Snapshot

Natural gas is the dominant fuel for new gas-fired power plants being built to serve AI data centers. The US produced 118.5 Bcf/d of marketed natural gas in 2025 (EIA), of which 36.6 Bcf/d came from Appalachia alone. The five producers listed here — EQT, Antero Resources (AR), Range Resources (RRC), Expand Energy (EXE, formerly Chesapeake+SWN, merged October 2024), and CNX Resources — collectively produced roughly 21 Bcf/d in 2025, or about 18% of national output. All operate primarily in Appalachia (Marcellus/Utica), except Expand Energy which also operates in the Haynesville shale.

These five producers generated roughly $22 billion in combined revenue and $4.8 billion in combined net income in FY2025 on $3.52/MMBtu average gas. Two structural demand shocks are arriving simultaneously: LNG export capacity is jumping from 11.9 Bcf/d (2024) to 16.4 Bcf/d (2026), and AI data center gas burn could add another 2–7 Bcf/d by 2030 est.. Gas price is what determines earnings: at $3/MMBtu these companies collectively earn roughly $5B; at $5/MMBtu, $12B+ est..

The product & how money is made

The product is natural gas (methane), measured in millions of cubic feet (Mcf) or British thermal units (MMBtu); 1 Mcf ≈ 1 MMBtu. Producers drill horizontal wells into shale formations (primarily Marcellus, Utica, and Haynesville), fracture the rock to release trapped gas, then sell the output at prices linked to regional hubs or the Henry Hub benchmark.

Revenue comes from three streams:

• Dry gas sales — the bulk for most Appalachian producers. Priced at Henry Hub minus a basis differential reflecting local pipeline congestion ($0.30–$1.00/MMBtu historically).
• Natural gas liquids (NGLs) — ethane, propane, butane from wet-gas wells. AR and RRC produce significant NGLs (38% and 31% of output respectively), diversifying away from pure gas pricing. NGL prices track oil more than gas.
• Hedging gains/losses — all five use swaps and collars to lock in future prices. AR has 1.3 MMBtu/d hedged at ~$3.92 for 2026; EQT has 25% of 2026 hedged with floors at $3.50–$4.25; CNX has 81% of 2026 gas hedged at $3.64.

Cash profit per Mcfe = realized price minus (royalties + operating costs + gathering/transport/processing + G&A + interest + DD&A). Full-cycle breakevens from Incorrys data (Q1 2022–Q3 2024): Haynesville ~$3.43/Mcf, Northeast Marcellus ~$3.52/Mcf, Utica ~$1.92/Mcf (NGL-rich wells benefit from liquids uplift) est.. At $3.50 Henry Hub, Haynesville and dry Marcellus are near breakeven; at $4.00+, all basins are strongly profitable.

Sources: EIA, Incorrys North American Natural Gas Breakeven Cost Study (Q1 2022–Q3 2024), company 10-Ks.

Demand

Contracted / visible demand

• LNG exports (contracted): US LNG export capacity jumping from 11.9 Bcf/d in 2024 to 16.4 Bcf/d by end of 2026 (EIA) — a 4.5 Bcf/d increase. Key facilities ramping: Plaquemines LNG (3.8 Bcf/d total), Corpus Christi Stage 3 (~1 Bcf/d), Golden Pass (2.1 Bcf/d, Train 1 started March 2026). By 2029, total capacity could reach ~28.7 Bcf/d est.. contracted
• Power sector (existing): 35.8 Bcf/d consumed in 2025, down 1.0 Bcf/d from 2024 as solar/battery displaced some gas generation. Largest single consuming sector. contracted
• US total consumption: 92.0 Bcf/d in 2025 (record), up from ~90.3 Bcf/d in 2024. contracted

Forecast / emerging demand

• AI data center gas demand: RBC Capital Markets (May 2026) estimates ~6.1 Bcf/d incremental gas from data centers by 2030, a ~20% increase to baseline power demand. Behind-the-meter gas generation totals ~101 GW announced, with 57 GW having disclosed equipment orders and 7 GW under construction. Largest single project: Homer City, Pennsylvania (4.4 GW, ~665,000 MMBtu/d). est.
• Enverus (more conservative): 2.1 Bcf/d high-confidence behind-the-meter demand by 2030, upper bound 4.1 Bcf/d if all new data center capacity were gas-fired. est.
• Front-of-meter gas plant pipeline: 64 GW of planned natural gas power projects in interconnection queues by 2030, up from 23.8 GW at start of 2025. New projects average 590 MW each. est.
• Industrial / residential: Industrial gas demand 23.6 Bcf/d in 2025 (+0.2 YoY). Residential and commercial rose on cold weather but are not structural growth drivers.
• Mexico pipeline exports: Growing ~1–2 Bcf/d as Mexican industrialization continues. est.

Sources: EIA STEO (March 2026), RBC Capital Markets (May 2026), Enverus Intelligence Research.

Supply

Current production capacity

• US total: 118.5 Bcf/d in 2025, up 5.3 Bcf/d (+4.7%) from 2024. EIA forecasts a new record in 2026.
• Appalachia (Marcellus/Utica): 36.6 Bcf/d (31% of US total), up 1.1 Bcf/d YoY. Growth throttled by pipeline takeaway, not geology.
• Haynesville: 14.9 Bcf/d (~13% of total), up ~4% YoY. EXE operates ~3.2 Bcf/d here, closest to Gulf Coast LNG demand.
• Permian (associated gas): 27.7 Bcf/d (23% of total), up 2.7 Bcf/d (+11%). Byproduct of oil drilling — grows regardless of gas price, adding price-insensitive supply.

Key bottleneck: Appalachian pipeline takeaway

Appalachian producers sit on the cheapest, most prolific gas resource in North America, but pipeline capacity to demand centers is at maximum utilization, depressing Appalachian basis differentials $0.30–$1.00 below Henry Hub. New capacity:

• Mountain Valley Pipeline (MVP): 2 Bcf/d, began operations June 2024. Owned by EQT. 0.5 Bcf/d expansion filed ("MVP Boost") plus MVP Southgate extension into North Carolina. contracted
• Borealis Pipeline (proposed): Up to 2 Bcf/d from western Appalachia to Gulf Coast via existing Texas Gas Transmission system. FID expected Q1 2026. paper
• Tioga Pathway: 190 MMcf/d out of Northeast Pennsylvania. paper

New interstate pipeline permitting takes 3–5+ years. Takeaway bottleneck caps Appalachian production growth at ~1 Bcf/d per year despite hundreds of Tcf of proved reserves.

Capital discipline

Post-2020, all major gas producers shifted to maintenance-mode spending, prioritizing FCF and shareholder returns over volume growth. EQT’s 2026 maintenance capex is $2.1–$2.2B to hold ~6.5 Bcf/d flat; growth capex of $580–$640M goes to pipeline and water infrastructure. AR guides 4.1 Bcfe/d on $1B D&C capital (a 20% production jump, but only because it curtailed in 2024–2025 at low prices). CNX runs on $556–$586M annual capex for ~1.7 Bcfe/d.

Sources: EIA (March 2026), EQT Q4 2025 press release, AR Q4 2025 press release, company investor presentations.

The gap

Henry Hub averaged $2.20/MMBtu in 2024 and rose 60% to $3.52 in 2025, driven by LNG ramp-up and cold weather. June 2026 spot: $3.17/MMBtu. If 10+ Bcf/d of incremental demand materializes against supply growth of 3–5 Bcf/d per year, the market must either incentivize new drilling (requiring $3.50–$4.00+ sustained) or ration demand via higher prices. The Haynesville breakeven at $3.43/Mcf acts as a soft floor — below this, rigs get dropped and supply falls. Complication: Permian associated gas grows ~2–3 Bcf/d annually regardless of gas price, adding supply that gas-focused producers cannot control.

Sources: EIA, RBC Capital Markets, Enverus, Incorrys.

The players

Sources: EQT Q4 2025 press release, AR Q4 2025 press release, RRC Q4 2025 press release, EXE Q4 2025 press release, CNX Q1 2026 earnings, StockAnalysis.com for market caps.

The price of exposure

Gas price sensitivity: Every $0.25/MMBtu change in Henry Hub shifts EQT’s annual revenue by ~$600M and FCF by ~$400M (pre-hedge, on ~6.5 Bcfe/d) est.. For CNX at 1.7 Bcfe/d, a $0.25 move is ~$155M in revenue est.. AR is partially insulated: ~39% of revenue comes from NGLs/oil, which don’t move with gas.

Capital return mechanisms:

• CNX — retired 38% of shares since 2016 (229M → 142M), continues aggressive buybacks. Cheapest per-dollar-of-earnings at 7.4x trailing P/E.
• RRC — $1.5B authorized buyback, raised dividend 11.1%. Net debt/EBITDAX 0.8x.
• EQT — prioritizing debt reduction ($7.7B → target $4.7B by end 2026) before material buybacks. Owns Mountain Valley Pipeline, generating midstream fee income independent of gas prices.
• EXE — base dividend $0.575/share ($2.30/yr, ~2.5% yield), targeting $1B+ annual debt reduction. Still integrating CHK/SWN merger.
• AR — strongest balance sheet (net debt/EBITDA ~0.5x), guiding 20% production increase in 2026 on flat capital. Returning cash via opportunistic buybacks.

Sources: Company press releases, StockAnalysis.com. P/E and FCF yield calculated from data above.

What to deep-dive next

• Hedging books in detail: CNX has 81% of 2026 hedged at $3.64 — if gas goes to $4.50, CNX captures far less upside than EQT (only 25% hedged). Hedge positions determine which company actually captures a gas price rally.
• Pipeline ownership economics: EQT’s Equitrans acquisition gave it the Mountain Valley Pipeline (2 Bcf/d, expanding to 2.5). Midstream cash flow profile, contract structure, and re-contracting opportunity deserve separate analysis.
• Expand Energy integration: EXE has the largest production base (7.5 Bcf/d guided) but carries $4.4B net debt and has only existed since October 2024. Haynesville wells need $3.43/Mcf to break even — the dual-basin model works differently from a pure Appalachian producer.
• Permian associated gas cap on pricing: Permian gas grows 2–3 Bcf/d annually regardless of gas price. How much this caps Henry Hub upside is the critical macro question for all gas producers.
• Behind-the-meter direct contracts: 101 GW of announced behind-the-meter gas generation for data centers. Fixed-price supply agreements directly with hyperscalers (bypassing Henry Hub) could create premium revenue streams. No producer has disclosed such a contract yet.

Sources & confidence