Durability

Bottom line

Plains owns assets that cannot be rebuilt and survived a 20% collapse in global oil demand with free cash flow intact. It also earns 42% less per barrel moved than in 2019, guides 2026 adjusted EBITDA down 3%, and books 96% of revenue as merchant crude sales whose dollar value is a price forecast. The conviction sources are real but partial; the year-10 gate does not clear.

The conviction sources, one at a time

The framework's year-10 gate draws conviction from five places: market structure, regulatory entry barriers, capital intensity, essentialness, and operating history. Three of the five apply to Plains with force. Two do not, and saying so is the point of running the list.

Market structure — oligopoly in assets, price-taker in economics

The corporate structure and segment mix are set out in Business; what matters here is share stability. Plains is one of several large owners of Permian crude infrastructure, not one of two. Its own Item 1 describes a market where "existing third-party owned pipelines with excess capacity in the vicinity of our operations also expose us to significant competition based on the relatively low operating cost associated with moving an incremental barrel," and states plainly that "as a result of multiple pipeline expansions in the Permian Basin and other areas… we continue to experience heightened competition for uncommitted barrels and contract renewals, which puts downward pressure on tariffs and margins" [1]. The named competitors run to "other crude oil and NGL pipeline and terminalling companies… major integrated oil companies and their marketing affiliates, independent gatherers, private equity backed entities, banks that have established a trading platform" [1].

Share is stable in barrels and unstable in price. Plains moved 9.68 million barrels per day of pipeline tariff volume in 2025, up from 6.61 million in 2019 [2] [3]. Over the same window, Crude Oil Segment Adjusted EBITDA fell from $2,753 million to $2,344 million [2] [3]. Volumes up 46%, segment earnings down 15%.

Derived: Crude Oil Segment Adjusted EBITDA divided by reported average crude oil pipeline tariff volume times 365 days; 2026 uses guidance midpoints. Sources: FY2025 10-K [2]; FY2023 10-K [4]; FY2021 10-K [3]; 2026 guidance [5].

Sources: FY2025 10-K [2]; FY2023 10-K [4]; FY2021 10-K [3]; 2026 guidance [5].

Two caveats on the ratio. The 2019 base year included a merchant book that captured unusually wide grade and location differentials, so part of the fall from $1.14 is the disappearance of a windfall rather than tariff erosion. And the volume denominator counts a barrel twice when it moves across two connected systems, a convention the filing states explicitly [6]. Neither caveat rescues the 2021-to-2025 stretch, where the base year is post-windfall and the convention is unchanged: $0.843 to $0.663, down 21%, on volumes up 56%.

A market structure that lets a participant grow volumes by half while earning less per unit is an oligopoly in physical assets and something closer to a commodity market in pricing. That distinction matters more at year ten than at year one.

Regulatory entry barriers — a rate ceiling, not a moat

Interstate liquids pipelines file tariffs with FERC under the Interstate Commerce Act, and most, including Plains, adjust rates within an annual index ceiling that FERC resets every five years [7]. This regime constrains what an incumbent may charge; it does not decide who may build. It is the opposite of the bank or insurer case the framework leans on, where the regulator's licence itself keeps entrants out.

What the regime does block is fast, cheap greenfield construction. Permitting under the Clean Water Act nationwide permit programme, the Endangered Species Act, and the National Environmental Policy Act is described in the filing as capable of making projects unviable through "lengthy regulatory review and approval requirements" [7]. And Plains cannot condemn allotted tribal land, which cuts both ways: it protects incumbency in some corridors and threatens Plains' own rights-of-way in others [8].

The honest grade: permitting friction is a genuine barrier to new long-haul capacity and a weak one against gathering-system competitors, who build inside a basin on private acreage. Regulatory entry barriers apply partially here, and not in the form the framework's precedents describe.

Capital intensity — the strongest source, and it is genuine

Plains carries $16.9 billion of net property and equipment, $900 million of linefill, $2.8 billion of equity-method investments and $1.8 billion of net intangibles [8]. In the Permian alone it operates over 5,600 miles of gathering pipeline representing roughly 3.9 million barrels per day of capacity, an intra-basin system of roughly 3.1 million barrels per day, and interests in long-haul systems totalling over 2.8 million barrels per day of takeaway [6]. Company-wide the footprint runs to over 20,000 miles of active pipeline [9].

Nobody replicates a wellhead-to-Gulf-Coast gathering and takeaway network for a marginal return. Gathering pipelines are further "supported by long-term acreage dedications" [6], which tie acreage rather than volumes and survive a producer's decision to slow drilling. This conviction source applies in full.

It also carries the framework's own warning attached: capital-heavy essentials survive, but they are not always good businesses. The per-barrel series above is what that looks like from inside.

Essentialness — tested in 2020, and it held

The most useful evidence on essentialness is not a demand forecast but a demand shock. Global consumption fell by roughly a fifth in the second quarter of 2020. Plains' pipeline tariff volumes fell 8%, from 6,613 to 6,082 thousand barrels per day; Crude Oil Segment Adjusted EBITDA fell 20%, from $2,753 million to $2,216 million [3]; reported free cash flow stayed positive at $772 million [10].

Sources: FY2021 10-K, Crude Oil Segment operating results [3] and subsidiary distributions [11]; consolidated cash flow data [10].

The distribution did not hold. PAA paid $1.38 per common unit in 2019, $0.90 in 2020 and $0.72 in 2021, a 48% reduction across two years [11]. That was a capital-allocation choice to delever rather than a cash-generation failure, and the payout policy is examined in Self-Help. For the durability question, the relevant fact is that the barrels kept moving and the cash kept coming.

Management's forward view is that demand grows: "we expect crude oil demand to continue increasing, driven largely by our view that hydrocarbon-based fuels are the most efficient fuels for the transportation of people and goods" [12]. That is an interested party's forecast, and it is placed directly above an EIA chart showing implied global stock builds of roughly 2 to 3 million barrels per day through 2027 [12]. Essentialness applies; abundance is the near-term condition.

Operating history — 28 years, through four cycles

PAA was formed in 1998 and has since completed and integrated over 100 acquisitions for roughly $17.5 billion, implemented investment capital projects totalling roughly $18.7 billion, returned roughly $21.0 billion to equity holders, and moved from non-investment-grade to investment-grade credit [13]. PAGP itself, the listed Class A vehicle, dates from the October 2013 IPO. Twenty-eight years of operating history spans 2008-09, the 2015-16 crude collapse, 2020, and the 2022 spike.

The framework asks for 30 to 50 years. Plains is just short of the lower bound at the asset level and twelve years into the listed structure. The record is long enough to be evidence and carries one scar the framework should see: those 100-plus acquisitions mean the operating history is partly a history of buying operating history.

The structural threats, hunted

Recontracting, quantified — the threat with a number attached

The most concrete year-10 threat is not distant. It is already in the 2026 guidance walk. Management's own bridge from 2025 to 2026 adjusted EBITDA takes $100 million out for "Cactus I Re-contracting" and $365 million out for the Canadian NGL sale, and puts $250 million back from the Cactus III acquisition plus $130 million from efficiency, optimisation and tariff escalation [14].

Source: Fourth-Quarter 2025 earnings presentation, Key Drivers 2025 to 2026 [14].

The 10-K names the mechanism without the number: 2025 crude segment results were held back by "the impact from certain Permian long-haul contract rates resetting to market in 2025" [15]. The CFO put it more directly on the fourth-quarter call: fourth-quarter crude segment adjusted EBITDA of $611 million reflected "a full quarter impact of recontracting on our long-haul systems" [16].

Sizing the year-10 exposure. The contracts that reset are the Permian long-haul agreements written in 2018-19, when takeaway was scarce and shippers paid for it. The weighted average remaining term of Plains' minimum volume commitments and acreage dedications is roughly five years [17]. On a ten-year view, therefore, effectively the entire contracted book resets twice. If a further two rounds of resets each carry a headwind of the order of the $100 million already booked, that is roughly $200 million of recurring adjusted EBITDA, or 7% of the $2,750 million 2026 guide [5], before any offsetting escalation or new commitment. That is a bounded estimate, not a forecast; it assumes the next two resets look like the last one.

The fair counter-fact sits in the same filings. The contracted revenue backlog has been refilled every single year and has grown, not shrunk. Total remaining performance obligations rose from $1,949 million at the end of 2021 to $3,430 million at the end of 2025, and the year-ahead figure has climbed steadily from $416 million to $639 million [18] [19] [20].

Sources: FY2025 10-K [18]; FY2023 10-K [21]; FY2021 10-K [19].

Each vintage traces the same downward slope and each new vintage starts higher. The roll-off shape is a disclosure artefact of a business that recontracts continuously, not evidence of decay. What the table cannot show is the price at which the refilling happens, and that is what the $100 million tells us.

One scale check on the backlog. Committed revenue of $639 million for 2026 sits against reported services revenues of $1,761 million for 2025 [22]. The filing explains the gap: the table excludes "expected revenues from legacy shippers not underpinned by minimum volume commitments" and excludes acreage dedications entirely, because they "require us to perform future services but do not contain a minimum level of services" [23]. Roughly two-thirds of fee revenue therefore rests on uncommitted, walk-up and dedication-based volume that reprices continuously. That is the exposure the recontracting number is measuring.

Regulatory reversal — FERC has proposed cutting the escalator

The index that governs annual tariff increases is up for its five-year reset on 1 July 2026. On 20 November 2025 FERC issued a Notice of Proposed Rulemaking proposing an index level of the Producer Price Index for Finished Goods minus 1.42% for the period 1 July 2026 to 30 June 2031 [24]. The index currently in force is PPI-FG plus 0.78%, reinstated by FERC in September 2024 after the D.C. Circuit vacated an intervening downward revision in LEPA v. FERC [24]. The proposed change is 220 basis points of annual escalation on indexed rates, compounding over five years to roughly 11% of tariff ceiling relative to the present index. FERC's methodology also compels a filing to lower rates in any year the index is negative [24].

Plains' own FY2025 10-K describes the LEPA litigation and warns that "the final resolution of these petitions could have an adverse effect on our cash flows" [7], and separately notes that FERC "could order PAA to reduce its rates and could require the payment of reparations to complaining shippers for up to two years prior to the complaint" [25]. It does not disclose the November 2025 five-year-review NOPR; the proposal is sourced here from a peer's filing on the same regime. The FY2025 guidance walk does credit "FERC recoup / Tariff escalations / Other" for $30 million in 2026 [14], so the escalator is a live contributor, not a rounding item.

Volume — the growth engine is guided to stall

Every year since 2020, rising Permian volume has offset falling per-barrel economics. For 2026 the company assumes Permian basin production is "relatively flat" at roughly 6.6 million barrels per day [5], and on the first-quarter 2026 call management confirmed "our assumption for the Permian this year was flat" [26]. The EIA's January 2026 Short-Term Energy Outlook is the same picture and then worse: Permian output of 6.6 million barrels per day in 2026, easing to 6.5 million in 2027, with total US crude falling from 13.6 to 13.3 million barrels per day, on rig activity declining as prices sit below stated Midland and Delaware breakevens.

Plains' own sensitivity puts a 100 thousand barrel per day change in total Permian production at $10 to $15 million of adjusted EBITDA, and a $10 per barrel move in WTI at $40 million [5]. Those coefficients are modest: a 500 thousand barrel per day basin decline costs $50 to $75 million, about 2% of guided EBITDA. The volume threat to year-10 cash flow is real but second-order; it matters mainly because volume growth is what has been paying for tariff compression.

Guided 2026 pipeline volumes of 10,350 thousand barrels per day are 7% above 2025 [5], on a flat basin. That increment is bought, not grown: it is Cactus III, acquired in 2025.

Growth by acquisition — the pattern, priced

Cash paid for acquisitions was $2,651 million in 2025, against $248 million in 2024 and $425 million in 2023 [27]. Since 2016 Plains has spent over $5.7 billion on acquisitions, roughly $2.8 billion of it in 2025 alone [28]. The 2026 walk shows where that spend lands: $250 million of acquired EBITDA is filling a hole created by $100 million of recontracting and $365 million of divestiture [14]. The company is buying replacement earnings at a rate that matters to the framework's adjusted-FCF definition, which deducts the five-year average of acquisition spend precisely to catch this. That arithmetic belongs to Yield; the durability observation is narrower and harder — an earnings base that requires continuous purchase to stay level is not the same asset as one that compounds on its own.

Execution here is good. The Cactus III integration was already at half its $50 million synergy run rate one quarter after closing [16]. Execution is not a moat, and a year-10 case that leans on management continuing to buy well has no structural protection in it.

Customer concentration

ExxonMobil and its subsidiaries accounted for approximately 31%, 31% and 27% of revenues in 2025, 2024 and 2023 [29]. The filing states the consequence: "if we were to lose one or more of these customers, there is risk that we would not be able to identify and access a replacement market at a comparable margin" [29].

The mitigant is real. The majority of that revenue is Crude Oil segment merchant activity — barrels bought and resold at multiple locations [29] — which passes through at a thin margin rather than contributing 31% of EBITDA. Concentration at 31% of revenue overstates the economic exposure, and the counterparty is investment grade. Rated a live but not decisive threat.

Substitution and technology

Plains' filing names the mechanism: demand fluctuates with "fuel conservation measures, alternative fuel adoption, governmental regulation, including climate change regulations, and technological advances in fuel economy and energy generation and storage technologies," and legislative action to reduce greenhouse gas emissions "could… accelerate the adoption of alternative energy technologies, thereby causing a reduction in the demand for such products" [30].

The framework's sharper question — is anyone's margin here an Amazon opportunity — has a specific answer for crude midstream, and it is mostly no. There is no software substitute for moving a physical barrel 500 miles, and the alternatives the filing names (truck, rail, barge) "typically cost more" [1]. The substitution risk is not to the transport function; it is to the product. On a ten-year view the honest statement is that US crude demand is a wide distribution and the Permian's share of US supply is high, so a slow-transition scenario damages volumes gradually rather than abruptly. This threat is real and slow-acting, and the volume sensitivities above bound it: it is smaller than the pricing threat over the same window.

Exclusion-relevant facts

Revenue is entirely North American — $39,761 million United States and $4,501 million Canada in 2025 [31]. No China revenue or asset dependence appears anywhere in the record. Structural-decline evidence is addressed directly below.

The disqualifier check

The framework's hard disqualifier is revenue declining at a high single-digit rate for three consecutive fiscal years after a long existence. The deterministic feature file reports three_year_hsd_decline: false and consecutive_decline_years: 1 (fit_features.revenue_trajectory). The full ten-year series behind that flag:

Source: reported consolidated revenues, FY2016-FY2025 [32]; FY2023-FY2025 as restated for continuing operations in the FY2025 10-K [33]. Series matches fit_features.revenue_trajectory exactly.

Source: as above; figures identical to fit_features.revenue_trajectory.per_year.

Declines occurred in 2019, 2020, 2023 and 2025. The longest run is two years (2019 and 2020, at -1.1% and -30.8%), and the most recent decline stands alone at -9.5%. The three-consecutive-year test is not met and is not close to being met.

The flag is correct and the flag is also nearly meaningless for this company, which is a point the gate needs made rather than buried. Of the $44,262 million of FY2025 revenue, crude sales were $42,408 million; transportation tariffs were $1,330 million and terminalling, storage and other fees were $349 million [22]. Roughly 96% of the top line is crude bought and resold, recorded gross. Revenue moves with the WTI price and the merchant book's size, not with the health of the pipeline network. The 31% collapse in 2020 and the 81% rebound in 2021 were price, not business.

The service line is the one that tests the business:

Source: reported services revenues by fiscal year [32]; FY2023 onward on a continuing-operations basis after the Canadian NGL business was reclassified to discontinued operations [33], so growth from 2022 to 2023 is understated on this basis.

Fee revenue has risen 49% since 2016 and has fallen in only two years, 2020 and 2021, at -11.8% and -3.0%. On the measure that tracks the actual pipeline franchise, there is no structural decline in evidence. X3 checked and absent.

FCF consistency

The framework's P2 test is the stability of the rolling five-year average of adjusted free cash flow — reported FCF less stock-based compensation less the five-year average of acquisition spend.

That series cannot be computed from the deterministic feature file. fit_features.fcf_stability returns an empty rolling_5y_avg array with cv_of_rolling_avg: null, and the stated reason is "fewer than five consecutive adjusted-FCF years"; the underlying adjusted_fcf.series shows sbc: null for all ten fiscal years and treats acquisitions as an "implicit zero" under a complete-row rule. Both inputs exist in the filings. Equity-indexed compensation expense was $50 million in 2025, $52 million in 2024 and $51 million in 2023, and cash paid for acquisitions net of cash acquired was $2,651 million, $248 million and $425 million over the same three years [27]. Treating a company that has spent over $5.7 billion on acquisitions since 2016 [28] as having spent zero materially overstates adjusted FCF. The correct figures are not substituted here; the gap is logged, and the adjusted computation is Yield's.

What can be shown on the primary record is reported free cash flow and its rolling five-year average:

Reported FCF from consolidated cash flow data [10], reconciling to the FY2025 10-K cash flow statement for FY2023-FY2025 [27]; rolling average derived. This is reported FCF, not the framework's adjusted FCF, which is not computable from fit_features.

The rolling average has risen in every window, from $783 million (FY2016-2020) to $2,048 million (FY2021-2025), and has never declined. Year-to-year FCF is volatile — the 2016-to-2017 swing was $2,088 million — which the framework tolerates. The five-year mean is not.

One negative year sits in the series. FY2016 reported FCF was -$616 million: $718 million of operating cash flow against $1,334 million of capital expenditure [10]. That is a growth-capex episode at the bottom of a crude cycle, not the underwriting-loss cadence the framework describes for insurers and banks — there is no mechanism here that periodically requires a loss year, and none has recurred as capital spending fell to a $408 to $643 million range across 2023 to 2025 [27]. On the reported basis, P2 reads as consistent. On the framework's adjusted basis, it cannot be assessed, and the year with $2.65 billion of acquisition spend is exactly the year where the two bases would diverge most.

The year-10 case, both ways

The strongest case that year-10 revenue and adjusted FCF are higher

The asset base is not reproducible and is not being challenged by a cheaper technology. Over 20,000 miles of active pipeline, $16.9 billion of net property and equipment, and Permian gathering supported by long-term acreage dedications that bind acreage rather than volumes [6] [8] [9]. Crude Oil Segment Adjusted EBITDA has risen in each of the last four fiscal years, from $1,909 million in 2021 to $2,344 million in 2025, and is guided to $2,640 million for 2026 [3] [2] [5]. The contracted backlog is 76% larger than four years ago [18] [19]. The business kept generating positive free cash flow through the worst demand shock in the history of the oil market, and reported FCF has risen from $772 million in 2020 to $2,288 million in 2025 [10]. Management is recontracting Cactus III "for term" and expects "to contract at higher rates than before with potentially new counterparties" [34], which, if Permian-to-Gulf-Coast utilisation stays as tight as current basin capacity implies, is a reasonable expectation rather than a hope.

The strongest doubt

Two legs of the gate carry independent problems.

The revenue leg is not underwritable in the framework's terms. Roughly 96% of reported revenue is crude bought and resold at prevailing prices [22]. "Year-10 revenue higher than today" for PAGP is a statement about the 2036 WTI price, and the record shows what that does: -31% in 2020, -17% in 2023, -9% in 2025 (fit_features.revenue_trajectory). No degree of business-quality analysis produces very high conviction about a commodity price a decade out.

The adjusted-FCF leg carries a directional, quantified problem. Per-barrel segment economics have compressed 21% since 2021 and 42% since 2019, on volumes up 56% and 46% respectively. The company's own 2026 bridge takes $100 million out for a single pipeline's recontracting and guides headline adjusted EBITDA down 3% [14]. The weighted average remaining term of committed contracts is roughly five years [17], so the book resets twice inside the window. FERC has proposed cutting the indexed escalator by 220 basis points from July 2026 [24]. The basin that supplied every barrel of volume growth is guided flat for 2026 and forecast by the EIA to decline in 2027. And the mechanism that has held the line — $2.65 billion of acquisitions in one year [27] — is the same spend the framework's adjusted-FCF definition subtracts.

The read

There is a genuine doubt, and it is specific: year-10 adjusted free cash flow requires either that per-barrel economics stop compressing after seven consecutive years of compression, or that acquisitions keep buying replacement earnings — and the framework's adjusted-FCF definition charges the second path against the very number it is meant to protect. Against that sits the fact that Crude Oil Segment Adjusted EBITDA has risen four years running and the contracted backlog is 76% larger than in 2021, so the compression has so far been outrun rather than lost to. The gate does not clear.

What would change this read, in order of decisiveness. A completed Cactus III and Cactus I recontracting cycle at rates at or above the expiring ones, disclosed with the dollar effect, would falsify the compression thesis directly — management said an update is due "in the coming quarters" [34]. A final FERC index for July 2026 to June 2031 at or above PPI-FG flat, rather than the proposed minus 1.42%, would remove the regulatory leg. Segment EBITDA per tariff barrel stabilising above $0.70 across two consecutive years without acquisition contribution would answer the core arithmetic. Conversely, a third consecutive annual recontracting headwind of $100 million or more, or Permian production declining faster than the EIA's 2027 path, would harden the doubt into a structural finding.