MPLX › analyze
MPLX · Analyze
Context. MPLX is a nomination to fill a stated gap in the 💰 Income–Yield Today sleeve: every current member (MAIN, TSLX, OBDC, ARCC, O, VICI) re-prices off the 10yr Treasury, which just hit a two-year high. This report tests whether MPLX is a good business, whether it is well priced today, and whether it actually diversifies the sleeve's rate-duration factor rather than just adding a seventh bond-proxy.
Knowledge Base Check
python .mcp/kb.py find returned no live notes on MPLX, midstream, or MLP — this is
first coverage, nothing to cite or contradict. Two pitfalls apply directly and are used below:
pitfall-fcf-definition-diverges-with-jv-partner-funds (generalized here to the
growth-vs-maintenance capex split, which does to MLP "free cash flow" what JV partner funds
do to Intel's) and pitfall-dyt-inverts-when-price-caused-the-yield (used in its mirror
direction — MPLX's yield is below its own band, and the question is whether the band or
the price is the stale number).
1. Fundamentals — General Health
Snapshot (2026-09-10, fin.py)
| Metric | Value | Metric | Value |
|---|---|---|---|
| Price | $59.63 | Mkt Cap | $60.46B |
| EV | $85.78B | 52wk range | $47.80 – $60.95 |
| P/E (ttm / fwd) | 12.96 / 12.30 | Beta | 0.46 |
| Div yield (ttm) | 7.15% | 5yr avg yield | 8.06% |
| Payout ratio (GAAP, vendor) | 90% | D/E | 183% |
| Gross / Op / Net margin | 56% / 38% / 39% | ROE / ROA | 34% / 7% |
The vendor 90% payout ratio is the wrong number and analysis_notes §4 says so explicitly: EPS is almost never the right denominator for a high-yield name; for an MLP it's distributable cash flow (DCF), not GAAP net income. On DCF the true payout is ~75% (see coverage below) — the GAAP figure overstates the strain by 15 points because MLP GAAP earnings absorb non-cash items (equity-method JV losses/gains, DD&A on a large asset base) that DCF adds back correctly.
Distributable cash flow & coverage — the real health question
| Quarter | DCF | Distribution paid | Coverage |
|---|---|---|---|
| Q2 2026 | $1,450M | ~$1,092M | 1.33x (co-reported 1.3x) |
| Q4 2025 | — | — | 1.3x |
| Q4 2024 | — | — | 1.5x |
| Q4 2023 | — | — | 1.6x |
Direction, not just level, matters here (per the High-Yield Assets overlay). Coverage has compressed in a straight line — 1.6x → 1.5x → 1.3x over three years — even as the payout itself accelerated (raises went from ~9.7-9.9%/yr to 12.5%/yr). That is not a coverage break: 1.3x is comfortably above the 1.0x floor and management is explicitly targeting 1.3x as the steady-state coverage level, not a violation. But the trend says the distribution is being grown up to the coverage ceiling, not organically outrunning it — worth watching, not yet a red flag.
A second, more consequential definitional trap sits underneath this, and it's the same
species as pitfall-fcf-definition-diverges-with-jv-partner-funds wearing a different
costume. Simple FCF (fin.py: OCF $5.91B − capex $1.81B = $4.10B, 3yr CAGR −0.9%) looks
like it's shrinking while the distribution grows — an apparent red flag. It isn't one: MLP
capex splits into maintenance (small, funded from OCF) and growth (large, funded by a
mix of retained DCF, debt, and equity, and excluded from the DCF the distribution is covered
by). Simple FCF fell because growth capex nearly tripled ($806M in 2022 → $1.81B in 2025) as
the company self-funds its Permian/Marcellus buildout — that's deliberate reinvestment, not
deterioration. Quoting "MPLX free cash flow shrank 0.9%/yr" without naming the maintenance/
growth split would be exactly the error the Intel pitfall warns against: technically true,
mechanistically wrong. DCF, the correct lens, is not shrinking — DCF/unit run-rate is
~$5.72 (annualized from Q2 2026's $1,450M ÷ ~1.014B units), up from a base of roughly $4.40-4.60
three years ago.
Capital allocation (FY2025, fin.py cash flow)
| Use | Amount | % of OCF ($5.91B) |
|---|---|---|
| Distributions | $4.02B | 68% |
| Growth + maintenance capex | $1.81B | 31% |
| Buybacks | $0.40B | 7% |
| Acquisitions (Delaware Basin sour gas $2.4B, Whiptail $237M, BANGL buy-in, Northwind) | $4.32B | 73% |
| Debt issued / repaid | +$6.59B / −$2.51B | net +$4.08B |
The tell: FCF ($4.10B) minus distributions ($4.02B) minus buybacks ($0.40B) is already −$0.32B before a dollar of the $4.32B acquisition spend is counted. Nearly all of 2025's M&A — and by extension the leverage increase below — was debt-funded, not FCF-funded. That's normal for a growth-phase MLP with IG access to capital, but it means the distribution raise, the growth capex step-up (2026 growth capital raised to $2.9B), and the M&A program are all drawing on the same balance sheet simultaneously. Leverage is the pressure valve to watch.
Leverage
| Metric | FY2024 | FY2025 |
|---|---|---|
| Net debt / EBITDA (consolidated, roic.ai) | 2.96x | 3.22x |
| EBITDA / interest expense | 6.96x | 7.08x |
| EBITDA (consolidated) | $6.57B | $7.29B |
Company's own reported "leverage ratio" was 3.7x at Q1 2026 — higher than the 3.22x consolidated figure above for the same reason the JV pitfall exists: the company's metric nets out non-controlling/JV attribution differently than a straight consolidated EBITDA calc. Use the company's own 3.7x for anything covenant- or rating-adjacent; it's the more conservative number. Either way this sits inside MPLX's own long-run target band (mid-3s to sub-4x) and interest coverage (7.1x) is strong — investment grade (Fitch confirmed BBB on a 2026 senior notes offering) — but the direction (2.96x → 3.22x/3.7x) is up, not down, while distribution growth, capex, and M&A all accelerate together. This is the single biggest structural risk in the file.
Shares/units outstanding & growth
- Units outstanding essentially flat (0.3% 3yr CAGR) — a modest buyback ($400M in 2025, ~$1.1B remaining authorization) roughly offsets unit issuance from equity-linked deals. Not a dilution story either direction.
- Revenue CAGR 2.9% (3yr), net income CAGR 7.6% (3yr, margin expansion from the IDR-free cost-of-capital structure and operating leverage), OCF CAGR 5.6% (3yr).
- Distribution history (quarterly rate,
get_stock_actions): $0.688 (Aug '21) → $0.705 ('22) → $0.775 ('23) → $0.850 ('24) → $0.957 ('25) → $1.077 ('26). Point-to-point 5yr CAGR ~9.4%, with the raise cadence accelerating: +9.9% (Nov '22) → +9.7% ('23) → +12.6% ('24) → +12.5% ('25), and 12.5%/12.5% guided for 2026-2027. Never cut, never frozen, in the company's history as a standalone entity.
§1 verdict: health is good and coverage clears the ≥1x bar with room (1.3x), but the combination of accelerating payout growth, rising growth capex, opportunistic M&A, and creeping leverage — all funded from the same debt-heavy well — is a real trend to track, not a one-quarter data point.
2. Moat & Competitive Position
Correcting the brief's own framing: the "MPC dropdown" story is largely over, not ongoing. MPC exchanged its GP economic interest and IDRs for MPLX units back in 2017-18 (the "IDR simplification"), which permanently lowered MPLX's cost of capital but also ended the era of MPC selling refining-logistics assets into MPLX for cash. Growth since has been funded by organic capex and open-market third-party M&A — the $2.4B Delaware Basin sour-gas treating acquisition (closed Aug 2025), the $237M Whiptail Midstream gathering deal (Four Corners), the BANGL pipeline buy-in, and Northwind Midstream — not by MPC dropdowns. MPLX has become a genuine third-party consolidator in Permian/Marcellus midstream, which is arguably a better growth story than a dropdown pipeline dependent on a single sponsor's asset sales, but it does mean M&A execution risk now sits with MPLX's own management, not a captive parent relationship.
What MPC still is: the anchor customer, not the growth engine. MPC owns ~64% of MPLX's LP units and provided 48% of MPLX's total 2025 revenue — durable, long-term minimum- volume/tariff commitments tied to MPC's refinery system, and the two companies' incentives are tightly aligned (MPC needs the logistics, MPLX needs the volume). But 48% from one counterparty is real concentration; if MPC ever rationalizes refining capacity for its own reasons (a live industry-wide risk as EV adoption and demand-side headwinds play out over a 10-20yr horizon), MPLX's Crude Oil & Products Logistics segment feels it directly, contract protections notwithstanding.
Adversarial stress-test. Could a well-funded rival replicate MPLX's position? Not easily and not soon. Physical gathering/processing networks (Marcellus at 94% utilization, expanding to 8.1 Bcf/d processing capacity via Harmon Creek III; Permian gas processing at 1.4 Bcf/d post-Secretariat I) took decades of permitting, right-of-way acquisition, and capital to build, and a rival would need all three simultaneously. The NGL/gas value chain (BANGL, Whistler, and the new Traverse pipeline JV with WhiteWater and Enbridge) layers switching costs (long-term take-or-pay contracts) on top of that physical scarcity. Moat sources: efficient scale + high replacement cost + contracted switching costs — a real, quantifiable moat, not a narrative one.
Evergreen assessment — mixed, and worth saying plainly. The natural gas/NGL side has a genuine structural tailwind (Permian associated gas growth, and the same AI/data-center gas demand thesis the sleeve-vacancy note is implicitly hedging against by looking outside the rate-duration cluster). The crude/refined-products logistics side is tied to MPC's refining volumes, which face a slow-but-real secular decline risk over a multi-decade horizon. Net: not a "forever business" in the purest sense, but a very long-duration one (multi-decade), with the growing half currently outpacing the slow-declining half. That's durable enough for an income holding on a 10-15yr horizon; it is not a compounder thesis, and it shouldn't be sold as one.
Fee-based insulation: ~85-90% of revenue is fee-based/take-or-pay/minimum-volume, which insulates near-term cash flow from commodity price swings but does not insulate it from volume decline if refining or drilling activity structurally contracts — a distinction the 90% figure alone doesn't convey.
3. Valuation
Graham's Intrinsic Value: √(22.5 × $4.60 EPS × $13.83 BVPS) = $37.84 vs. $59.63 spot — price is 58% above Graham IV. This model is weighted low here on purpose: MLP book value is distorted by decades of accumulated D&A on a heavily-levered, replacement-cost-understated pipeline network, and GAAP EPS is (per §4 of analysis_notes) the wrong earnings denominator for this asset class regardless. Treat $37.84 as a Graham floor / sanity check, not a fair value estimate.
Dividend Yield Theory — the pitfall applies in its mirror direction. Naive DYT: current
yield 7.15-7.2% vs. 5yr average 8.06% says MPLX is trading rich to its own history (the
opposite problem from the NVO case the pitfall was written on, but the same diagnostic
question). Per pitfall-dyt-inverts-when-price-caused-the-yield's guard: decompose why the
yield moved. Over the 5yr band period, the distribution grew ~9.4%/yr while the unit price
roughly doubled — the yield compression is price-driven, not caused by a slowing payout.
That alone doesn't validate the naive $53.45 ($4.308 ÷ 0.0806) target, though, because the
question the pitfall poses next is whether the band itself is stale. It is: the 2021-2023
period baking into that 8.06% average was a stretch when MLPs traded cheap on generalist/index
avoidance (K-1 stigma, post-2020 oil-crash scar tissue, pre-IG-rating uncertainty). MPLX has
since re-rated as a lower-risk, IG-rated, self-funding large-cap midstream name — the "peer
class it is becoming" (EPD ~6.7% yield / ~10.5x EV/EBITDA) supports a structurally lower
fair-value yield than its own 2021-vintage history, not a reversion up to 8%. Resetting the
band to 7.0-7.75% gives a fair-value range of $55.60-61.50 — this is the number carried
forward, and the naive $53.45 is explicitly discarded per the pitfall's own prescribed repair.
EV/EBITDA peer comp. At $85.78B EV against $7.1-7.3B of forward adjusted EBITDA, MPLX
trades ~11.7-12x — richer than EPD (~10.5x) and much richer than ET (~7.6x forward), in
line with its higher-quality/lower-leverage positioning. (fin.py's own 14.03x print is an
outlier versus this and versus web-reported 11.0x comps — likely the same
consolidated-vs-attributable EBITDA definitional gap flagged in §1; the roic.ai/peer-anchored
11.7-12x is used here as the more defensible number.) Applying a 10.5-12x band to $7.1-7.3B
EBITDA and netting ~$24B of debt implies an equity value of roughly $53-64/unit.
P/DCF cross-check. $59.63 ÷ ~$5.72 DCF/unit run-rate = 10.4x — mid-band for the midstream group, neither a screaming discount nor obviously rich.
Fair value range: $54-62, central ~$58. Three independent approaches (reset DYT, EV/EBITDA peer comp, P/DCF) converge inside a few dollars of each other, which is itself a useful confidence signal. Spot ($59.63) sits at the upper half of that range, near the 52-week high ($60.95), with effectively no margin of safety. This is a good business priced fairly-to-slightly-rich, not a discount — per the framework's own first-principles table, that's the "wait / watch" quadrant, not the buy quadrant.
Entry: $52-56 (yield 7.7-8.3%, P/DCF ~9.1-9.8x) — a real pullback, not a rounding move.
Trim: $68 — ≈11.9x DCF/unit ($5.72), ≈6.3% yield. Dollar-form is deliberate: the site
derives a multiple-trim as mult × price ÷ P/E, i.e. off GAAP EPS ($4.60), and MLP
economics run through DCF ($5.72/unit) — the same GAAP-EPS-vs-true-cash-metric gap that
forced VICI and NLCP's trims into dollar-form. A bare "12x" would render off GAAP EPS
(~$55.20) — below spot, which would invert the call. Re-set the ~12x DCF judgment at each
/analyze.
4. Sentiment
Coverage is broadly constructive and largely income-media (24/7 Wall St., Motley Fool
"$10,000 into income" pieces, Zacks) — the retail narrative is entirely about the 7%+ yield,
which is exactly the naive framing the DYT decomposition above corrects. Zacks carries a
"buy" recommendation consensus (fin.py); analyst mean target $62.67, modestly above spot.
No adverse insider signal: the one open-market transaction on record in the lookback was a
small CFO sale (4,000 units, $189K, Nov 2024) offset by a small officer purchase (4,000
units, $211K, Mar 2025) — net neutral, immaterial size either way, the bulk of insider activity
is routine RSU vesting (per pitfall-yahoo-insider-purchases-counts-rsu-grants, correctly
excluded here). No red-flag news items in the recent window — the Q2 2026 print was
in-line-to-slightly-soft on volumes but held guidance, consistent with the "5% EBITDA growth"
framing in the sell-side slide commentary.
Explicit Answers
1. §1 health bar + distribution coverage ≥1x? Yes on both counts. FCF/DCF generation is solid, leverage is investment-grade (though rising: 2.96x → 3.22x consolidated, 3.7x on the company's own metric), and DCF coverage is 1.3x — comfortably above 1.0x, though it has compressed from 1.6x three years ago as payout growth accelerated. Not a break; a trend to watch.
2. Buy zone or watch? Watch. Three independent valuation approaches converge on $54-62 fair value; spot ($59.63) sits at the top of that band near the 52wk high with no margin of safety. Good company, fair-to-rich price — the framework's own "wait" quadrant, not "value."
3. Does it beat NLCP (HOLD 5.0), and does it genuinely diversify the rate-duration cluster? Yes on both, with a caveat on the second. MPLX beats NLCP decisively on every quality axis that matters for an income sleeve: investment-grade balance sheet vs. NLCP's 1.8%-debt-but-tiny-and-illiquid OTCQX profile, a stable-to-accelerating distribution history vs. NLCP's frozen-9-quarters payout, 48%-single-counterparty concentration that is at least a majority-owner-aligned counterparty vs. NLCP's fragmented tenant base with one in restructuring, and $60B of market cap/liquidity vs. a micro-cap. On diversification: MPLX's earnings driver is genuinely different in mechanism from the rest of the sleeve — contracted pipeline/gathering volumes and inflation-escalated tariffs, not a floating-rate loan book (BDCs) or a net-lease cap-rate (REITs) — so it does address the stated gap (a different fundamental risk factor). But it is not rate-immune: as a high-yield equity it still carries some bond-proxy price sensitivity to the 10yr, just via a different transmission mechanism than MAIN/TSLX/OBDC/ARCC/O/VICI. Diversifies the driver, doesn't eliminate all rate sensitivity — a real but partial answer.
4. Proposed conviction: 6.0. Above NLCP (5.0), below the sleeve's better-priced/held names (MAIN 7.0, VICI 5.5-adjacent-but-held-for-coupon). Verdict WATCH, not ACCUMULATE, because the business case is genuinely good but the entry isn't — this is a name to size on a pullback to $52-56, not to chase at the top of its own fair-value band.
⚠️ K-1 / UBTI — Structural Tax Consideration (explicit flag)
MPLX is a publicly traded limited partnership. It issues a Schedule K-1, not a 1099-DIV, and this has two real consequences for account placement:
- Taxable accounts: K-1s typically arrive later than 1099s (often March, sometimes with extensions), can require multi-state filing awareness (MPLX's own K-1 is reportedly simpler than most MLPs — a single K-1 rather than multiple state schedules — but the partnership still operates across many states), and carry Section 751 "hot asset" recapture treatment on sale that converts part of any gain to ordinary income. This is manageable but is genuinely more tax-prep friction than a C-corp dividend payer.
- IRA/Roth/tax-advantaged accounts — the sharper issue. As a limited partner, an IRA holding MPLX earns its pro-rata share of the partnership's unrelated business taxable income (UBTI), reported on the K-1 (commonly line 20V). The first $1,000/year of UBTI across all the IRA's MLP-type holdings combined is exempt; above that, the custodian must file Form 990-T on the IRA's behalf and the IRA itself owes tax at compressed trust rates — a real cost and a real paperwork burden that most retail custodians pass through clumsily. At MPLX's scale, a single meaningful position can plausibly push a small IRA over the $1,000 UBTI threshold on its own.
Recommendation: if this position is added, hold it in the taxable account, not an IRA/Roth — the K-1 friction is manageable there, and long-term capital gains + qualified treatment on the return-of-capital portion of distributions is the more efficient outcome anyway. If the intent was specifically to add IRA-eligible midstream income exposure, the better vehicles are the C-corp midstream names (ONEOK, Williams, Kinder Morgan) — same sector exposure and yield profile, standard 1099-DIV, no K-1/UBTI issue, no Form 990-T risk. That's a real fork in the road worth deciding explicitly, not defaulting past.
Biggest Risk
Leverage is rising while three things draw on the same balance sheet at once: accelerating distribution growth (12.5%/12.5% guided), a stepped-up growth capex program ($2.9B for 2026, up from $1.8B in 2025), and opportunistic M&A ($2.4B+ in 2025 alone) — nearly all debt-funded, since FCF after distributions and buybacks was already modestly negative before a dollar of acquisition spend. Net debt/EBITDA has moved from 2.96x to 3.22x (consolidated) / ~3.7x (company's own metric) inside twelve months, still inside the IG target band but trending toward its ceiling. If gas/NGL growth decelerates or a large project underdelivers while this leverage trajectory continues, DCF coverage — already compressed from 1.6x to 1.3x over three years — has less room to absorb a shock before it tests the 1.0x floor that keeps the whole income thesis (and the IG rating that funds it) intact. Secondary risks: 48% single-counterparty revenue concentration (MPC); multi-decade secular refining-volume decline risk to roughly half the business; the K-1/UBTI account-placement friction above.