Williams (WMB)

A Growth Story That Needs More Money Than It Makes

Williams owns the best-placed gas transmission system in the United States and is spending faster than it earns to make it bigger. FY2025 adjusted EBITDA was $7,750 million and leverage finished the year at 3.71× net debt to adjusted EBITDA. Both look comfortable. The cash statement is where the tension shows: $5,898 million came in from operations, $4,893 million went out on capital spending, and the common dividend took $2,442 million, so the year ended $1,437 million short and the difference was borrowed.

That is not a scandal, and for a company building contracted pipeline into visible demand it may be the right call. It is simply the fact the headline metric hides. Williams reports available funds from operations, AFFO, and covered its dividend 2.40 times on it in 2025. Nothing in that ratio knows that the building programme exists.

The label deserves a moment, because it has no equivalent at the partnerships Williams is screened against. Williams defines AFFO as operating cash flow with working-capital swings taken out, less preferred dividends and cash paid out to minority holders. No capital spending is deducted, maintenance or expansion. That is why AFFO of $5,858 million lands within $40 million of the $5,898 million the business generated from operations. Compare it with Enterprise Products' $8,000 million of distributable cash flow, struck after maintenance capital, or Kinder Morgan's $2,891 million of free cash flow, struck after every dollar of capex including expansion. Three companies, three lines drawn in three different places. Williams draws its furthest upstream, so its number is the largest by construction. And note what DCF means in this corner of the market: distributable cash flow, the partnership payout metric, not discounted cash flow.

Business Overview

The flagship asset is Transco, the pipeline running from the Gulf Coast up the Atlantic seaboard to New York. Transco is rate-regulated, which changes the risk shape: expansions are approved projects earning a defined return on invested capital, closer to a utility build than to the gathering systems that sit around it. Williams also gathers and processes gas in the major producing basins, and the two halves do not behave alike.

Roughly 7.1 Bcf/d of transmission expansion was in execution as of February 2026, including SESE, Line 200, NESE and Power Express, with a further ~14.3 Bcf/d of identified backlog carrying in-service dates from 2027 to 2033 and around $15 billion of potential capital behind it. The backlog is an opportunity set rather than a commitment, and projects leave it as well as enter it.

Two 2026 decisions changed the shape of the company. In July, Blackstone agreed to put $5.34 billion into a joint venture holding Williams' Power Innovation business, the behind-the-meter generation it is building for data-centre load. In August, Williams agreed to buy Momentum Midstream for up to $5.5 billion, about $3.5 billion in cash and debt and $2 billion in stock, at roughly 8.5 times projected 2027 EBITDA. Momentum brings more than 4,000 miles of Haynesville gathering pipe and three take-or-pay pipelines with 4.05 Bcf/d of transport capacity, connecting that basin to Gulf Coast LNG. It has not closed: antitrust clearance is outstanding, and Williams expects to complete later in 2026. Williams raised 2026 adjusted EBITDA guidance to $8.3–$8.5 billion partly on the back of it.

Over 90% of earnings are fee-based on the company's 2024 mix disclosure, which is worth reading precisely. Fee-based means the revenue does not move with the gas price. It does not mean the revenue does not move with volume, and it is a wider and weaker category than take-or-pay, where the shipper pays whether it ships or not. The contract quality guide works through the difference. A separate 93% figure appears in the NGL processing footnotes and refers to volumes under fee-based processing contracts, not to the share of company earnings.

On demand: the EIA's short-term outlook put US LNG exports at an average 15.1 Bcf/d in 2025, rising to 17.0 Bcf/d in 2026 and 18.2 Bcf/d in 2027. Williams' Atlantic-side system links Gulf Coast supply to Southeast and Mid-Atlantic demand, and its newer projects are aimed squarely at power generation.

How the Economics Work

A Transco expansion is a straightforward proposition. Shippers sign long-term contracts, the regulator approves a rate, Williams builds, and EBITDA arrives when gas flows. The commercial risk is concentrated in timing and permitting rather than in the gas price, which is what the fee-based share is telling you. Slippage is the thing that hurts: the debt is drawn during construction, so a delayed in-service date raises leverage before it raises earnings.

Because AFFO deducts no capital spending, it cannot answer whether the dividend and the building programme fit together. The items that move it without moving adjusted EBITDA are cash interest, cash taxes, preferred dividends, payments to minority holders and the difference between joint-venture earnings and joint-venture cash actually received. Those are worth tracking. But the ratio they produce is not a coverage ratio in the partnership sense, and it should not be read as one.

The test that survives is the one no accounting choice can reclassify: operating cash flow, less all capital spending, less the dividend. In 2025 that was negative $1,437 million. For 2026 the company guides to $7.3–$7.9 billion of growth capital and $850–$950 million of maintenance capital, against $8.3–$8.5 billion of expected EBITDA, so the gap widens before it narrows. Williams is funding it with debt, with the $5.34 billion Blackstone joint-venture investment, and with the stock it is issuing for Momentum. That is a legitimate way to build. It is also the reason leverage and the equity count are the numbers to watch rather than the dividend cover.

As a corporation, Williams pays tax on its profits before it pays a dividend, and holders receive a Form 1099 rather than a Schedule K-1. The MLP versus C-corp guide sets out what that changes for the person holding the shares.

Valuation Framework

The usual anchor is EV divided by adjusted EBITDA, and large-cap midstream works to a rough 8–10× band. Two index prints get quoted around it: the Alerian MLP Infrastructure Index at 8.57× forward and the broader Alerian Midstream Energy Select Index, which holds corporations as well as partnerships, at 10.94×. It is tempting to read those 2.4 turns as what a corporate wrapper is worth to Williams. Most of it is not. The two indices hold different companies with different assets and geographies, so the spread mixes structure with mix and cannot separate them. Anchor to named peers instead, and to what the assets are: a rate-regulated trunkline with a contracted expansion book earns a fuller multiple than a gathering system exposed to one basin's drilling.

An AFFO yield gets quoted too, roughly 6.7% on FY2025 figures ($5,858 million against a market capitalisation near $87.5 billion). Treat it as a gross cash yield and nothing more. Because AFFO comes before capital spending, that 6.7% is not money available to shareholders; the dividend takes about 2.8 points of it and the capital programme takes the rest and then some. A free cash flow yield on the same figures is negative.

Where a discounted cash flow does better here is in giving the backlog somewhere to sit. Expansions with signed contracts and approved rates have a shape a multiple on this year's EBITDA cannot express, and the outer-year backlog is worth less than the projects already in execution because more of it will never be built.

What to Watch in the Financials

The cash statement, not the coverage ratio. Operating cash flow less capital spending less the dividend was negative $1,437 million in 2025. Whether that gap closes as projects enter service, or widens as new ones are sanctioned, is the single most informative number on the page, and no reclassification can move it.

Leverage against the capital programme. 3.71× at the end of 2025, 3.67× at the half-year, and management flagged roughly 3.75× once Momentum settles. Debt funds construction ahead of the earnings it creates, so a slipped in-service date shows up here first.

In-service dates on the projects in execution. The ~7.1 Bcf/d already under construction is where near-term growth comes from. The ~14.3 Bcf/d beyond it refreshes at each investor day as projects sanction or fall away, so check the current list before underwriting anything past 2027.

Share count. Roughly $2 billion of stock is going out for Momentum. Growth funded by issuing equity has to clear a higher bar than growth funded from retained cash, because it is diluting the holders it is meant to benefit.

Peer Context

Kinder Morgan is the closest comparison: same tax form, same gas-transmission emphasis, FY2025 adjusted EBITDA of $8,391 million. It abandoned a distributable cash flow bridge in favour of free cash flow, which is struck after all capex, and reported $2,891 million against a $2,604 million dividend. That $287 million of headroom is the same cross-check applied to Williams, and Williams is on the other side of it, because Kinder Morgan is spending far less on expansion. Kinder Morgan also separates its contract quality more carefully, at 96% take-or-pay, fee-based or hedged, of which 65 points are take-or-pay.

Energy Transfer is a partnership, files a K-1 and reports coverage of about 1.80× on partner-level distributable cash flow, against consolidated adjusted EBITDA of $15,984 million. Run the residual test there and the comfort disappears too: what it retains after distributions falls well short of its own growth budget. Williams and Energy Transfer are doing the same thing behind different metrics.

Enterprise Products is the one that funds its programme out of its own cash flow, on the lowest headline coverage of the three at 1.7× and $8,000 million of DCF. Its assets are NGL and Gulf Coast weighted rather than long-haul gas transmission, so it is not a like-for-like on the business, but it is the standard for how the cash should stack up.

Key Risks

Funding the build. The 2026 capital guide of $7.3–$7.9 billion of growth capital sits well above what the business generates after its dividend. Debt, joint-venture partners and issued stock cover the difference, and each has a cost that rises if credit markets tighten or the share price falls.

Project execution and permitting. Everything in the backlog assumes regulatory approval and on-time construction. North-eastern gas pipeline history is a long catalogue of projects that took years longer than sanctioned, or never got built at all.

Counterparty credit and recontracting. Fee-based contracts move volume risk onto whoever signed them, so what matters is who that is and for how long. Long-dated commitments from investment-grade utilities and LNG developers are the strength of the story; the weighted-average remaining contract life, and what happens when the first expansions come up for renewal at whatever the market rate is then, are the exposures worth sizing.

A deal that has not closed. Momentum is pending antitrust clearance. Until it clears, the 2026 guidance that reflects it is conditional, and the assets being bought sit in one basin rather than across a network.

Reading AFFO as if it were coverage. Williams defines AFFO itself, and 2.40× cover on a metric struck before all capital spending is not the same claim as an MLP covering its distribution 2.40 times. The distributable cash flow versus free cash flow guide covers where each of these measures draws its line.