; Investing.com EXE Q1 slides; AInvest EXE; Range IR 2026 guidance; Antero 1Q2026 deck (PDF, cloudfront).

Tools used: WebSearch, WebFetch. Generated: 2026-07-22.