Yield

Bottom line

On the framework's basis — reported free cash flow less stock compensation less the five-year average of acquisition spend — PAGP's FY2025 adjusted FCF is $1,537 million consolidated, and $865 million after the preferred and joint-venture minority claims that sit ahead of the common. That is $1.23 per look-through PAA unit, a 4.7% yield at $26.33. Net debt of 3.9x EBITDA places the name on the 25% levered reference line.

What the adjustment removes

The framework's yield basis is adjusted FCF: reported free cash flow, minus stock-based compensation, minus the trailing five-year average of acquisition spend. fit_features.adjusted_fcf returns not_computable for every year of this run — the structured cash-flow feed behind it carries no stock-compensation or acquisitions field. The reported-FCF series in that feature file is populated, and it reconciles exactly to the filed statements: FY2025's $2,288 million is $2,931 million of net cash from operating activities less $643 million of additions to property, equipment and other [1]. The two missing rows come from the same filed pages.

Stock compensation appears as "Equity-indexed compensation expense": $50 million in FY2025, $52 million in FY2024, $51 million in FY2023 [1], $41 million in FY2022 and $24 million in FY2021 [2], $16 million in FY2020 and $35 million in FY2019 [3]. At roughly 2% of operating cash flow, this is the smallest of the three deductions — a partnership pays its people mostly in cash.

Acquisition spend is the line that moves the number. "Cash paid in connection with acquisitions, net of cash acquired" ran $50 million (FY2019), $310 million (FY2020) [3], $32 million (FY2021), $149 million (FY2022) [2], $425 million (FY2023), $248 million (FY2024) and $2,651 million (FY2025) [1]. The FY2025 figure — the EPIC Crude and Cactus III transactions — is larger than the prior six years combined, and it drags the trailing five-year average to $701 million.

Adjusted FCF = reported FCF − stock-based compensation − trailing five-year average acquisition spend; derived from company filings. Sources: FY2025 10-K, Consolidated Statements of Cash Flows [1]; FY2022 10-K [2]; FY2021 10-K [3].

Two honest limitations sit on that table. The acquisition window is incomplete before FY2023: the corpus holds PAGP 10-Ks back to FY2021 only, so FY2019 through FY2022 use one-, two-, three- and four-year averages rather than five, and their adjusted figures are correspondingly generous. And the reported-FCF series in fit_features pairs total-company operating cash flow with continuing-operations capital expenditure for FY2023–FY2025, because the FY2025 filing restated capex for the pending Canadian NGL divestiture while leaving operating cash flow on a total-company basis. Adding back the discontinued-operations investing outflows of $162 million, $170 million and $197 million [1] lowers adjusted FCF to $1,908 million, $1,581 million and $1,340 million for those three years — a 13% cut to FY2025. Every yield below is computed on the fit_features basis; the alternate is carried through in the sensitivity at the end of the yield section.

Whose cash flow it is

PAGP is not the operating company. It "does not directly own any operating assets"; its sole source of cash is an indirect investment in Plains All American Pipeline, and at 31 December 2025 it held an approximate 85% limited partner interest in Plains AAP through 197.9 million AAP units [4]. AAP in turn owns roughly 233.0 million PAA common units, about 31% of PAA's common units and Series A preferred units combined [4]. PAGP nonetheless consolidates 100% of PAA. Every cash-flow figure above is PAA's.

Dividing $1,537 million of consolidated adjusted FCF by PAGP's own market capitalisation — 197,904,124 Class A shares [5] at the $26.33 close of 31 July 2026, or $5.21 billion — produces 29.5%. That number is an artifact of consolidation, not a yield, and it is what a screen run on this ticker returns.

The look-through arithmetic is exact, and the distribution record proves it. PAA paid $1.5200 per common unit in 2025, of which $354 million went to AAP; AAP passed $301 million of that to PAGP [6], and PAGP paid its Class A shareholders $301 million, also $1.5200 per share [7]. One Class A share carries the economics of one PAA common unit. So the correct denominator is the whole common-unit base — $1,070 million of total 2025 common distributions at $1.5200 implies 703.9 million units [6] — and the numerator has to clear two claims that rank ahead of the common:

• Preferred distributions. $154 million to PAA's Series A holders and $71 million to Series B in 2025 [8].
• Consolidated joint-venture minorities. $447 million paid out to the 35% Permian JV partner, the 30% Cactus II partner and the 33% Red River partner in 2025 [6]. Their cash is inside consolidated operating cash flow; none of it belongs to PAA's common.

FY2025: $1,537m adjusted FCF − $225m preferred − $447m joint-venture minorities = $865 million, or $1.229 per unit on 703.9 million units.

The yield, three ways

Preferred and joint-venture minority distributions from the FY2025 10-K [6] [8], FY2022 10-K [9] [10] and FY2021 10-K [11]; unit counts derived as total common distributions divided by distribution per unit; year-end prices from the daily price feed.

Current. $1.229 per unit against the $26.33 close of 31 July 2026 is 4.67%. On the fully-exchanged share base — 197.9 million Class A plus the 35.1 million Class B shares exchangeable one-for-one into Class A [12] — the per-unit figure is unchanged, because the exchange is one AAP unit for one Class A share and the look-through unit count does not move.

Three-year average. FY2023–FY2025 per-unit adjusted FCF of $2.140, $1.528 and $1.229 averages $1.632. On today's price that is 6.20%.

The company's own baseline. Measured at each fiscal year-end price, the seven readings run 7.5%, 6.0%, 17.7%, 14.6%, 13.4%, 8.3% and 6.4%. The median of the six years through FY2024 is 10.9%.

Derived: adjusted FCF attributable to PAA common, per common unit, divided by the last close on or before each fiscal year-end. Sources: filed cash-flow statements [1] [2] [3] and distribution notes [6].

The framework's jump test asks whether a formerly stable yield has doubled off its own baseline. Here the movement runs the other way. At 4.67% the current reading is 0.43x the six-year median and the lowest of the seven, and the price that produced it is a high, not a low: fit_features.capitulation_gauge records a January 2025 peak of $22.13, an October 2025 trough of $16.85 — a 23.9% drawdown — and a current close of $26.33, which is 19% above the pre-drawdown peak and 56% above the trough. The compression is mostly price. On FY2023's per-unit figure of $2.140, today's price would still yield 8.1%; the fall from 13.4% to 4.7% is roughly 55% price and 45% a smaller numerator.

Sensitivity. On the alternate FCF basis that charges discontinued-operations investing outflows against total-company operating cash flow, FY2025 adjusted FCF is $1,340 million, $668 million to the common, $0.949 per unit — a 3.60% yield. Charging the FY2021–FY2024 average acquisition spend of $214 million instead of the $701 million five-year average lifts the FY2025 figure to $1.921 per unit — 7.30%. The plausible band on today's price is roughly 3.6% to 7.3%, with the framework's own convention at 4.67%.

Which bar applies

fit_features.balance_sheet_class returns unknown — the structured feed's balance sheet reports only $1,099 million of long-term debt for FY2025, which is the term loan alone, not the debt stack. The filed Note 11 gives the full figure: total debt of $11,259 million at 31 December 2025, comprising $563 million short-term and $10,696 million long-term [13], against $329 million of cash [5].

Net debt = $11,259m − $329m = $10,930 million. Adjusted EBITDA attributable to PAA for FY2025 was $2,833 million [14]. Net debt / EBITDA = 3.86x, above the framework's 3.0x levered threshold, so the reference line is 25%.

The company's own measure agrees and then some: crediting preferred equity at 50% debt treatment, it printed a 3.9x leverage ratio at 31 December 2025 against a 3.25x–3.75x long-term target range [15], and 4.1x pro forma at 31 March 2026, expecting roughly 3.5x once the Canadian NGL divestiture closes [16]. The deleveraging is real but it does not change the classification: at ~3.5x the name stays on the levered line.

Position against the bar, in plain arithmetic:

• Against the 25% levered line that the balance sheet selects: 4.67% is 2,033 basis points short.
• Against the 10% default line, had the balance sheet been moderate: 533 basis points short.
• Against the 8–9% fortress line: 333 to 433 basis points short.

No classification rescues the position. The three-year average of 6.20% and the most generous sensitivity of 7.30% are also below all three lines.

The framework's float-retirement check points the same way. The consolidated common equity — 706.6 million PAA common units at the January 2026 declared distribution of $0.4175 per unit on a $295 million total [6] — is worth $18.6 billion at $26.33. Against $865 million of adjusted FCF to the common, retiring the float takes 21.5 years. The framework's absurdity marker is roughly three. The executed repurchase record is covered in Self-Help; the cash number is $8 million of common units bought back in 2025, with $190 million of authorisation left [7].

Normalized mid-cycle yield

Revenue swings violently here — down 30.8% in FY2020, up 80.7% in FY2021, down 9.5% in FY2025 — but almost all of that is commodity price running through PAA's buy/sell arrangements, not volume or margin. The cash-earnings line is far steadier. Adjusted EBITDA attributable to PAA ran $3,227 million in FY2019, $2,546 million in FY2020, $2,196 million in FY2021 [17], $2,510 million in FY2022, $2,711 million in FY2023 [18], $2,779 million in FY2024 and $2,833 million in FY2025 [14]. Peak to trough over seven years is 32%, and the trough was the 2020–21 demand collapse.

That history says FY2025 is not a cyclically depressed year. The seven-year mean of $2,686 million sits 5.2% below the FY2025 actual; restricting to the post-COVID window FY2022–FY2025, the mean of $2,708 million sits 4.4% below. A mid-cycle normalization on this business lowers the yield rather than raising it.

The normalization, with every assumption stated so it can be recomputed under an adjacent window:

1. Base. 2026 guided Adjusted EBITDA attributable to PAA of $2,880 million (±$75 million) [16]. This is the first full year carrying the EPIC Crude and Cactus III assets acquired in 2025, so it is the only base that is like-for-like with the current asset footprint.
2. Cycle haircut. Apply the 5.2% gap between the FY2019–FY2025 mean and the FY2025 actual: $2,880m × 0.948 = $2,730 million mid-cycle EBITDA. The FY2022–FY2025 window instead gives 4.4% and $2,753 million — a $23 million difference, immaterial to the answer.
3. Conversion to cash. Guided 2026 Adjusted Free Cash Flow of approximately $1,850 million on $2,880 million of guided EBITDA is a 64.2% conversion, after cash interest, cash taxes, $350 million of investment capital, $185 million of maintenance capital and distributions to the consolidated joint-venture minorities [16]. The company's Adjusted FCF measure is already net of those minority distributions [19]. Applied to mid-cycle EBITDA: $2,730m × 0.642 = $1,753 million.
4. Framework deductions. Less $50 million of equity-indexed compensation. Less an acquisition allowance, where the choice of allowance is the live assumption: the framework's trailing five-year average of $701 million, or the pre-2025 run rate of roughly $322 million a year implied by "over $5.7 billion" of acquisitions from 2016 through 2025 of which approximately $2.8 billion fell in 2025 [20].
5. Preferred. Less approximately $205 million of Series A and Series B distributions, being $144 million on the post-repurchase Series A base (the February 2026 payment was $36 million [8]) and roughly $61 million on Series B at the current floating rate.

Mid-cycle result, on 706.6 million units and a $26.33 price:

• With the $701 million allowance: ($1,753m − $50m − $701m − $205m) = $797m, $1.128 per unit, 4.28%.
• With the $322 million allowance: ($1,753m − $50m − $322m − $205m) = $1,176m, $1.664 per unit, 6.32%.

The mid-cycle band of 4.3% to 6.3% brackets the spot reading of 4.67% and sits 1,868 to 2,072 basis points below the 25% line. A reader who prefers a different cycle window can substitute it directly at step 2; because the EBITDA history is tight, a ±10% swing in the mid-cycle EBITDA assumption moves the yield by roughly ±50 basis points — not enough to change the position against any of the three reference lines.

The consensus check

fit_features.consensus_forward_yield carries Capital IQ consensus free-cash-flow means but no yields, because the feature file could not resolve a market cap. The absolute figures are usable directly. There is no direct consensus on "adjusted FCF attributable to PAA common"; the proxy is CapIQ's consolidated free cash flow mean for PAGP, vintage 3 August 2026, from which the same preferred and joint-venture minority claims must be deducted to reach the common. The deduction used below is a flat $655 million a year — $205 million preferred plus $450 million of joint-venture minority distributions, the latter held at the FY2025 actual of $447 million [6].

Derived: Capital IQ consensus free-cash-flow means (fit_features.consensus_forward_yield, vintage 3 August 2026), less $655m of annual preferred and joint-venture minority claims, divided by 706.6m look-through units and the $26.33 close. Minority-distribution anchor: FY2025 10-K [6].

Consensus does not clear the bar, and does not get near it. FY2026 at $1,580.2 million computes to 4.97%; FY2029, the last year with an estimate, to 7.46%. Four years of forward consensus put the yield below even the fortress line, on a balance sheet that selects the levered line. On a 25% bar, consensus would need adjusted FCF to the common of roughly $4.65 billion — about five times the FY2025 figure and more than PAA's entire consolidated Adjusted EBITDA. On a 10% bar it would need roughly $1.86 billion to the common, or $2.5 billion of consolidated free cash flow — 58% above the FY2029 consensus mean.

Because consensus sits below the bar, the framework requires an explicit mean-reversion underwrite rather than an assertion. The honest version is that there is no reversion to underwrite. This is not a name where consensus has cut estimates and anchored the price to a trough: the consensus FCF path rises every year from FY2026 to FY2029, Adjusted EBITDA guidance was raised at the 1Q26 call [16], and the share price is above its pre-drawdown peak. The yield gap is a price gap, not an earnings gap. Closing it to the 10% line from here would require the price to fall roughly 53%, to about $12.30, with FY2027 cash flow held flat — a level PAGP last traded at in 2022. My estimated probability that the look-through adjusted yield exceeds 10% within three years on the current price is below 10%: it requires either a halving of the multiple or a doubling of cash flow, and the mechanism for the second — Permian volume growth plus the acquired EPIC and Cactus III assets — is already inside the consensus path shown above. The 25% line is not reachable on any assumption the record supports.

The reference the market actually pays attention to is the distribution: $0.4175 per share declared for the fourth quarter of 2025 [7] annualises to $1.67, a 6.3% distribution yield at $26.33 — above the 4.67% adjusted-FCF yield. The partnership is distributing more than the framework's adjusted measure earns, which is what the 2025 acquisition year does to a five-year average.

Cash conversion

Derived: reported FCF and revenue from fit_features, reconciled to the filed cash-flow statements [1]; Adjusted EBITDA attributable to PAA from the MD and A non-GAAP reconciliations [14] [18] [17].

The two conversion measures point in opposite directions, and both readings are true.

Reported FCF over revenue improves: 3.9% in FY2019 to 5.2% in FY2025, with FY2025 the best of the seven years. That series is close to meaningless on its own — revenue is $44.3 billion of mostly pass-through crude purchases against $2.9 billion of operating cash flow, so the ratio moves with the oil price as much as with the business. FY2020's 3.3% trough came in the year revenue fell 31%.

Adjusted FCF to the common over Adjusted EBITDA attributable to PAA deteriorates, and this is the measure that matters for the framework: 58.7% in FY2021, 55.2% in FY2023, 38.6% in FY2024, 30.5% in FY2025. Two forces drive it, and neither is a one-year accident. Joint-venture minority distributions rose from $14 million in FY2021 to $447 million in FY2025 [10] [6] as the Permian JV and Cactus II were consolidated — EBITDA that appears in the headline but is not PAA's to spend, running at $541 million of FY2025 Adjusted EBITDA attributable to those minorities [14]. And the acquisition deduction has risen with a genuinely more acquisitive posture: $2.8 billion of the $5.7 billion of deals completed since 2016 fell in 2025 alone [20].

The counter-fact worth holding against that reading: the acquisitions bought EBITDA, and 2026 guidance of $2,880 million is $47 million above FY2025 with the NGL business — $170 million of guided EBITDA — on its way out the door [16]. If the acquisition cadence returns to its pre-2025 run rate, the conversion ratio recovers toward the mid-40s and the mid-cycle yield toward the 6.3% upper case. That is the arithmetic the upper end of the normalization band already assumes, and it still leaves the name well inside every reference line the framework draws.

Data gaps

fit_features returned not_computable for adjusted FCF, adjusted FCF yield, yield baseline, balance-sheet class, FCF stability, float-retirement years and the consensus forward yields — the structured feed carries no stock-compensation or acquisitions row, no share count, and a long-term-debt figure ($1,099 million) that is the term loan rather than the debt stack. Every figure above is rebuilt from the filed pages and the price feed, with the workings shown. The one input that could not be sourced at all is acquisition spend for FY2017 and FY2018: the corpus holds PAGP 10-Ks back to FY2021 only, so the trailing five-year acquisition window is complete from FY2023 forward and short by one to four years before that.