Williams Companies
Fee-based, take-or-pay regulated/contracted infrastructure: ~90% of EBITDA from fixed-fee transmission & gathering capacity, increasingly augmented by long-term (10-yr) take-or-pay power-supply contracts with hyperscalers. C-corp (not an MLP).
The thesis on this name
State of Data-Center Power
Gas-to-power at the meter — $5.1B 'power innovation' portfolio building modular gas plants AT datacenter sites (Project Socrates online H2 2026).
State of Data-Center Power
$5.1B power-innovation portfolio building on-site datacenter gas plants; lowest-beta, dividend-paying diversifier.
Earnings, margins, COGS & capex
Williams is a fee-based natural-gas infrastructure compounder delivering record results — FY2025 revenue $11.83B (+10%) and record Adjusted EBITDA $7.75B (+9%, ~9% 5-yr CAGR), with Q1 FY26 net income up 25% YoY to $865M and Adjusted EBITDA up 13% to a record $2.254B (fact). Growth is driven by Transco transmission expansions, gathering volumes, gas-marketing margin, and the new behind-the-meter power business. The model is ~90% fee-based take-or-pay, so EBITDA is highly contracted; the trade-off is heavy, debt-funded capex during the current datacenter build-out cycle that pushes free cash flow negative.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~34¢ is cost of goods and ~44¢ operating expense, leaving ~22¢ of operating profit (~22¢ net).
Revenue trend
Margins
up
up
flat
down
COGS structure
COGS is dominated by the cost of natural gas and NGLs purchased for the Gas & NGL Marketing Services segment (a low-margin commodity pass-through that inflates both revenue and COGS), plus pipeline fuel/operating costs, compression power, plant O&M, and depreciation on a ~$60B+ asset base. The high-margin transmission & gathering fee business carries low incremental COGS; reported revenue/COGS is heavily distorted by commodity marketing flow-through, which is why EBITDA (not gross margin) is the management metric.
Capex
FY2026 guidance: growth capex $7.0–7.6B + maintenance $0.85–0.95B (fact). Funds Transco expansions (Southside Reliability, Socrates pipelines), Louisiana Energy Gateway / Gulf gathering, Northeast G&P, and the $5.1B+ power-innovation portfolio (Socrates, Apollo, Aquila, NEO behind-the-meter plants). This is a step-up vs historical ~$2-3B — the datacenter cycle is materially raising capital intensity.
Latest earnings
Beat — Adjusted EPS $0.73 topped consensus; management said FY26 tracking to the upper half of guidance (fact)
FY2026 raised: Adjusted EBITDA $8.05–8.35B, Adjusted EPS $2.20–2.38, AFFO $6.085–6.315B; growth capex $7.0–7.6B; leverage ~4.0x; dividend +5% to $2.10 annualized (fact)
- Adjusted EBITDA (Q1 FY26)
- $2.254B (+13% YoY, record)
- AFFO (Q1 FY26)
- $1.77B (+22% YoY)
- Transco contracted capacity
- 33.4 Bcf/d (record, FY2025)
- Power portfolio committed
- $5.1B+ behind-the-meter
Growth drivers
- Transco interstate pipeline expansions — record 33.4 Bcf/d contracted capacity (FY2025); Southeast/Gulf demand pull from LNG, power, datacenters (fact)
- $5.1B+ behind-the-meter power-innovation portfolio — Socrates (400MW, Meta), Apollo+Aquila ($3.1B, online H1 2027), NEO (682MW, H2 2028), all on ~10-yr take-or-pay (fact)
- AI/datacenter electricity demand bypassing the congested grid — structural multi-year demand for on-site gas generation (fact/trend)
- LNG-export feedgas — Gulf Coast transmission & gathering volumes tied to rising US LNG exports (fact)
- Sequent gas-marketing platform leveraged as a 'growth multiplier' to bundle supply+transport+storage+power for hyperscalers (fact)
- Potential upstream gas-asset acquisitions to offer hyperscalers a single integrated supply-to-power solution (estimate — exploratory, reported)
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-24. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
Williams owns the irreplaceable Transco backbone and is the clearest scaled midstream way to play AI/datacenter power demand — pairing ~90% fee-based pipeline cash flows with a fast-growing, contracted behind-the-meter generation business at premium returns.
- Transco is a structurally advantaged, demand-pull, FERC-regulated franchise feeding the fastest-growing US gas demand (LNG, power, datacenters) — record 33.4 Bcf/d contracted (FY2025)
- $5.1B+ behind-the-meter portfolio (Socrates/Meta, Apollo, Aquila, NEO) on 10-yr take-or-pay adds a new, high-return growth leg with blue-chip hyperscaler counterparties (fact)
- Record FY2025 EBITDA $7.75B and Q1 FY26 +13%; FY26 guide raised and tracking upper half — visible double-digit growth with ~90% fee-based stability
- 17+ consecutive years of dividend growth (+5% to $2.10, FY26) with AFFO coverage; durable income plus growth
- Integrated supply→transport→storage→power offering (incl. Sequent and potential upstream M&A) is hard for any single competitor to replicate
The stock already prices in the AI-power optionality at ~17-18x EV/EBITDA and ~33x forward P/E — rich for a capital-intensive, ~$30B-net-debt midstream whose new power business carries execution risk and rides hyperscaler-capex cyclicality.
- Valuation is stretched — ~17-18x EV/EBITDA and ~33x forward P/E vs. midstream-utility norms; consensus avg target (~$74-83) sits near/below the price (fact), limited margin of safety
- Free cash flow is negative through the build-out — $7.9–8.6B FY26 capex exceeds ~$5.9B operating cash flow, funded by debt; ~4.0x leverage
- Behind-the-meter power is a newer competency — EPC, turbine-supply (slots tight through 2030), and plant-operating execution risk on multi-billion projects
- Datacenter/AI demand is itself a cyclical bet — a hyperscaler capex pullback or grid-policy shift could strand or delay power projects
- Crowded field — Energy Transfer, Kinder Morgan, utilities and turbine-rich new entrants chase the same loads; commoditization could compress the premium returns the bull case assumes
What it is worth
EV/EBITDA and DCF/AFFO-yield for a contracted-infrastructure name, cross-checked vs. dividend yield and analyst targets
~$60-67 (multiple de-rates toward ~14-15x on a datacenter-capex slowdown or rate-driven multiple compression)
~$78-83 (consensus avg ~$74-83
mid-single-digit upside; EBITDA growth offsets a flat-to-slightly-lower multiple)
~$95-98 (Morgan Stanley $98 high target — sustained power-portfolio wins + multiple holds at ~18-19x EV/EBITDA)
Trades ~17-18x EV/EBITDA and ~33x fwd P/E on ~$8.2B FY26 EBITDA guide — a premium midstream multiple already crediting the AI-power optionality; ~2.8-3.5% dividend yield. (fact: metrics; estimate: scenarios)
SWOT
Strengths
- Transco — the largest-volume US interstate gas pipeline (~33 Bcf/d contracted, FY2025), an irreplaceable, FERC-regulated, demand-pull asset with structural advantage to the high-growth Southeast/Gulf
- ~90% fee-based, take-or-pay cash flows insulate EBITDA from commodity prices; 9% 5-yr EBITDA CAGR with record FY2025 results
- First-mover scale in behind-the-meter datacenter power — $5.1B+ committed, marquee Meta (Socrates) anchor, 10-yr contracts
- Investment-grade balance sheet supporting heavy capex — ~4.0x leverage with 17 consecutive years of dividend growth
Weaknesses
- Heavy, debt-funded capex ($7.9–8.6B FY26) pushes free cash flow negative during the build-out; ~$30B net debt
- Power-generation / EPC execution is a newer competency vs. core pipeline operations — construction, turbine-supply, and plant-operating risk
- Gas-marketing (Sequent) and commodity exposure introduce earnings volatility quarter to quarter
- Capital-intensity step-up means returns depend on contracts being honored and projects landing on-time/on-budget
Opportunities
- Multi-year AI/datacenter power demand bypassing the constrained grid — Williams positioned as integrated supply-to-power provider
- Bundle pipeline + generation + storage + Sequent marketing into one-stop hyperscaler offering ('infrastructure solutions provider')
- Potential upstream gas-asset M&A to close the supply loop and capture more of the value chain
- Continued Transco/LNG-feedgas expansions and Pacific NW (Silver Spur/Rockies Columbia Connector) — first major PNW expansion in 20+ years
Threats
- Hyperscaler capex/AI-demand cyclicality — a datacenter spending pullback would strand or delay power projects
- Competition for the same datacenter loads from Energy Transfer, Kinder Morgan, utilities, and turbine-constrained new entrants
- Permitting/FERC and environmental opposition to new pipelines and on-site gas plants (emissions, local siting)
- Interest-rate sensitivity given ~$30B debt and a yield-oriented investor base; equity already rich at ~17-18x EV/EBITDA
Moats, dependencies & bottlenecks
Moats
Transco interstate pipeline network (right-of-way + FERC franchise) Largest-volume US gas pipeline; right-of-way and permitting make it effectively impossible to replicate — the core moat (fact)
Capacity reservations and 10-yr power contracts insulate EBITDA from volume/commodity swings
transport, storage, marketing, now generation) One-stop integrated offering for hyperscalers is hard to match, but each piece individually has competitors
Incumbency at demand centers / first-mover in behind-the-meter datacenter power Marquee Meta (Socrates) anchor and a $5.1B head start, but new entrants and turbine OEMs are crowding in
Investment-grade balance sheet + low cost of capital for mega-projects Enables financing $8B/yr capex others can't; eroded if rates rise or leverage climbs
Dependencies
Power-innovation returns hinge on hyperscalers honoring 10-yr contracts and keeping AI-buildout demand; a capex pullback is the key downside (fact — Meta anchors Socrates)
permitting) Transco rate cases and pipeline/plant permits gate every expansion; opposition/delay risk is structural (fact — Socrates South approved Jun-2025)
Siemens Energy, GE Vernova) Behind-the-meter plants depend on turbine supply; OEM slots are tight through 2030, a real build-out bottleneck (fact)
Haynesville, Permian producers) Needs feedgas for pipelines and on-site plants; exploring upstream M&A to internalize this (estimate)
~$30B net debt and ~$8B/yr capex make it sensitive to rates and credit access (fact)
Advantages
- Owns the Transco backbone — irreplaceable right-of-way feeding the fastest-growing US gas-demand corridors (LNG, Southeast power, datacenters)
- ~90% fee-based take-or-pay cash flows give EBITDA stability rare among AI-exposed names
- First-mover scale ($5.1B+) and a marquee Meta anchor in behind-the-meter datacenter power
- Integrated supply→transport→storage→marketing (Sequent)→generation stack hyperscalers can buy as one solution
- Investment-grade balance sheet + low cost of capital to finance ~$8B/yr of projects competitors can't
- 17+ years of dividend growth with AFFO coverage — durable income alongside the growth story
Weaknesses
- Negative free cash flow through the build-out; ~$30B net debt and ~4.0x leverage
- Power-generation EPC/operating is a newer, less-proven competency than pipeline transport
- Commodity/marketing (Sequent) introduces quarter-to-quarter earnings volatility
- Rich valuation (~17-18x EV/EBITDA, ~33x fwd P/E) leaves little margin of safety
- Concentration of the power story in a handful of large hyperscaler counterparties/projects
- Interest-rate and capital-markets sensitivity given the capital intensity and yield-oriented holder base
Bottlenecks
- Turbine and reciprocating-engine supply — OEM order slots are tight through 2030, constraining how fast behind-the-meter plants can be built
- Permitting/FERC and local siting timelines for new pipelines and on-site gas plants
- Capital — $7.9–8.6B FY26 capex exceeds operating cash flow, so growth is debt-funded and balance-sheet-constrained
- Skilled EPC and plant-operations execution capacity for a power-generation business that is newer than the core pipeline operation
- Feedgas availability and pipeline interconnects at specific datacenter sites in grid-constrained regions
Top signals & trends
Top signals
First real proof the gas-to-power model works at scale and on schedule; slippage would dent the thesis (fact)
Q1 beat and raised guide signal momentum; watch each quarter for confirmation (fact)
Apollo, Aquila, NEO and further wins extend the growth runway; disclosed take-or-pay terms validate returns (fact)
Growth is good but funded by debt — leverage drifting above ~4.0x would pressure the credit and equity (fact)
Could deepen the integrated offering but adds commodity/upstream risk outside the fee-based core (estimate)
Supply constraints could delay the power build-out and the associated EBITDA (fact)
Trends
Core tailwind — drives behind-the-meter on-site gas generation, Williams' new growth leg (fact/trend)
Gas is the fastest dispatchable on-site power; structurally favors midstream-built plants (fact)
Feeds Transco/Gulf transmission and gathering growth (fact)
Validates demand but constrains build pace and can raise project costs (fact)
~$30B debt + ~$8B/yr capex make rising rates a headwind to both financing and valuation multiple (fact)
Long-term policy risk to new gas infrastructure, partly offset by gas's role as reliability backbone (estimate)
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Titan 250 / PGM 130 turbines + Cat 3520 reciprocating engines for Socrates behind-the-meter plants
SGT-400 turbines in the Socrates power-generation configuration (German-listed; ADR SMEGF)
Gas-turbine OEM for on-site datacenter power (tight supply slots through 2030)
US natural-gas producers (Marcellus/Haynesville/Permian) Feedgas suppliers; e.g. EQT, Expand Energy — Williams exploring upstream M&A to internalize supply
Build-out of modular plants and pipeline expansions
Anchor behind-the-meter customer — Socrates (400MW, Ohio) via affiliate Sidecat under a 10-yr agreement
Apollo, Aquila, NEO power projects on 10-yr take-or-pay (counterparties incl. top-tier hyperscalers/utilities)
Transco/Gulf feedgas and transmission customers tied to rising LNG exports
Local distribution companies (LDCs) & utilities Long-term firm transportation contracts on Transco/Northwest Pipeline
Transmission capacity to existing grid power plants (Southeast/Gulf)
Large diversified midstream MLP; signed direct gas-supply deals to Texas datacenters — chases the same datacenter load
Largest US gas-pipeline network by miles; pursuing RNG/datacenter gas strategy — closest pipeline peer
Scaled gas-gathering/processing & NGL midstream; overlaps in G&P and gas logistics
Major North American gas transmission + utilities; competes for gas-infrastructure and power-adjacency deals
Turbine OEM also enabling rival on-site power deals (e.g., Crusoe); both supplier and a route competitors use to bypass Williams
Independent power and datacenter-power specialists building behind-the-meter and grid generation for AI loads