WMB 10-K Analysis: $5B Power Pivot, 243% FCF Payout, and the EBITDA Illusion

Williams Companies, America's largest natural gas pipeline operator, reported record Modified EBITDA of $7.7 billion in FY2025 while its free cash flow collapsed 58% to $1 billion — a paradox that forces a question the earnings call doesn't answer: when a company's dividend costs 243% of its free cash flow, is the 3.3% yield a sign of confidence or a warning?

The headline numbers look strong. Revenue grew 13.8% to $11.9 billion. Operating income jumped 22.7% to $4.2 billion. WMB raised its quarterly dividend 5% to $0.525 per share in January 2026 — its second consecutive annual increase. The company's 33,000-mile Transco pipeline system, which moves roughly 30% of America's natural gas under FERC-regulated rates, generated 89% of operating income at a 68% margin. And the stock trades at $60.11 — a $73.4 billion market cap and approximately $100 billion enterprise value — reflecting the market's belief that WMB is both a regulated utility and a data center growth story.

But the 10-K reveals structural tensions invisible from the earnings transcript. The data center power contracts that anchor the bull case are 10-12.5 years — half the 20-year duration management cited in investor presentations — with zero quantified return metrics in the filing. Nearly $1 billion of the company's preferred EBITDA comes from unconsolidated joint ventures invisible to GAAP, creating a 17% valuation gap depending on which denominator you trust. And the $4.9 billion capital expenditure surge that compressed free cash flow is just the beginning: management guides $6.1-6.7 billion for FY2026, meaning the dividend will be funded by debt issuance for at least two to three more years.

The Capex Paradox — Record Earnings, Negative Free Cash Flow

Williams is reporting the best financial results in its history while bleeding cash faster than at any point in the past decade. The disconnect is arithmetic, not accounting: operating cash flow grew 18.6% to $5,898 million in FY2025, but capital expenditures nearly doubled from $2,573 million to $4,893 million — a 90.2% surge that swallowed the incremental cash generation and then some. The result was free cash flow of just $1,005 million, down 58.1% from $2,401 million in FY2024.

The FCF payout ratio tells the story in a single number. At 96% in FY2024, Williams was already spending essentially all of its free cash flow on dividends. At 243% in FY2025, the company is funding its $2.4 billion annual dividend entirely through debt — issuing $4,940 million in long-term debt during the year while simultaneously paying $2,442 million to shareholders. On a net income basis, the payout ratio actually improved from 104% to 93%, which is the number management emphasizes. But net income doesn't pay dividends — cash does.

"Williams' growth capital and investment expenditures in 2026 are expected to range from $6.1 billion to $6.7 billion, excluding acquisitions and certain long-lead time equipment for power innovation projects which are backed by reimbursement from the customer if the equipment order is cancelled."

The filing's $6.1-6.7 billion capex guidance for FY2026 — a further 25-37% increase over FY2025's already-elevated $4.9 billion — means this FCF compression is not temporary. The company will likely issue $4-5 billion in additional debt in FY2026 while carrying a $2.9 billion working capital deficit and just $63 million in cash. Total available liquidity of $3.1 billion ($63 million cash plus $3,050 million in credit facility capacity) provides adequate short-term coverage, but WMB is operating with essentially no cash buffer. Share repurchases have been suspended entirely since FY2023, with all capital directed to growth investment. The $398 million South Mansfield asset sale closed in January 2026 and will produce a gain in Q1, but it barely dents the funding need.

This is not a company in distress — it is a company making a deliberate bet. Williams Companies' free cash flow collapsed 58% to $1 billion in FY2025 as capital expenditures nearly doubled to $4.9 billion, forcing the company to fund its $2.4 billion annual dividend entirely through debt issuance at a 243% FCF payout ratio.

The EBITDA Illusion — What $965 Million in Invisible JV Income Means

The number Williams wants investors to focus on is Modified EBITDA — approximately $7,673 million in FY2025, up 15.3% from $6,656 million the prior year. It is the metric management reports in earnings releases, the denominator in covenant calculations, and the foundation of the company's valuation narrative. But Modified EBITDA exceeds GAAP EBITDA ($6,543 million) by approximately $1.1 billion, and the gap is almost entirely explained by a single accounting construct: $965 million in proportional equity-method EBITDA from unconsolidated joint ventures.

"This measure is further adjusted to include Williams' proportionate share (based on ownership interest) of Modified EBITDA from its equity-method investments, including its indirect share from interests owned by equity-method investees."

This is income that WMB partially owns but does not consolidate on its financial statements. It never flows through revenue, operating income, or net income on the GAAP income statement — it appears only as "equity earnings" of $760 million. The Modified EBITDA figure ($965 million) exceeds GAAP equity earnings ($760 million) by $205 million because it adds back depreciation within the JVs. The segment-level breakdown reveals where this invisible income lives — and the concentration is striking.

The Northeast G&P segment is the most telling example. Widely cited as a "93.5% margin cash cow," the segment reports $2,028 million in Modified EBITDA. But $640 million of that — 31.6% — comes from unconsolidated joint ventures. The remaining $1,388 million is what WMB actually consolidates. The segment is profitable, but its economic footprint is materially different from what the headline margin suggests.

The choice of denominator determines three things simultaneously: whether WMB looks expensive or fairly valued (15.3x vs 13.0x EV/EBITDA), how much covenant headroom exists (0.7 turns vs 1.27 turns at year-end), and whether the $2.8 billion January 2026 debt issuance pushes leverage into uncomfortable territory (4.10x on Modified EBITDA, with 0.90 turns of headroom against the 5.0x covenant limit — or 5.5x during acquisition periods). Williams Companies' $965 million in equity-method JV proportional income inflates its preferred Modified EBITDA by 15% relative to GAAP, creating a 17% valuation gap — 13.0x on management's metric versus 15.3x on GAAP — that determines whether the stock looks fairly valued or overpriced.

The Power Pivot Reality — 10-Year Contracts, Not 20

Williams' transformation from pure midstream transport to integrated gas-to-power infrastructure is the centerpiece of the bull case. The company has committed approximately $5.1 billion to build natural gas-fired power plants adjacent to data centers, backed by fixed-price power purchase agreements with a single "large, investment-grade company." No other midstream operator has attempted anything like this. But the 10-K tells a materially different story than management's investor presentations.

"The Socrates project involves the construction of the Socrates North and South power generation facilities in New Albany, Ohio. Williams has agreed to provide committed power generation and associated gas pipeline infrastructure for the project, which is expected to provide a combined 400 megawatts of onsite power generation capacity to the customer. The project is backed by a 10 year, primarily fixed-price power purchase agreement, with an option for the customer to extend the term of the agreement."

The Socrates project — WMB's lead Power Innovation facility — has a 10-year PPA. Apollo and Aquila are 12.5-year agreements. Socrates the Younger is 10 years. All include customer extension options, but the base contracted periods are roughly half the 20-year duration management cited in investor presentations. The filing also contains zero mention of the "5x return multiples" or "18-month paybacks" that have appeared in management presentations to investors. These return claims remain unverified against the formal filing — a conspicuous silence for a company committing over $5 billion to a new business line.

The risk mitigants are real. All four projects are backed by fixed-price PPAs with the same investment-grade counterparty. Long-lead equipment orders totaling $4.3 billion carry a customer reimbursement clause if cancelled, limiting stranded-cost risk. And the staggered delivery schedule — Socrates in late 2026, Apollo and Aquila in 2027-2028, Socrates the Younger in late 2028 — creates a visible revenue ramp rather than a single binary delivery event. WMB also invested $378 million in Louisiana LNG (a 10% equity interest) and an 80% interest in Driftwood Pipeline, extending its value chain from domestic gas transport into LNG export infrastructure.

But a 10-year contract versus a 20-year assumption reduces visible contracted revenue by approximately 8-10 years at what will become a $400+ million annual run-rate per project, shifting re-contracting risk forward to the mid-2030s — when Williams will carry $35 billion or more in debt. Williams Companies' 10-K reveals its four data center power contracts are 10-12.5 year agreements — half the 20-year duration management cited in investor presentations — with $5 billion committed but zero quantified return metrics in the filing.

The Monopoly Engine — Transco's 89% Profit Concentration

Every analysis of Williams Companies ultimately arrives at the same conclusion: this is a Transco story. The FERC-regulated pipeline system — spanning 33,000 miles from the Gulf Coast to the Northeast — generated $3,720 million in operating income on $5,443 million in revenue in FY2025, producing a 68.2% operating margin and accounting for 88.7% of the company's total operating income. No other segment comes close to this level of profitability or strategic importance.

The economic engine behind Transco's margins is fixed-cost leverage. WMB's incremental operating margin — the conversion rate of marginal revenue to marginal profit — was 53.5% in FY2025, calculated as $775 million in operating income growth on $1,447 million in incremental revenue. This means each new dollar of revenue from expansion projects or rate increases generates over 53 cents of operating profit, reflecting the infrastructure-heavy cost structure where pipelines, once built, carry gas at near-zero marginal cost.

"A $291 million increase in Transco's revenues primarily associated with expansion projects placed in service, notably Regional Energy Access in August 2024, Southside Reliability Enhancement in November 2024, Texas Louisiana Energy Pathway in April 2025, and Southeast Energy Connector in April 2025; and transportation and storage rate increases."

The $291 million in organic revenue growth from four named expansion projects placed into service during FY2024-2025 is the proof of concept for the capex-driven growth model. These are physical pipelines that now generate regulated revenue under long-term contracts — exactly the kind of asset that justifies capital spending. The Transco rate case, settled with customers during Q3 2025 and approved by FERC on December 30, 2025, becomes effective March 1, 2026, providing additional multi-year revenue visibility for the system's most profitable segment.

The risk is concentration. Gas & NGL Marketing's apparent turnaround — Modified EBITDA swinging from -$124 million to +$311 million — is almost entirely a $473 million unrealized commodity derivative swing ($138 million gain in FY2025 versus -$335 million loss in FY2024), not an operational improvement. West segment took $212 million in impairments. Strip away these volatile contributions and Transco stands essentially alone as WMB's profit engine. Any regulatory disruption — an adverse rate case outcome, a change in FERC policy, or a major environmental incident — would be existential. Williams Companies' Transco pipeline generated 89% of the company's $4.2 billion operating income at a 68% margin in FY2025, with $291 million in organic revenue growth from four expansion projects placed into service during the year.

What to Watch

At $60.11, Williams trades at 13.0x Modified EBITDA and 15.3x GAAP EBITDA — a premium to midstream peers that typically trade at 9-11x. The filing supports the growth narrative: 15.3% Modified EBITDA growth, four named Power Innovation projects with staggered 2026-2028 delivery, $291 million in organic Transco expansion revenue, and a FERC-approved rate case effective March 2026. But it complicates the income narrative: a 243% FCF payout ratio, a $2.9 billion working capital deficit, a dividend funded entirely by debt issuance, and $6.1-6.7 billion in guided FY2026 capex that will push leverage higher before it normalizes.

The investment case resolves to a single question: does the capex supercycle produce enough EBITDA growth to normalize free cash flow before debt accumulation becomes self-defeating? Credit ratings offer a partial answer — BBB+ (S&P, stable), Baa2 (Moody's, positive), BBB (Fitch, positive) — with two of three agencies on positive outlook despite the leverage increase. But the market is pricing WMB for execution, not optionality.

Five metrics define the next four quarters:

1. Socrates project delivery (Q3-Q4 2026) — the lead Power Innovation project. On-time delivery validates the capex cycle; a delay beyond Q1 2027 extends the negative FCF period by at least a year and signals execution risk across the three remaining projects.
2. Modified EBITDA trajectory toward $8.2-8.5B — observable in quarterly 10-Qs (Q1 2026 expected May 2026). Growth below 5% in FY2026 would signal the capex-to-earnings bridge is weakening. Above 10% confirms the supercycle thesis.
3. Covenant leverage at or below 4.3x — post-January 2026 leverage of approximately 4.10x leaves 0.90 turns of headroom. A move above 4.5x at any quarter-end would constrain WMB's ability to fund remaining capex without an equity raise.
4. Transco quarterly transmission service revenue above $1.3B — the rate case effective March 2026, combined with expansion projects, should produce a visible step-up from the approximately $1.2 billion quarterly run-rate. Below $1.2 billion in Q2 2026 would signal the rate settlement included a net decrease.
5. Dividend action — management has raised the quarterly dividend twice (to $0.500 then $0.525) during the capex surge. A third increase in January 2027 would signal genuine confidence that the FCF bridge is holding. A freeze would be the first warning sign.

At $60.11, the market implies WMB's Modified EBITDA will grow 5-6% annually for five years. The filing supports that assumption — but with FCF deeply negative through 2027-2028, the 3.3% yield is a leveraged bet on the capex cycle, not a cash-flow-covered dividend. If Power Innovation projects deliver on schedule and capex normalizes post-2028, free cash flow recovers to $3-4 billion and the current price is justified. If any project slips materially, the debt pile grows while EBITDA growth decelerates — and the 99.4x EV/FCF multiple has no floor.

Frequently Asked Questions

Is Williams Companies' dividend safe?

WMB's $2.4 billion annual dividend represents 243% of its $1.0 billion FY2025 free cash flow, meaning the dividend is entirely funded by debt issuance. On a net income basis, the payout ratio improved from 104% in FY2024 to 93% in FY2025. Management raised the quarterly dividend 5% to $0.525 per share in January 2026, signaling confidence. Dividend safety depends on whether the $6.1-6.7 billion annual capex program generates enough incremental EBITDA to normalize FCF by 2028-2029. The most likely threat is not a dividend cut — management has demonstrated strong commitment with two consecutive 5% increases — but rather prolonged debt accumulation that compresses the stock's valuation multiple.

What are WMB's data center power projects and when do they start generating revenue?

Williams has four named Power Innovation projects: Socrates (two power plants totaling 400 MW in New Albany, Ohio, expected Q3-Q4 2026), Apollo (Ohio, H2 2027), Aquila (Utah, H2 2027/H1 2028), and Socrates the Younger (Ohio, H2 2028). All are backed by primarily fixed-price power purchase agreements with a single "large, investment-grade company." The projects provide onsite gas-fired power generation capacity for data centers. Total committed capital is approximately $5.1 billion, with $4.3 billion in outstanding PP&E commitments at December 2025. Long-lead equipment orders are backed by customer reimbursement if cancelled, reducing stranded-cost risk.

Why is Williams' EV/EBITDA higher than midstream peers?

WMB trades at 13.0x EV/Modified EBITDA or 15.3x EV/GAAP EBITDA, compared to roughly 9-11x for major midstream peers like KMI, OKE, and EPD. The premium reflects three factors: (1) Transco's FERC-regulated monopoly position generating 89% of operating income at 68% margins; (2) the data center gas-to-power strategy, which is unique among midstream companies and provides optionality on AI and data center demand growth; and (3) strong revenue growth of 13.8% in FY2025, compared to low-single-digit growth at most peers. However, $965 million (12.6%) of Modified EBITDA comes from proportional JV income that other companies may not include in comparable metrics, making apples-to-apples peer comparison difficult.

What is the difference between Williams' Modified EBITDA and GAAP EBITDA?

Williams' Modified EBITDA of approximately $7,673 million in FY2025 exceeds its GAAP EBITDA of $6,543 million by approximately $1.1 billion. The gap is almost entirely explained by $965 million in proportional equity-method EBITDA from unconsolidated joint ventures — income from JVs that WMB partially owns but does not consolidate on its financial statements. By segment: $640 million from Northeast G&P, $147 million from Transmission, $142 million from West, and $36 million from Gas & NGL. This JV income appears in the GAAP income statement only as "equity earnings" ($760 million), not as operating revenue. Modified EBITDA exceeds equity earnings by $205 million because it adds back depreciation within the JVs. This is the metric used in covenant calculations and management guidance.

How close is Williams to breaching its debt covenants?

WMB's credit facility requires a debt-to-EBITDA ratio no greater than 5.0x (or 5.5x during acquisition periods). At December 2025, the covenant ratio was approximately 3.73x ($28.7 billion total debt divided by $7.7 billion Modified EBITDA, subject to credit agreement definition), providing 1.27 turns of headroom. After the $2.8 billion January 2026 debt issuance, the ratio rises to approximately 4.10x, leaving 0.90 turns of headroom. With $6.1-6.7 billion in FY2026 capex guidance, WMB will likely add $3-4 billion in net debt during 2026. Staying within covenants requires Modified EBITDA to grow toward $8.2-8.5 billion — achievable if Power Innovation projects and Transco expansion deliver on schedule, but with limited margin for error.

What happened to the 20-year contracts Williams cited for data center power?

Management investor presentations cited 20-year contracts for data center power projects. The 10-K tells a different story: Socrates has a 10-year primarily fixed-price PPA, Apollo and Aquila have 12.5-year agreements, and Socrates the Younger has a 10-year contract. All include customer options to extend the term. The discrepancy is significant — it halves the base contracted revenue duration and introduces re-contracting risk in the mid-2030s rather than the mid-2040s. The 10-K also contains zero mention of the "5x return multiples" or "18-month paybacks" that management has cited in investor presentations. These return claims remain unverified against the formal filing.

Why did Williams' free cash flow collapse 58% despite record revenue?

FY2025 operating cash flow grew 18.6% to $5,898 million, but capital expenditures nearly doubled from $2,573 million to $4,893 million (+90.2%), resulting in free cash flow of just $1,005 million — down 58.1% from $2,401 million in FY2024. The capex surge reflects the beginning of WMB's largest-ever infrastructure build cycle: Power Innovation projects, Transco pipeline expansions, and bolt-on acquisitions totaling approximately $1.0 billion. Management guides $6.1-6.7 billion in growth capital for FY2026, suggesting FCF will remain compressed or negative for at least two to three more years. The 3.3% dividend yield is funded by $4.9 billion in annual debt issuance, not operating cash flow.

How does the Transco rate case settlement affect Williams?

Transco filed a general rate case with FERC in August 2024 seeking an overall increase in rates. During Q3 2025, Transco reached a settlement agreement with customers, which FERC approved on December 30, 2025. New rates become effective March 1, 2026. The settlement includes a rate refund liability that Transco has accrued. The net effect should be positive: rate cases for FERC-regulated pipelines typically result in rate increases that reflect the higher asset base from expansion projects placed in service. Transco's FY2025 transmission service revenue of $4,826 million should see a step-up from the new rate schedule, providing multi-year revenue visibility for the company's most profitable segment.

Should investors use GAAP or Modified EBITDA to value Williams?

Both metrics are informative, but for different purposes. For covenant analysis and management guidance, Modified EBITDA (~$7.7 billion, 13.0x EV/EBITDA) is the relevant metric — it is what the credit facility uses and what management reports against. For peer comparison, GAAP EBITDA ($6.5 billion, 15.3x) provides a more conservative and comparable baseline, because other midstream companies may define their adjusted EBITDA differently. The 17% gap creates a "denominator effect" where bulls citing 13.0x and bears citing 15.3x are both technically correct. The resolution: understand that $965 million (12.6%) of Modified EBITDA comes from proportional JV income and decide whether that income stream is durable and cash-generative enough to warrant full EBITDA credit.

What would cause Williams to cut its dividend?

The most likely catalyst for a dividend cut would be a covenant breach or near-breach that forces capital allocation prioritization. At 4.10x covenant leverage (post-January 2026) versus a 5.0x limit, WMB has 0.90 turns of headroom. A scenario where FY2026 Modified EBITDA growth stalls below 5% while $6.5 billion in capex adds $4 billion or more in net debt could push leverage toward 4.8-5.0x, forcing management to choose between the dividend ($2.4 billion annually) and growth investment. A Power Innovation project delay or cancellation could also trigger reassessment, though the customer reimbursement backstop on equipment orders limits the capital at risk. Management is more likely to issue equity or slow capex than cut the dividend.

What is Williams doing in LNG exports?

In 2025, WMB invested $378 million to acquire a 10% equity interest in Louisiana LNG (a fully permitted LNG export facility) and an 80% interest in Driftwood Pipeline. WMB will operate the pipeline, manage gas supply for the LNG facility, and purchase approximately 10% of the LNG produced. This creates a vertically integrated chain from wellhead gathering through Transco transport to LNG export — a strategic diversification beyond the core pipeline business. The investment is early-stage and will not contribute materially to earnings for several years, but it extends WMB's business model from domestic gas infrastructure into international energy markets.

How does Williams' segment rename to 'Transmission, Power & Gulf' matter?

The addition of "Power" to the segment name signals that management views power generation as a core, permanent business line — not an experimental side project. Power generation revenue and costs will be reported within the Transmission segment (the highest-margin, FERC-regulated segment), which could blend the power business's economics with legacy pipeline economics. If power generation carries lower margins than regulated gas transport — which is likely, given it is unregulated and capital-intensive — the blended segment margin could decline even as revenue grows, potentially masking the power pivot's actual economics. Investors should watch whether Power Innovation project margins are disclosed separately or are subsumed within the 68% Transmission segment margin.

Methodology

Data Sources

This analysis is based on The Williams Companies' FY2025 Annual Report (10-K) filed February 24, 2026 with the SEC (CIK: 0000107263). Quantitative data was sourced from the MetricDuck metrics pipeline, which processes SEC XBRL filings, and from filing text extraction of MD&A, risk factors, and financial statement footnotes. Management presentation claims (20-year contracts, 5x return multiples, 18-month paybacks) are sourced from public investor conferences and presentations and are flagged as unverified against the 10-K where applicable.

Limitations

• Maintenance versus growth capex is not disclosed. The filing does not break out maintenance capital expenditure, which limits the ability to calculate "true" free cash flow using only maintenance capex. The 243% FCF payout ratio uses total capex ($4,893 million).
• Covenant EBITDA definition carries uncertainty. The filing references "ratio of debt to EBITDA no greater than 5.0 to 1.0" but does not reproduce the exact credit agreement definition of EBITDA. The covenant leverage calculations assume Modified EBITDA ($7,673 million) is the relevant denominator, which may differ from the precise credit agreement measure.
• Power Innovation project-level economics are not disclosed. Per-project costs, returns, and margin profiles are not available in the filing. Total committed capital ($5.1 billion) is cited from management presentations, not from the 10-K.
• No peer pipeline data. Peer comparisons reference approximate industry benchmarks, not MetricDuck pipeline-extracted financials. Peer-specific numbers (KMI, OKE, ENB, EPD EV/EBITDA multiples of 9-11x) are directional estimates and should not be cited as precise values.
• Forward projections are estimates. Filing Preview projections are based on guided capex, disclosed project timelines, and trailing growth rates. They are testable predictions, not forecasts.
• Stock price as of filing date. Valuation metrics use the $60.11 price from pipeline data at the time of the filing. The valuation framework is structural, not a time-sensitive price target.

Disclaimer:

This analysis is for informational purposes only and does not constitute investment advice. The author does not hold positions in WMB. Past performance and current metrics do not guarantee future results. All data is derived from public SEC filings and may contain errors or omissions from the automated extraction process. Forward-looking statements represent analytical projections based on filing data and should not be relied upon as predictions. Numbers tagged as derived should be verified against the original 10-K before use in investment decisions.