Self-Help

What this tab establishes

Plains can comfortably outlast its problem: investment-grade ratings, no senior note maturity between 2027 and 2028, a 5.00x covenant against 4.1x actual leverage, and $3.3 billion of divestiture cash received in May 2026. What it does not have is a repurchase engine. Cash spent on buybacks over the last three years totals $8 million; the unit count has risen since January 2023; and management ranks repurchases last among four uses of capital.

Debt and Maturities

At 31 December 2025 total debt stood at $11,259 million against $329 million of cash — up from $7,618 million a year earlier, the increase funding the EPIC Crude and Cactus III purchases [1]. The stack is $9,118 million of PAA senior notes net of discounts, $970 million of commercial paper, a $1,099 million term loan and $72 million of other borrowings [1].

The maturity schedule is the reason the balance sheet is not the binding constraint. The debt footnote gives the next five years and a residual; the company's own investor deck breaks the residual out year by year.

Sources: FY2025 Annual Report (Form 10-K), Note 11 Debt — Maturities, which gives 2026 $750M, 2027 nil, 2028 nil, 2029 $1,000M, 2030 $750M and $6,683M thereafter [2]; the post-2030 breakout is from the May 2026 investor presentation [3].

The weighted average maturity of the senior notes is approximately ten years and the weighted average coupon approximately 5.0% [2] [3]. Only one senior note matures before 2029: the $750 million 4.50% issue due December 2026, which management has said will be repaid from divestiture proceeds rather than refinanced [8].

Covenants. The credit agreements and the term loan require a trailing-four-quarter debt-to-EBITDA ratio no greater than 5.00 to 1.00, widening to 5.50 to 1.00 for three quarters following an acquisition above $150 million; letters of credit and hedged-inventory borrowings are excluded from the calculation. The agreements prohibit distributions and unit repurchases only while a default is continuing — so long as Plains is in compliance, its ability to distribute available cash is unrestricted. Plains was in compliance at year-end 2025 [2]. Against that 5.00x limit, the reported leverage ratio at 31 March 2026 was 4.1x on a definition that already gives the preferred units 50% debt treatment, versus a long-term target range of 3.25x to 3.75x, with Fitch at BBB, Standard and Poor's at BBB and Moody's at Baa2 [3].

Liquidity. At year-end 2025, availability under the $1,350 million senior unsecured revolver plus $1,298 million under the senior secured hedged inventory facility, less $970 million of commercial paper outstanding, left $1,678 million of undrawn capacity; with cash, total liquidity was $2,007 million [4]. The revolver runs to August 2029 and the inventory facility to August 2027 [5].

Refinancing at current rates. New money in 2025 was priced at 4.70% for 2031 paper, 5.60% for 2036 paper and 5.95% for the 2035 issue [6]. The notes coming due next carry lower coupons — 3.55% in 2029 and 3.80% in 2030 — so refinancing $1.75 billion at roughly 125 basis points more costs about $22 million a year against adjusted EBITDA attributable to Plains of $2,833 million in 2025. Cash interest paid was $377 million in 2023, $351 million in 2024 and $431 million in 2025 [9].

The $1.1 billion term loan drawn in December 2025 to retire assumed EPIC debt matures November 2027 and carried a mandatory prepayment trigger on the closing of the Canadian NGL divestiture [5]. That sale closed on 12 May 2026 for approximately $3.3 billion of net cash proceeds, earmarked to repay indebtedness and expected to bring leverage toward the middle of the target range [7].

On the first half of the question — can the company outlast the problem — the answer is unambiguous. On the second half, the allocation of that headroom is the constraint, and it is a chosen one rather than an imposed one.

Repurchases Executed

Ruchir's test is cash actually spent, not authorisations announced. The Board approved a $500 million common equity repurchase programme in November 2020, covering PAA common units and PAGP Class A shares interchangeably. Five and a half years later, $190 million of that authorisation remained unused at 31 December 2025 [10].

Sources: repurchase cash from the Consolidated Statements of Cash Flows in the FY2021 [11], FY2023 [12] and FY2025 [9] Form 10-Ks; PAA common unit distributions from the FY2021 [15] and FY2025 [14] Form 10-Ks.

The record year by year: $50 million in 2020 and $178 million in 2021 [11], $74 million in 2022 and nil in 2023 [12], nil in 2024 and $8 million in 2025 [9]. Total spend across six years is $310 million. The deterministic feature file records the same two most recent figures — nil for FY2024 and $8 million for FY2025 (fit_features.share_count_trend.buyback_cash_per_year).

Prices paid were good. The 2020 tranche was 6,222,748 units for $50 million, or $8.03 per unit; the 2021 tranche 18,061,583 units for $178 million, or $9.86 [13]. The 2025 tranche was approximately 0.5 million units for $8 million, near $16 [10]. Against a 31 July 2026 close of $26.33, every unit bought back was bought cheap. The buying simply stopped: over 2023, 2024 and 2025 combined, $8 million went to repurchases while $2,709 million went to common distributions.

One repurchase of size did happen in the period, and it was not of common units. In January 2025 PAA bought back approximately 12.7 million Series A preferred units — 18% of the class — from EnCap Flatrock Midstream, an entity affiliated with a member of the board, at the $26.25 issue price for $333 million plus $10 million of accrued distributions [10] [27]. At the 9.375% reset rate that is $31 million of annual preferred distributions retired for $333 million — a defensible use of cash, and 42 times what the common got that year.

The unit count

PAA does not publish a units-outstanding roll-forward in the PAGP filings, but each 10-K states the January-declared quarterly distribution in both dollars and dollars per unit, which pins the count at a consistent record date every year.

Derived as the declared fourth-quarter distribution divided by the per-unit rate, from each Form 10-K: $127M at $0.18 (Jan 2022) [15], $187M at $0.2675 (Jan 2023) [16], $223M at $0.3175 (Jan 2024) [17], $267M at $0.38 (Jan 2025) [18] and $295M at $0.4175 (Jan 2026) [14]. Rounding of the dollar totals gives roughly plus or minus one million units per point.

The shape is a shallow V. Distributions paid on a full-year basis imply about 727.5 million units in 2019 and 727.8 million in 2020 [15]; the $302 million of repurchases in 2020 to 2022 took the count to a trough of 699.1 million at the January 2023 record date; it has risen every year since, to 706.6 million in January 2026. That is 7.5 million units added, 1.07% over three years, or roughly 0.36% a year.

The driver is equity compensation, not acquisition currency. Equity-indexed compensation expense ran $51 million in 2023, $52 million in 2024 and $50 million in 2025 [9]. Ruchir's framework treats a share count that keeps rising on stock compensation or serial M&A as disqualifying, and on the letter of that test the count is rising. Two facts sit against the severity of the reading, and both belong in it: the drift is 0.36% a year, an order of magnitude smaller than the cases the rule was written for; and the acquisitions have not been paid for in units. Plains has made 17 bolt-on acquisitions since 2022 for approximately $4.3 billion of cumulative net investment [19] and funded them with debt and divestiture proceeds; the 10-K states plainly that in the near term the company does not plan to issue common equity to fund such expenditures [31].

What is absent is the offset. A programme that retires 0.36% of the count a year would neutralise the drift for roughly $67 million annually at the current price. The last three years produced $8 million.

Where the Next Dollar Goes

Brokers have asked the question on every recent call, and management's answer has been consistent and explicit about ordering. On the Q1 2026 call the CFO set it out in full:

On capital allocation, with the proceeds from NGL, we anticipate paying down a little over $3 billion of debt, which would be the term loan, the outstanding CP we have, and a $750 million note that matures later this year. Post that, we expect to be right at the midpoint of our leverage range, about 3.5x… Our capital allocation priorities remain: maintaining distribution growth; funding investments, whether organic or M&A-related; taking out preferreds should leverage remain at or below the bottom end of the range; and opportunistic share repurchases.

— Al Swanson, CFO, Q1 2026 earnings call, 8 May 2026 [8]

Repurchases are fourth of four, and gated twice — behind roughly $3 billion of debt reduction, and then behind a leverage condition of at or below 3.25x that the company does not expect to satisfy before it has first funded distribution growth and investment. Three months earlier, asked directly whether the priorities had shifted, the same executive gave the same ranking: "Our primary way of returning cash to shareholders is going to be through distribution growth… Secondly, we do have some preferred securities as well as common unit repurchases. Those will be more on an opportunistic basis" [21]. And on the Q3 2025 call, asked what happens once leverage is back at the midpoint: "quite honestly, you're sitting at the midpoint of the leverage range and still seeing potential opportunities to deploy capital with good return, we'll be more biased towards looking at the bolt-ons at that point" [22].

The company's own materials rank it the same way. The May 2026 deck lists four levers to drive growth — growth capex, the $100 million cost programme, bolt-on M&A at 13% to 15% IRR, and last, "Capital Optimization: Preferreds (~$2.3 Bln) / Opportunistic buybacks" [20].

This is not an accident of the current cycle. The stated preference for other uses has been stable for years: in November 2023, asked whether the preferreds might be taken out, the CFO said "there's been no change in our thinking around the preferreds… we won't sacrifice our financial flexibility," pegging the weighted average cost of capital at 11% to 12% [32].

Insider buying. The Form 4 record holds 127 transactions between May 2023 and August 2025, of which two are open-market purchases, both by the Chairman and CEO: 75,000 shares at $13.247 on 8 May 2023 for $994 thousand, and 62,000 shares at $16.237 on 28 September 2023 for $1.007 million [25]. No insider bought in the open market during the January-to-October 2025 drawdown.

The Levered Exception

Ruchir's framework tolerates leverage only where the adjusted yield is extreme — roughly 25% to 40% — and where two further conditions hold alongside it. The company qualifies as levered: net debt of $10,930 million against 2025 adjusted EBITDA attributable to Plains of $2,833 million is 3.86x, and the company's own reported ratio was 4.1x at 31 March 2026 [1] [3]. All three legs have to compute together, and two do not.

Leg 1 derived from reported figures as set out below; leg 2 from the January-declared distributions in the FY2021 to FY2025 Form 10-Ks [14]; leg 3 from reported free cash flow and revenue, fit_features.adjusted_fcf.series and fit_features.revenue_trajectory.per_year.

Two of three is not the exception. The yield gap is the decisive part of the arithmetic — the levered bar sits at roughly three times where the adjusted yield computes, and no reasonable treatment of the acquisition deduction closes that distance.

Years to Retire the Float

The feature file cannot answer this one. fit_features.float_retirement_years is null, because market_cap is null ("no positive annual period-end or outstanding share count") and adjusted_fcf.latest_adjusted is null ("missing SBC for FY 2016 through 2025; no complete consecutive five-year acquisition window with SBC"). Both inputs are in fact in the filings — the cash-flow statement reports equity-indexed compensation expense of $50 million and cash paid for acquisitions of $2,651 million in 2025 [9], and 197,904,124 PAGP Class A shares were outstanding at 20 February 2026 [30] — so the gap is one of data plumbing rather than disclosure. The arithmetic below is computed from those filed figures and is not the deterministic feature; it is logged as a data gap.

The float that matters is the whole common-unit stack, because the cash flow being measured is PAA's. At the 31 July 2026 close of $26.33, 706.6 million PAA common units are worth $18,605 million.

Adjusted FCF = reported FCF minus stock-based compensation minus the trailing five-year average of acquisition spend; reported FCF of $2,288M is the feature-file figure (fit_features.adjusted_fcf.series, FY2025), equal to operating cash flow of $2,931M less $643M of continuing-operations capital additions. Compensation, acquisitions (2021 $32M, 2022 $149M, 2023 $425M, 2024 $248M, 2025 $2,651M, averaging $701M) and the noncontrolling-interest distribution split are from the FY2023 [12] and FY2025 [9] Form 10-K cash-flow statements, the PAA common distribution table [14] and the preferred distribution tables [27].

The JV-minority deduction is the residual of the $1,441 million paid to noncontrolling interests in 2025 after the $716 million to public common unitholders, $53 million to AAP's outside partners and $225 million to the two preferred classes — $447 million leaking to the 35% partner in the Permian JV, the 30% partner in Cactus II and the 33% partner in Red River.

So: $18,605 million of common equity divided by $1,537 million of adjusted FCF is 12.1 years, and divided by the $875 million that actually reaches common holders, 21.3 years. The 2025 acquisition figure is an outlier that pulls the five-year average up; substituting the 2021 to 2024 average of $214 million raises adjusted FCF to $2,024 million and the common-available figure to $1,362 million, giving 9.2 and 13.7 years respectively. Ruchir's absurdity threshold is around three years. The range here is nine to twenty-one.

Distribution Safety

At $1.67 per share annualised against the 31 July 2026 close of $26.33, the distribution yield is 6.3% [10]. At that level the yield is most of the near-term return, so it carries the full test.

Source: May 2026 investor presentation, Increasing Return of Capital to Equity Holders; 2026 is guidance furnished 8 May 2026 [28].

Coverage. On the company's distributable-cash-flow definition, coverage has fallen from about 270% in 2022 to a guided 160% for 2026, and the threshold itself was cut from 160% to 150% in February 2026 alongside a $0.15 per unit, 10% distribution increase [29]. On the framework's adjusted-FCF basis the picture is tighter: the $1,070 million paid to common unitholders in 2025 against $875 million of adjusted FCF available to common is 0.82x cover, and against the M&A-normalised $1,362 million, 1.27x. In a year with $2,651 million of acquisitions and a $3.3 billion divestiture pending, part of the distribution was funded by the balance sheet. Headroom exists on the company's own measure; it is thin on Ruchir's.

The record through a downturn is a cut. Distributions per common unit paid were $1.38 in 2019, $0.90 in 2020 and $0.72 in 2021 — a 48% reduction across the 2020 collapse [15]. The 2025 rate of $1.52 is still 10% below the 2019 level six years later. A yield case here has to underwrite a distribution that was halved once inside the last six years, and the coverage buffer is the only thing standing between the payout and a repeat.

Commitment language and what would force a cut. Management has named distribution growth its primary return-of-capital vehicle on every recent call and has delivered the $0.15 per unit annual step for four consecutive years [21] [28]. The covenants do not threaten it: distributions are unrestricted while Plains is in compliance, and 4.1x sits well inside the 5.00x limit [2]. What would force a cut is a sustained EBITDA decline large enough to push leverage toward the covenant or to threaten the BBB rating — the 2020 pattern, when volume and margin collapsed together. The coverage buffer absorbs roughly a 37% fall in distributable cash flow before the 2026 distribution is uncovered on the company's own definition.

Promise Against Delivery

The sample is the five most material forward commitments made two to four years ago, checked against what happened.

Sources: Q4 2021 earnings call, 9 February 2022 [23]; Q2 2024 earnings call, 2 August 2024 [33]; Q4 2024 earnings call, 7 February 2025 [34]; 3Q25 earnings presentation, 5 November 2025 [35]; distribution and coverage history from the May 2026 investor presentation [28] and the February 2026 earnings deck [29].

The pattern is legible. Plains is reliable on what it controls — leverage targets, deal execution, the cost programme, the distribution cadence — and unreliable on basin volume forecasts and on the last mile of its own EBITDA range, having missed a range it had narrowed one quarter earlier. On the repurchase half of the February 2022 commitment, the follow-through was absent. Asked on that same call how the 25% equity-holder allocation would split between distributions and buybacks, the CEO answered that "our goal is to get our leverage down. But once that happens, it gives a significant amount of capacity to return to unitholders" [24]. The leverage did come down; the capacity went to distributions.

That is a delivery record, not a promotional one. The distinction Ruchir draws is between managers who talk their book without owning it and managers whose forecasts are simply hard. The ownership evidence points firmly to the second. As of 23 March 2026 the named executive officers held approximately 10.9 million units of PAA and PAGP equity worth about $238 million, and all executive officers and directors together held approximately 57.8 million units worth about $1.3 billion; each named executive held securities worth more than 14 times base salary at year-end 2025; and hedging, pledging and margin-account holdings of company securities are prohibited outright [26]. The CEO added $2.0 million of stock with his own money in 2023 [25]. CEO total compensation of $12.3 million against a 91:1 pay ratio is unremarkable for the asset base. Checked across the transcript archive and the proxy, the promotional-management exclusion does not fire here.

What would change this read

Three things, in order of how quickly they would show. A repurchase authorisation increase with cash behind it once leverage reaches the bottom of the target range would convert the fourth priority into a real one — the existing $190 million of unused capacity is the tell to watch [10]. A unit count that turns down for two consecutive January record dates would answer the share-count condition on evidence rather than intent. And an adjusted yield reaching the framework's levered bar would require either a price roughly two-thirds below today's or adjusted FCF available to common roughly tripling — neither is in the consensus path shown in Yield.

The counter-case is worth stating in the same breath: on the company's own measures the balance sheet is stronger than at any point in this record, the divestiture has landed $3.3 billion of cash, and the distribution has compounded at 19% a year for four years. An investor who wants a covered 6.3% yield from an investment-grade crude midstream is being offered exactly that. An investor who wants the repurchase flywheel is being offered the fourth line item.