Capture Hits 112% and Every Region Clears $20 a Barrel; the Stock Still Only Matched the Index

Results vs. Consensus

The same caution we applied to the Q1 print applies here, for the same reason. Published estimates for MPC's second quarter ranged from roughly $12.94 to $14.52 of EPS and from $34.83B to $41.16B of revenue, because the crack environment moved faster inside the quarter than sell-side models were revised. The table below anchors on the most widely cited mark, $14.52, which also happens to be the most conservative characterization of the beat available. Against the low end of the published range the beat was 37%. Against the high end it was 22%. We use 22%.

Note the shape of the segment beats. The one that cleared consensus by the widest relative margin is the smallest segment, Renewable Diesel. The one that carried the dollars is Refining & Marketing, which beat by 14.75%. Midstream, the segment whose trajectory we flagged as the weak link in May, beat by 5.51%. All three cleared, which is the first quarter of our coverage where that is true.

Year-over-Year Comparison

Two lines in that table deserve more attention than they usually get. Cost of revenues rose 43.4% while revenue rose 53.5%, which is the entire quarter in two numbers: crude and feedstock costs went up a great deal and product prices went up more. And the noncontrolling interest line, which took $400M against $394M a year earlier, absorbed 7.2% of consolidated net income this quarter against 24.5% last year. The MPLX minority claim is fixed in dollars and shrinking in proportion whenever refining earns. That is a structural feature of the consolidation worth carrying into any per-share model.

Sequential Comparison

The sequential move is the largest in our coverage of this name, and it is not a like-for-like comparison: Q1 carried a deliberate turnaround pull-forward and a derivative timing loss that both reversed here. Read the table as the sum of a genuine margin expansion and the unwind of two known Q1 drags.

First-Half Comparison

The half-year view strips out the Q1/Q2 timing swaps and is the cleanest read on what has actually changed. It also sets up the second-half arithmetic that governs several of management's outstanding commitments.

Quality of the Beat

• Revenue: Price, with an unusually clean read this quarter because volume moved the other way. Net refinery throughput fell to 2,944 mbpd from 3,060, down 3.8%, while refined product sales volume was essentially flat at 3,842 mbpd against 3,835. A 53.8% increase in operating revenue on a 3.8% decline in throughput is the crack spread and the crude price, full stop. There is no acquisition contribution in the refining top line and no accounting change.
• Margins: High quality on the revenue side, mixed on cost. R&M margin of $36.33 per barrel did the work and capture of 112% says a meaningful slice of it was self-generated rather than handed over by the indicator. But refining operating cost of $5.72 per barrel missed management's own $5.65 guide and rose from $5.34 a year ago, and the explanation given in the release is a volume-denominator one: "primarily driven by decreased utilization due to planned downtime in the Mid-Con, compared to the prior year quarter." Distribution cost of $5.88 also rose from $5.52. Both controllable per-barrel cost lines are now higher year over year for the second consecutive quarter.
• EPS: No engineering whatsoever, which is the cleanest possible answer. There were zero pre-tax adjustments in the quarter, so adjusted net income of $5,138M equals reported net income attributable to MPC, and adjusted diluted EPS of $17.73 equals GAAP diluted EPS. The tax rate went the wrong way for the company and the right way for credibility, rising to 20.7% from 14.3%. Share count fell 5.5% year over year through genuine repurchase rather than through the diluted-share convention. If anything the quarter is understated relative to peers who would have found something to add back.
• Cash: The headline is the flattered part. Consolidated cash rose $5,617M to $7,768M, but $3.8B of the quarter's cash generation was a working-capital source, and management was explicit on the call that part of it is owed: "We have a liability and a repayment obligation for these crude exchanges, and we will be watching as well our payables." Operating cash flow excluding working capital was $6.6B, which is the durable number. The sensitivity disclosed on the call is that every $10 move in crude swings working capital by roughly $550M in either direction.

Segment Performance

The mix inverted. A year ago Refining & Marketing was 53.8% of segment EBITDA and Midstream was 46.7%; this quarter refining is 76.6% and midstream 20.5%. That is not a change in the midstream business, which grew. It is refining tripling underneath it. Anyone valuing MPC on a blended multiple should note that the blend just moved a very long way toward the cyclical half, and will move back.

Refining & Marketing

Segment adjusted EBITDA of $6,655M against $1,890M a year ago, on a margin of $36.33 per barrel against $17.58. Crude capacity utilization was 94%, in line with the guide management set in May, on crude runs of 2,798 mbpd against a guide of 2,795. That is the volume commitment from last quarter delivered essentially to the barrel. Total throughput came in slightly light at 2,944 mbpd against a guided 2,990, and the entire shortfall is in other charge and blendstocks (146 mbpd against roughly 195 implied), not in crude.

The disclosure that matters most is capture, which reached 112% from 99% in Q1, taking the first half to 108% against a FY2025 figure of 105%. Management decomposed it on the call rather than asserting it.

"This quarter, in particular, the 112 was driven by strict inventory discipline in a backward-dated market." — Maryann Mannen, Chairman, President and CEO

The Gulf Coast decomposition given by the CFO named four items, one of which is explicitly a Q1 reversal rather than new value creation.

"In the Gulf Coast, margin capture was positively impacted by the advantage crude we ran in the quarter, increased jet fuel margins, and the physical offset related to the derivative losses in the first quarter. These tailwinds were partially offset by secondary products, which remain a market-driven headwind as prices of secondary products lag higher clean product prices." — Maria Khoury, EVP and CFO

Assessment: Three of the four named drivers are repeatable and one is not. The derivative reversal was flagged as a Q1 timing item at the last call and duly arrived, which is a delivered commitment, but it is a one-time credit to Q2 capture and its size was never disclosed. Strip it out and capture is still comfortably above 100%. The secondary-product drag, which we complained last quarter was never sized, is still never sized, now for the second consecutive quarter. It is the only recurring capture headwind management identifies and the only one they will not quantify.

Regional Performance

MPC began publishing regional adjusted EBITDA per barrel in Q1, tied roughly 20% of the annual cash bonus to regional competitiveness, and thereby created a scoreboard it can be held to. This is the first quarter where the scoreboard reads well everywhere at once.

Gulf Coast

Segment EBITDA of $27.01 per barrel against $5.65 a year ago and $6.20 last quarter, on 100% crude utilization. Refining operating cost of $4.30 per barrel is the lowest in the system and barely moved year over year, down $0.04 from $4.34. Turnaround expense was almost nil at $0.15 per barrel against $1.93 in the Mid-Continent. The region ran clean, cheap and full.

"First, we continue to talk about our lost capacity due to internal problems being, frankly, at the lowest level that we've seen in a decade. Reliability and execution really mattered, that, I would say, is one of the drivers. Second, you heard us talk about Garyville obviously having incremental jet capacity. You saw those jet-to-diesel margins." — Maryann Mannen, Chairman, President and CEO

Assessment: This is the cleanest regional result MPC has published since it started publishing regional results, and its composition is favourable: cost flat, utilization at the ceiling, turnaround absent, and a yield project delivering into the spread it was designed for. The caveat is the turnaround line. At $0.15 per barrel the Gulf Coast carried essentially no maintenance burden this quarter, and management has told us Q3 turnaround activity is "mainly focused on conversion units in the Gulf Coast and MidCon regions." Some of this quarter's Gulf Coast advantage is borrowed from next quarter.

Mid-Continent

The region we flagged in May as the exception to the regional-leadership pillar, at $1.50 per barrel of segment EBITDA, printed $20.96. That is the single largest sequential swing in the table and it clears the FY2025 system average of $5.63 by nearly four times. It is the condition we said would support an upgrade, and on the margin line it was met without ambiguity.

The rest of the region's disclosure is less flattering. Utilization was 87%, the lowest of the three regions. Crude runs fell to 1,030 mbpd from 1,165, down 11.6%. Refining operating cost rose to $6.31 per barrel from $5.04, an increase of 25.2% and the worst cost line in the system. Turnaround expense nearly doubled to $1.93 per barrel from $1.04. So the region earned $20.96 on a shrinking, more expensive, more heavily maintained asset base, and the reason is that margin per barrel went to $33.68 from $17.86.

Assessment: Mid-Continent recovery is confirmed on the metric we said we would track, and we will not move the goalposts. But the recovery is a price recovery, not an operating one. On the two lines management actually controls, volume and cost, this region got worse year over year while the market did the work. When the Mid-Continent crack normalizes, the cost base it normalizes onto is $1.27 per barrel higher than it was a year ago.

West Coast

Segment EBITDA of $27.26 per barrel against $8.18, on the highest regional margin in the system at $41.28. This is the region where refining operating cost fell furthest year over year, to $8.08 from $8.62, and the only one where distribution cost fell at all, to $5.94 from $6.42. Crude runs rose to 515 mbpd from 485 and distillate yield rose to 215 mbpd from 179, a 20% increase, at 93% utilization.

The regional supply story was put plainly in Q&A: the state cannot import its way to balance while Asian cargoes are diverted.

"the Jones Act waiver is allowing us and the rest of the industry to make movements from the Gulf Coast into the West Coast, but those movements aren't enough to overcome the lack of Asian imports that are not coming in as they usually would due to the Middle East conflict." — Rick Hessling, Chief Commercial Officer

Assessment: The West Coast is simultaneously MPC's highest-margin and highest-cost region, and this quarter it improved on both. It is also the region most exposed to a specific and reversible cause: Asian import displacement driven by the conflict. Structural California tightness predates the conflict and survives it; the current spread does not. Note also that management flagged competitor turnarounds in the region during Q3, which is a near-term support and a reminder of how much of this margin is other people's downtime.

Midstream

Segment adjusted EBITDA of $1,778M against $1,641M, up 8.3%. This is the inflection the Q1 print did not have, and it is the reason our second bull pillar moves off its risk flag. The composition needs care, because the physical throughput data does not obviously support it.

"Segment adjusted EBITDA increased $137 million compared to the second quarter of 2025. The increase was primarily driven by higher rates and throughput, including growth from equity affiliates and acquisitions, partially offset by the divestiture of non-core gathering and processing assets." — Maria Khoury, EVP and CFO

Assessment: Two of the five disclosed physical metrics went down, including the largest one. The growth is therefore rate, equity-affiliate contribution and acquisition, which the release says explicitly. That is a legitimate way to grow a midstream business and MPLX has been buying assets that fit, but it is not the same thing as the base business expanding, and the disclosure does not let an outside investor separate the three. The divestiture drag, which we asked to have sized in May, is again named and again unsized. Third quarter is when the argument gets settled: first-half Midstream EBITDA of $3,376M annualizes to $6,752M against FY2025's $6,750M, so on a half-year basis this segment is exactly flat, and the entire mid-single-digit growth commitment has to be delivered in the second half. At a 5% full-year outcome the second half needs $3,712M, roughly 9.5% above the $3,389M it earned in the second half of 2025.

Renewable Diesel

Segment adjusted EBITDA of $258M against $(19)M, a $277M swing and the best result since the segment was broken out. Renewable Diesel margin rose to $321M from $49M. Operating cost rose modestly to $74M from $66M and distribution cost to $32M from $25M, so essentially the whole improvement is margin. Martinez ran at 95% utilization after its Q1 turnaround, ahead of the low-90s management guided in May.

"Following the completion of the Martinez turnaround in the first quarter, utilization increased to 95%, reflecting a strong operational availability." — Maria Khoury, EVP and CFO

The capital posture has not changed, and management repeated the point without being asked to soften it despite the improved returns.

"The capital that we put to work in RD is really for efficiency only, there's no change in the way we think about the allocation of capital today to the renewable segment." — Maryann Mannen, Chairman, President and CEO

Assessment: A segment that lost $110M across FY2025 has earned $296M in six months, and management is declining to add capital to it. That is the correct answer. The earnings here are a function of regulatory credit values and a short RIN balance, not of an asset that got better, and the discipline to run the asset well without expanding it is the same discipline the refining business is being credited for. The risk is that the segment's contribution is now large enough (3.0% of segment EBITDA, and a $277M year-over-year swing) that a policy reversal would be visible in consolidated results.

Corporate

Corporate expense of $256M against $243M a year ago and $274M last quarter, and guided to $260M for Q3. Management had guided $240M for this quarter in May, so this is a $16M miss on a small line.

What is missing is the explanation. The Q1 release attributed its Corporate increase to two identified items: performance-based stock compensation remeasurement and Martinez environmental remediation. The Q2 release says only: "Corporate expenses totaled $256 million in the second quarter of 2026, compared with $243 million in the second quarter of 2025." No driver, no split.

Assessment: This is a disclosure regression on the one line item we identified in May as structurally pro-cyclical. If Corporate expense scales with the share price through stock-compensation remeasurement, then in a quarter where the stock rose it should have gone up, and it did, and the release declines to say so. Small dollars, but it is the second consecutive quarter above guide on a cost line management fully controls, and the second consecutive quarter where the remediation accrual goes unquantified.

Product Yields: the Q1 Question, Answered

We closed our Q1 note with an unanswered question: management described a system running maximum diesel and jet, while the disclosed yield table showed heavy fuel oil up 51 mbpd and gasoline down 71. The Q2 table settles it.

Heavy fuel oil fell to 29 mbpd from 125 in Q1 and 61 a year ago, taking it from 4.3% of system yield in Q1 to 1.0% here. Distillates rose to 37.9% of yield from 36.1% a year ago while gasoline fell to 48.2% from 49.3%. And all of that happened while sour crude throughput was held at 48%, up from 45% a year ago. The company ran the same heavy, sour, discounted slate and moved 96 mbpd of the bottom of the barrel up into clean products.

Jet specifically was quantified for the first time.

"That resulted in a 3% increase in Jet yield year-over-year. In fact, since 2024, we've now increased our Jet yield capability from 8% to 12%." — Julian Stoll, SVP of Value Chain Optimization

Assessment: This is the most important operating disclosure in the release and the one that should update a thesis. Our Q1 concern was that the crude advantage was being paid for in yield, which would make it a trade on light-heavy differentials rather than a durable capability. The Q2 data says the opposite: same sour slate, 96 mbpd less residual fuel, 108 mbpd more distillate, and a jet yield ceiling that has risen from 8% to 12% of the barrel in two years through capital that is already spent. Q1's yield mix looks in retrospect like a turnaround-quarter artifact rather than a structural cost of the crude strategy. That is a genuine strengthening of the third bull pillar and it is checkable in the filings rather than asserted on the call.

Key Topics & Management Commentary

Overall Management Tone: Confident and noticeably more distributed than last quarter, with the commercial, planning and refining leaders each given substantive airtime and each arriving with specific, checkable content rather than framework language. The posture on the macro shifted from cautious constructiveness in May to an explicit multi-year call, which is the largest tonal change of the two quarters we have covered. Management was least convincing in exactly the same place as last quarter: asked twice about repurchase pacing, the answer both times restated priorities without describing a method, and the contrast with the granularity offered on crude sourcing and yields was again conspicuous.

1. The outage was re-measured, and the denominator changed

The single number that governs this thesis is how much refining capacity is offline and for how long. In May, management sized it as roughly 6 million barrels per day, close to 6% of global refined product capacity, attributable to the Middle East conflict. This quarter the framing is different.

"Globally, there is over 9 million bpd of planned and unplanned refined capacity downtime, approximately 4 million bpd above historical norms, reflecting ongoing Persian Gulf disruptions and accelerated Ukrainian attacks on Russian infrastructure." — Maryann Mannen, Chairman, President and CEO

These are not the same measure. The Q1 figure was conflict-attributable capacity offline. The Q2 figure is total downtime, planned and unplanned, against a historical norm. Nobody on the call asked management to bridge the two, and management did not volunteer it. Taken at face value the incremental outage is now described as 4 million barrels per day rather than 6, which would be a de-escalation, but the two figures are not comparable and no reconciliation exists.

The composition, though, did change in a way that matters, and management was specific about it.

"That's a third of Russia's refinery capacity offline, and the country's fully banned diesel export. We certainly see that it will take quite some time for the infrastructure to be repaired." — Maryann Mannen, Chairman, President and CEO

Assessment: The reversal risk did not go away, but its character shifted. Middle East capacity comes back when flows resume, which can happen on a headline. Russian capacity damaged by drone strikes comes back when it is physically rebuilt, which cannot. A quarter ago the outage was mostly of the first kind; management is now describing a meaningful component of the second. That is a genuine, if partial, hardening of the earnings base, and it is the strongest argument yet advanced for the durability of the current environment. It does not change the fact that MPC has still never published what the business earns when the capacity returns.

2. An explicit multi-year call, made twice

Last quarter management would not go past describing the timeline as unknowable. This quarter they put a horizon on it, in prepared remarks and again in Q&A.

"Looking ahead, we expect to remain in an enhanced mid-cycle environment through the end of the year and into 2027." — Maryann Mannen, Chairman, President and CEO

Pressed on whether mid-cycle itself could be structurally higher for a decade, the answer stayed constructive but stopped short of endorsing the decade framing.

"We remain constructive well into 2027. As you mentioned, the global supply, it's extremely dynamic. As we've seen, it is difficult for us to have a level of prediction as to when it will resolve." — Maryann Mannen, Chairman, President and CEO

The inventory backdrop was cited as the supporting evidence: "U.S. gasoline inventory remains well below the five-year range, while distillate inventory is at the bottom of its five-year range, underscoring continued market tightness."

Assessment: The phrase to underline is "enhanced mid-cycle," which is management conceding that the current print is above whatever they think mid-cycle is, while claiming the excess persists into 2027. That is a more honest formulation than most refiners offer at a cycle peak and it is a falsifiable one. It is also, notably, not a claim that $17.73 quarters recur. An enhanced mid-cycle that ends in 2027 and a share price that has doubled off a mid-cycle base are two different propositions.

3. Capture at 112%, and what inside it repeats

Capture went to 112% from 99%, taking the first half to 108% against 105% for all of FY2025. The most useful exchange of the call was management separating the sustainable component from the borrowed one, unprompted by the question's framing.

"There were clearly some things in that first half of the year that we think are sustainable, right? That is both the operations planning, commercial and operation excellence that allow us to optimize both in the short term and the long term, capture the prompt, and continue to generate sustainable change." — Maryann Mannen, Chairman, President and CEO

"Obviously, the volatility that was created in Q1 on derivatives, we saw the benefit of that unwind in the second quarter." — Maryann Mannen, Chairman, President and CEO

On the technology and process underneath it, the value-chain organization made a claim about scope rather than about outcome.

"As we take this capability, we're long past optimizing single refineries. We optimize regional refineries, and now we're optimizing cross-regional across the entire system." — Julian Stoll, SVP of Value Chain Optimization

Assessment: Management deserves credit for volunteering the derivative reversal as a non-repeating contributor, having flagged it as a Q1 timing item three months earlier. That is the disclosure loop closing properly. What they did not do is size it, so an investor cannot compute capture excluding it. Since the Q1 derivative drag was disclosed at roughly $500M of unrealized losses, the Q2 credit is likely to be of comparable order, which on a $9,735M R&M margin is not trivial. We will continue to track the reported 112% rather than any ex-items version, but a reader modelling forward capture should assume the sustainable figure is somewhat below it.

4. Crude sourcing broadened again, and it hedges the bear case

The commercial account of the quarter was the same shape as Q1's, with a wider set of sources. Direct SPR purchases, more than double the Venezuelan volumes of the prior quarter, record Canadian heavy on the Gulf Coast, and twice the normal California crude at Los Angeles.

"We ran more than double Venezuelan crudes in 2 Q versus what we ran in 1 Q. We ran record amounts of Canadian heavy in the Gulf Coast in the second quarter, thus reducing our exposure to Brent-based crudes." — Rick Hessling, Chief Commercial Officer

"Out on the West Coast, we ran twice as many California-based crudes than normal with advantaged economics due to the recent pier closures that we've seen in that market, making these barrels available." — Rick Hessling, Chief Commercial Officer

The most analytically important remark of the entire call came at the end of a discussion of quality differentials, and it identifies a hedge inside the thesis that we did not credit in May.

"Lastly, I would say when the conflict gets solved and you get more Middle East barrels coming into the Gulf Coast, specifically into North America, that's going to continue to put pressure on every competing grade in the Gulf Coast. When that day happens, it'll be another tailwind." — Rick Hessling, Chief Commercial Officer

Assessment: Read that quote against the bear case and it is a partial offset that is easy to miss. The event that compresses MPC's crack, medium sour barrels returning to the Atlantic basin, is the same event that widens MPC's feedstock discount, because it puts price pressure on every competing heavy grade the company already runs at scale. The two do not net to zero and the crack effect is far larger, but the crude-sourcing advantage is not simply a beneficiary of the dislocation; part of it is a hedge against the dislocation ending. That is a better structural position than we described three months ago.

5. Quality differentials are moving in MPC's favour, not against it

Our May upgrade test asked whether the crude and logistics advantage would persist at narrower light-heavy differentials. On the evidence of this call, that test has not been run, because the differentials are widening rather than narrowing.

"When we look at WCS, today, let's call it WTI, $-14 a barrel. When we look forward, if you look at Q4, Theresa, on the forward curve, it's north of $16 a barrel." — Rick Hessling, Chief Commercial Officer

Two further supply pressures were named. Venezuelan barrels arriving in volume will price to clear, which pressures competing grades directly. And roughly 110 to 111 million barrels of SPR crude have been released year to date, with management estimating "potentially another 38 million bbl that they could release yet this year. We believe that'll apply some pressure on differentials, which will be a definite tailwind for us from a feedstock perspective."

On the choice between the two heavy sources, the ranking was given without hedging: "I would tell you eight out of 10x , heavy wins when we look at the economic advantage of a heavy Canadian barrel versus Venezuela."

Assessment: This is favourable and it is also a deferral. The forward curve moving roughly $2 per barrel in MPC's direction into Q4 supports the near-term earnings case, but it means the specific question we posed in May, whether this edge survives a normalized differential, remains unanswered because the environment has not tested it. We are not going to score a test that did not happen as a pass. What can be scored is breadth: four distinct advantaged sources this quarter against three last quarter, in three different regions.

6. The capital programme is converting on schedule

Two refining projects were placed in service during the quarter, both of them on the timeline given in May: the El Paso yield improvement and the Robinson product flexibility investment. Robinson adds roughly 10 mbpd of jet capability; El Paso upgrades the fluid catalytic cracker and alkylation units to produce specialty gasolines for El Paso, Phoenix and Mexico. Management restated the hurdle these have to clear, "our targeted return of 25% or above."

Six further projects carry disclosed in-service dates, none of them earlier than year-end 2027: the Galveston Bay 90 mbpd distillate hydrotreater, Garyville feedstock optimization (30 mbpd of incremental crude throughput), and Garyville product export flexibility (10 mbpd of incremental export premium gasoline).

Assessment: Three consecutive quarters of refining projects landing in the quarter they were guided to, each one pointed at the product spread that was actually open when it arrived. The pattern is now long enough to be a capability rather than a coincidence. What is still missing, for the third consecutive quarter, is a single realized return on anything already in service. Management restates the 25% hurdle at every opportunity and has never published an outcome against it.

7. The $3.8B working-capital inflow is partly borrowed

Consolidated cash rose to $7,768M from $2,151M at March 31. The composition is the interesting part: operating cash flow excluding working capital was $6.6B, and working capital contributed a further $3.8B.

"Working capital was at $3.8 billion source of cash for the quarter, driven by higher payables, the timing benefit of crude exchanges, and inventory draws." — Maria Khoury, EVP and CFO

Both the CFO and the CEO were explicit that a portion of it reverses, and the CEO tied it directly to the SPR purchases that helped the quarter's capture.

"We are already starting to rebuild inventory, right? We have a liability and a repayment obligation for these crude exchanges, and we will be watching as well our payables." — Maria Khoury, EVP and CFO

"The only thing that's slightly different, and Maria mentioned it as well, with the SPR barrels, we have an obligation to repay those barrels in the future. That's embedded in the working capital that she spoke about." — Maryann Mannen, Chairman, President and CEO

The disclosed sensitivity is $550M of working-capital movement per $10 change in crude.

Assessment: Management handled this well and said the uncomfortable part out loud twice without being pushed. The practical consequence is that the $7.8B cash balance is not $7.8B of deployable cash, and anyone treating the balance-sheet build as a buyback war chest is double-counting the SPR barrels. It also sharpens an inconsistency: management describes the cash the business needs as "roughly $1 billion," MPC-standalone cash is around $6.7B, and yet the pace of return is described as unchanged. Either the excess is committed to the exchange repayment, in which case say how much, or it is available, in which case the pacing question has an answer.

8. MPLX inflects, and pulls 2027 capital into 2026

Midstream grew 8.3% after shrinking 7.1% last quarter, and the project list that underwrites the distribution commitment moved forward on every line. Secretariat I entered service in April, Blackcomb began commissioning in July with full commercial service expected in the fourth quarter, and Harmon Creek III began operations the week of the call, taking MPLX processing capacity to 8.1 Bcf/d and de-ethanization above 800 mbpd. Permian sour gas treating exceeded 150 MMcf/d for a second consecutive quarter against a year-end target above 400.

The new item is a capital increase that is really a schedule change.

"MPLX is increasing its 2026 growth capital spending outlook by $500 million, to $2.9 billion, primarily reflecting the accelerated execution of the Gulf Coast fractionation project to meet global demand for U.S. energy." — MPC Q2 2026 earnings release

"The increase primarily reflects the accelerated execution of the ongoing Gulf Coast fractionation project, pulling forward capital MPLX previously expected to deploy in early 2027." — Maryann Mannen, Chairman, President and CEO

Midstream capital expenditure at the MPC consolidated level rose 47.8% year over year in the quarter, to $1,021M from $691M, which is where the acceleration shows up in the numbers.

Assessment: Pulling capital forward to accelerate a fractionation project into a tight NGL market is a defensible use of a strong balance sheet, and it is not incremental spending in aggregate. The timing is what deserves scrutiny. MPLX is committing to 12.5% distribution growth in both 2026 and 2027 while spending an extra $500M in 2026 and while first-half segment EBITDA is flat year over year. That combination draws on coverage in the year it can least afford to. The offset is that the same acceleration should make 2027 both cheaper and higher-earning, which is the year the second half of the distribution commitment lands. On balance this moves our second bull pillar off its risk flag, but the pillar is now resting on a second-half delivery that has not happened yet.

9. Renewable Diesel's best quarter depends on a rule that has not been written

The segment earned $258M and management's forward case rests on the RIN balance staying short. The commercial view was that it will.

"Renewable diesel margin continues to look favorable given the fundamental RIN balance. As you probably well know, the market is short. We believe it will continue to stay short." — Rick Hessling, Chief Commercial Officer

The forward policy dependency was described in Q&A as the Set 3 rule, expected by the middle of next year, which will set 2028 and 2029 obligations. The company's position is that current obligations are unsustainable against actual domestic supply and that the market is "pulling on the bank," meaning drawing down accumulated RIN inventory rather than being served by current production.

Assessment: Management is arguing simultaneously that the RIN market is short (which is why the segment is earning) and that the obligations creating that shortage should be lowered (which would reduce what the segment earns). Both positions are internally consistent as advocacy, and both are honest, but an investor should notice that the company's stated policy preference is adverse to its own current segment economics. The right modelling posture is to treat the $258M as a policy-window result rather than a new baseline, which is also, to management's credit, exactly how they are treating it with capital.

10. Capital return: $2.5B, well executed, still unframed

MPC returned over $2.8B in the quarter, including $2.53B of repurchase, against $750M of repurchase in Q1. Shares outstanding fell to 283M from 293M, a 3.4% reduction in a single quarter, and remaining authorization stands at $6.1B from $8.6B pro forma at the end of March.

The execution deserves to be scored on its own terms. The 10 million share reduction against a $2.53B outlay implies roughly $253 per share before any offsetting issuance, in a quarter whose daily volume-weighted average price was about $245. The stock closed today at $312.61. Whatever else is true about the absence of a framework, the shares now trade roughly 24% above the price at which that capital was deployed.

The framework question was nonetheless put twice, from two angles, and produced the same non-answer both times.

"No, our capital allocation priorities are consistent. No change there. We believe the return of capital via share buyback to our shareholders continues to be the right vehicle." — Maryann Mannen, Chairman, President and CEO

"No change in the way that we think about the return of capital. Timing matters, as you can imagine." — Maryann Mannen, Chairman, President and CEO

Assessment: Our May criticism was that no method exists to slow repurchase if the cycle turns, and that criticism stands unchanged after two more direct questions. But the honest scoring of this quarter has to include the outcome: intensity rose 3.4x, into a price that has since proven to have been low. That is the pattern of a company buying when cash is available rather than when the stock is cheap, and this time those two conditions coincided. They will not always. "Timing matters" is the closest thing to a framework offered in two quarters, and it is a sentence, not a policy.

11. Management pre-guided the third quarter lower, in three separate places

The most underappreciated content on the call was a coordinated effort to lower expectations for Q3, delivered by three different executives. The CFO put a structural reason on it inside the guidance walkthrough.

"With activity mainly focused on conversion units in the Gulf Coast and MidCon regions, limiting our ability to upgrade certain products and creating a headwind to capture." — Maria Khoury, EVP and CFO

The CEO added the seasonal base rate, which is a genuinely useful disclosure and one most refiners do not volunteer.

"Third quarter, just keep in mind, if you look over 2023 to 2025, it has tended to be the most muted quarter of all four quarters. It averages 95%, which means some are higher, some are lower." — Maryann Mannen, Chairman, President and CEO

And the commercial organization gave the freshest data point available, a month into the quarter.

"While margins are good, they're not to the level they were. When I say margins, I'm specifically speaking to product margins. They're not to the level that we saw in 2Q through a month or so in of the quarter. We've seen a pullback in the jet-to-diesel spread." — Rick Hessling, Chief Commercial Officer

A related warning ran underneath the gasoline discussion: the industry-wide shift to maximum distillate is now pressuring the product it is shifting away from. "We're seeing most everyone taking the appropriate signals, and we're all in max diesel mode, and we've watched this play out before. When you're in max mode of any one product for a significant period of time, it puts pressure on other products, and we're starting to see that pull through on the gas crack as I think you're seeing in the indicators."

Assessment: Three executives independently walking the Street down from a 112% capture quarter, with a named mechanism (conversion-unit turnarounds), a named base rate (95% in Q3), and a named live observation (jet-to-diesel narrowing) is about as clear a sequential warning as a company gives without issuing formal guidance. It is creditable behaviour and it is also the most important forward-looking content in the release. The market's reaction today priced the beat; it did not obviously price this.

Guidance & Outlook

MPC does not guide to revenue or earnings. It guides to a short list of R&M cost and volume inputs, one quarter at a time, which makes each guide a falsifiable commitment. Below is how the second-quarter guide, set in May, actually landed, followed by the new third-quarter outlook.

Grading the Second-Quarter Guide

Five of nine items missed, but the two that mattered most were met precisely: crude runs to within three thousand barrels a day and utilization on the number. The misses cluster in cost, and two of them, refining operating cost and Corporate, are repeats of the same misses we flagged in May. The distribution-cost dollar figure is implied from the disclosed per-barrel rate and actual throughput; MPC guides that line in dollars and reports it per barrel, which makes the comparison one an investor has to construct rather than read.

Third-Quarter 2026 Outlook

Implied quarter-over-quarter shape: volume up 2.1% and refining operating cost down $0.12 per barrel, offset by higher turnaround expense, higher distribution cost per barrel (roughly $5.97 implied against $5.88 delivered), and higher Corporate. The controllable ledger is therefore roughly neutral to slightly negative sequentially, and the entire Q3 outcome will be decided by margin and capture, both of which management has verbally guided lower.

Where the second half sits on turnaround: the full-year refining turnaround plan was $1.35B when management reaffirmed it in May. First-half actual was $805M and the third quarter is guided at $290M, which leaves $255M implied for the fourth quarter. The Q2 release does not restate the full-year figure, and nobody asked. That is the arithmetic to check against the Q3 print.

Guidance style: volume-conservative and cost-optimistic, consistent with the pattern across both quarters we have observed. MPC has now met or beaten its crude-run guide twice in a row while missing its refining operating cost guide twice in a row. The correct modelling posture is to take the volume guide at face value or slightly better, and to add roughly $0.05 to $0.10 per barrel to the cost guide.

Analyst Q&A Highlights

What is actually inside a 112% capture rate

The opening exchange, and the one the rest of the call kept returning to, was an attempt to convert an unusually strong capture number into something modellable. The answer separated a repeatable operating capability from a set of quarter-specific market events, and named both, which is more than the question asked for.

Q: "The capture was very strong at 112%. You called out a couple things, including crude optimization and the strength of the product margin. Maybe you could just help us understand what the formula for success around the capture rate was as we think about modeling this on the go forward, too."
— Neil Mehta, Goldman Sachs

A: "One- Would be our ability to capture, take advantage of the opportunities in the prompt. Things like dislocations, supply constraints, regional asset disruptions, crude sourcing, as you saw in this quarter, and they're just a few, and we should be able to optimize around those."
— Maryann Mannen, Chairman, President and CEO

Assessment: The formula given is real but it is a description of capability, not a rate. Management then listed the quarter's specific contributors, including the derivative reversal, and declined to size any of them. An investor leaves the exchange knowing that 112% contains a non-repeating component and not knowing how large it is. That is the same gap as last quarter, in the opposite direction: in Q1 the unquantified item hurt capture, in Q2 it helped.

Whether mid-cycle itself has re-based higher

The most consequential question of the call asked management to endorse a decade-long re-rating of the industry's mid-cycle, on the argument that physical damage to Russian and Middle Eastern capacity plus depleted global inventories cannot be repaired quickly. Management stayed constructive and declined the decade.

Q: "Is there a possibility here that for the next 10 years or so, even the mid-cycle is higher given the actual unplanned downtime which will go on for some time and the global product inventory depletion, if you could talk about that?"
— Manav Gupta, UBS

A: "We remain constructive well into 2027. As you mentioned, the global supply, it's extremely dynamic. As we've seen, it is difficult for us to have a level of prediction as to when it will resolve."
— Maryann Mannen, Chairman, President and CEO

Assessment: This is the single most important boundary management drew all quarter, and it went almost unremarked. Offered a ten-year structural bull case on a platter, the CEO answered with 2027 and repeated that the resolution timing is unpredictable. Management's own horizon on the enhanced environment is roughly eighteen months. Any valuation that capitalizes current earnings into perpetuity is more bullish than the company is.

Whether a fuel-specification waiver flattered the capture number

A technically precise line of questioning tried to attribute the capture beat to a policy artifact, the seasonal vapour-pressure waiver that permits more butane blending into gasoline and that normally lifts fourth-quarter capture. Management conceded the benefit existed and sized it down twice in three sentences.

Q: "Are you benefiting from that butane blending uplift in your capture in the second quarter, and should we therefore expect that also in the third quarter?"
— Doug Leggate, Wolfe Research

A: "That benefit, certainly it would be a part of 2Q, but it is just a really small element of all of the contributors that we were trying to articulate in the second quarter. Certainly a benefit, but not one that is significant in the second quarter."
— Maryann Mannen, Chairman, President and CEO

Assessment: A clean question and a direct answer, and the answer is credible on its own arithmetic: butane blending economics move capture by tenths of a point, not by thirteen. What the exchange does establish is that the capture beat is not attributable to any single identifiable artifact, which cuts both ways. It is not a trick, and it is also not decomposable, which is why it cannot yet be modelled as a durable rate.

Capital-return pacing and the growing cash balance

Two separate lines of questioning, from two firms, converged on the same point: with cash at $7.8B against a stated operating need near $1B, is there a pacing method, is there a ceiling, and does the reluctance to deploy everything signal doubt about the durability of the margin environment. The response to both was a restatement of priorities.

Q: "Starting with the assumption that you'll be willing and able to buy back this much stock, should we think of that $3 billion-$3.5 billion as kind of a reasonable upper bound for the buybacks as you work down this cash balance?"
— Multiple analysts incl. John Royall, Piper Sandler; Doug Leggate, Wolfe Research

A: "No change in the way that we think about the return of capital. Timing matters, as you can imagine."
— Maryann Mannen, Chairman, President and CEO

Assessment: This is the third and fourth direct attempt across two quarters to obtain a repurchase framework, and the fourth non-answer. The one substantive thing that emerged is a reason for the cash to sit: the SPR crude exchanges carry a repayment obligation embedded in the working-capital balance. That is a legitimate constraint and it should have been the answer to the question rather than something disclosed separately. A company that said "we hold X billion against the exchange obligation and return the rest ratably" would have satisfied both questioners in one sentence.

Whether the full-year capture target moves up

With capture tracking at 108% through six months against 105% for all of last year, a natural question was whether the long-term target should be revised and whether recent strength is self-help or transitory. The answer used the opportunity to walk the third quarter down rather than to raise the bar.

Q: "Is it fair to say that this year's full-year capture should look even higher than last year's 105%? If that's the case, are you thinking of the long-term targets any differently today?"
— John Royall, Piper Sandler

A: "Third quarter, just keep in mind, if you look over 2023 to 2025, it has tended to be the most muted quarter of all four quarters. It averages 95%, which means some are higher, some are lower."
— Maryann Mannen, Chairman, President and CEO

Assessment: Management was handed an invitation to raise a public target and used it to lower near-term expectations instead. The seasonal base rate of 95% for the third quarter is a disclosure the company did not have to make and it is precisely the kind of thing that makes the rest of the capture commentary more believable. No long-term target was revised, which is the correct answer three quarters into an abnormal environment.

Heavy crude economics and where quality differentials go next

A detailed exchange on feedstock ranked the two heavy sources against each other, then extended into the forward curve and the policy supply that is pressuring differentials. It is the clearest public statement of MPC's feedstock position we have seen.

Q: "Would you provide some additional color on your views on quality differentials from here and key dynamics to watch for the balance of 2026 and beyond? On the Gulf Coast consumption of heavy barrels in particular, would you elaborate on the economics of running WCS versus Venezuelan crude, and how does that compare across your Gulf Coast refining system currently?"
— Theresa Chen, Barclays

A: "I'll start with how heavy compares to Venezuelan in our system. I would tell you eight out of 10x , heavy wins when we look at the economic advantage of a heavy Canadian barrel versus Venezuela."
— Rick Hessling, Chief Commercial Officer

Assessment: The full answer went further than the question, giving a current WCS differential, a fourth-quarter forward level roughly $2 per barrel wider, an estimate of remaining strategic-reserve releases, and the observation that returning Middle East barrels would pressure competing grades further. Taken together it describes a feedstock position that improves in most of the scenarios an investor would worry about, including the one where the refining crack normalizes. This was the most valuable eight minutes of the call and it received the least attention.

How much of the working-capital inflow reverses

The $3.8B working-capital source was the largest single line in the cash build and drew an immediate question about its durability. The answer was unusually specific and volunteered the sensitivity rather than waiting to be asked for it.

Q: "Working capital is such a big nut this quarter. How do we think about the unwind of that item?"
— Neil Mehta, Goldman Sachs

A: "It's good to keep in mind the sensitivity that we have communicated. For every $10 move in crude, that is about a $550 million change in working capital one way or another."
— Maria Khoury, EVP and CFO

Assessment: The best-answered question on the call. The CFO said plainly that inventory rebuild has already started, that the crude exchanges carry a repayment obligation, and that payables are being watched, then gave a quantified sensitivity. That is three specific reversals and a coefficient in about ninety seconds. It also, indirectly, undercuts the framing of the cash balance as freely deployable, which makes the unanswered buyback-pacing question look less like evasion and more like a disclosure that landed in the wrong part of the call.

What They're NOT Saying

1. The bridge between last quarter's outage number and this quarter's: In May the outage was sized at roughly 6 million barrels per day, close to 6% of global refined product capacity, attributable to the conflict. This quarter it is described as over 9 million barrels per day of total downtime, roughly 4 million above historical norms. The measures are different, the direction of the implied incremental outage is downward, and no reconciliation was offered or requested. This is the most important number in the investment case and its definition changed between quarters without comment.
2. Any scenario for what the business earns when capacity returns: Second consecutive quarter, second consecutive omission. Management now describes an "enhanced mid-cycle environment," which concedes the existence of a mid-cycle the current print exceeds, and still declines to say what that mid-cycle produces. A company willing to publish regional EBITDA per barrel and a seasonal capture base rate could publish this.
3. The size of the derivative reversal inside the 112% capture: The Q1 headwind was quantified at roughly $500M of unrealized losses when an analyst asked for it directly. The Q2 credit was named three times as a capture contributor and never sized. That asymmetry means capture is disclosed net of a helpful item this quarter and was disclosed net of a hurtful one last quarter, with only one of the two quantified.
4. The secondary-product drag, again: Named for the second consecutive quarter as the principal offset to capture, in both quarters left unquantified in dollars or in capture points. It is the only recurring capture headwind management identifies and the only capture component they will not measure publicly.
5. Any repurchase pacing framework: Asked twice this quarter and twice last quarter. The nearest thing to a method offered in six months is the phrase "timing matters." With MPC-standalone cash near $6.7B against a stated need of roughly $1B, and $6.1B of authorization remaining, the absence of a stated framework is now a two-quarter pattern rather than a single deflection.
6. Whether the full-year turnaround plan is still $1.35B: Reaffirmed in May, unmentioned in the Q2 release. First-half actual of $805M plus the $290M third-quarter guide leaves $255M for the fourth quarter. A silent full-year number after a quarter that came in $25M under its own guide is exactly the setup in which scope quietly moves, and management's May insistence that "we didn't change scope" makes the silence more notable, not less.
7. Why Corporate expense rose, and what the Martinez remediation accrual is: The Q1 release named two drivers for the Corporate increase. The Q2 release names none, on a line that missed its own guide by $16M and that we identified in May as scaling with the share price through stock-compensation remeasurement. The environmental remediation obligation at Martinez has now gone unquantified for two consecutive quarters.
8. The contribution of acquisitions and equity affiliates to Midstream growth: The release attributes the 8.3% increase to "higher rates and throughput, including growth from equity affiliates and acquisitions, partially offset by the divestiture of non-core gathering and processing assets." Two of the five disclosed physical throughput metrics fell. Without a split between rate, affiliate, acquisition and divestiture, the organic trajectory of the segment carrying the 12.5% distribution commitment is unrecoverable from the disclosure.
9. Realized returns against the stated 25% refining hurdle: Third consecutive quarter of restating the hurdle, third consecutive quarter with no realized return published for any completed project. Garyville jet flexibility has now been in service for two quarters, El Paso and Robinson for one. The company has the data.

Market Reaction

• Pre-print setup: MPC closed at $307.03 on August 3, entering the print up 88.8% year to date against +11.0% for the S&P 500, up 82.8% over trailing twelve months, and up 15.3% over the trailing 30 days. The 52-week closing range was $158.59 to $319.76.
• The record was already set, two weeks before the print: The 52-week closing high of $319.76 was made on July 21, after a run from $266.35 on July 2, a gain of 20.1% in under three weeks. The stock then drifted lower into the release.
• Reaction session: MPC opened at $308.83, a 0.6% gap, traded a $301.57 to $315.12 range, and closed at $312.61, up 1.8% or $5.58. Volume was 2.5M shares against a 2.2M 30-day average, a 1.1x day.
• Against the index: The S&P 500 rose 1.79% on the same session. MPC rose 1.82%. Excess return on a 22% earnings beat: three basis points.
• Peer reaction: The refining group did not follow. Valero closed +0.4%, Phillips 66 -0.1%, HF Sinclair -0.5%, PBF -3.0%, and Exxon Mobil -0.7%.
• Affiliate reaction: MPLX units rose 2.7%, outperforming the parent. That is the reverse of the Q1 print, when MPLX fell 2.6% on a day MPC rose 3.2%.
• Positioning: The close at $312.61 sits 2.2% below the July 21 record close. The print did not make a new high.

The cleanest way to read this session is that the stock earned nothing from the earnings. A 22.1% beat on the headline number, a record quarter by every operating measure the company reports, and the shares returned three basis points more than a passive index fund. That is not a disappointing reaction to a good quarter; it is the market saying the quarter was already in the price, and the July path shows exactly when it got there. Between July 2 and July 21 MPC rose 20% on a tape where cracks were visible in real time to anyone watching the indicators. The print confirmed what the tape had already discounted.

The peer divergence sharpens rather than softens that reading. In May, MPC outperformed Valero and Phillips 66 by 220 to 250 basis points on its print, and we argued the differentiated portion of the move was company-specific content. This time the group was flat to lower while MPC merely matched the index. Whatever MPC delivered that its peers did not, the market declined to pay a premium for it on the day. Either the operating outperformance is now assumed, or investors are looking through it to the same reversal risk the company will not quantify.

The MPLX reversal is the most informative signal in the table and it is a vote on a specific disclosure. Last quarter midstream shrank 7.1% and the units fell on the print. This quarter midstream grew 8.3%, the project list moved forward on every line, and the units outran the parent by nearly a full percentage point. Investors are marking the two halves of this enterprise separately and in the correct directions two quarters running, which is a reason to take the market's flat verdict on the refining half seriously rather than dismiss it as inattention.

Street Perspective

Debate: Has the earnings base hardened, or just moved?

Bull view: The bull case being made on the Street is that the composition of the supply outage changed this quarter in a way that materially extends its life. Middle Eastern capacity returns when flows resume; Russian capacity damaged by sustained drone attack, now described as a third of the country's refining base with diesel exports fully banned, returns only when it is rebuilt. Add gasoline inventories below the five-year range and distillate at the bottom of it, and the argument is that the tightness now has a physical floor rather than a political one.

Bear view: The bear camp contends that the company changed the way it measures the outage between quarters and that the new measure, roughly 4 million barrels per day above historical norms, is smaller than the old one. On this view the re-framing is a quiet admission that the dislocation is narrowing, the "enhanced mid-cycle" language is management pre-positioning for a step-down, and three separate executives walking the third quarter lower on the same call is the tell.

Our take: The bull has genuinely gained ground since May and the bear has kept the more important point. Russian capacity is a harder outage to reverse than Gulf capacity and that is a real change to the shape of the risk, not a talking point. But the company's own stated horizon is "well into 2027," not a decade, and management declined an explicit invitation to endorse a permanently higher mid-cycle. When a management team that has published regional per-barrel economics and a seasonal capture base rate still will not publish a reversal scenario after two quarters of being asked, the reasonable inference is that the answer is unattractive rather than unknown.

Debate: Is the buyback a discipline problem or was that always the wrong frame?

Bull view: Some sell-side desks argue the second quarter settles this empirically. MPC spent $2.53B retiring 3.4% of its shares at an implied average near $253, in a quarter whose volume-weighted average price was about $245, and the stock now trades at $312.61. Demanding a formal pacing policy is asking a company to publish a market-timing framework it would then be judged against; the outcome is the framework, and the stock now trades roughly 24% above the average price paid.

Bear view: The bear camp contends that a good outcome in one quarter is not a policy, and that the pattern being established is intensity scaling with available cash rather than with valuation. Repurchase went 3.4x sequentially in the quarter cash arrived, not in the quarter the stock was cheapest, and the same company is now holding $6.1B of authorization against a stock that has doubled. Refiners have a long record of buying hardest at the top for exactly this reason.

Our take: We raised this concern in May and we owe the company the update: on execution, this quarter was good, and we are not going to grade a favourable outcome as if it were a bad one. The criticism narrows rather than disappears. What remains is that the deployment decision this quarter was driven by cash availability, and cash availability at a refiner peaks with the cycle. The single sentence that would resolve it already exists inside the company's own disclosure: hold a stated amount against the crude-exchange repayment obligation and return the balance on a stated cadence. Management gave both halves of that answer on this call, in different sections, and never joined them.

Debate: Is MPLX finally worth what MPC says it is?

Bull view: A growing consensus view is that the second quarter is the proof point the structure needed. Midstream grew 8.3% after a shrinking quarter, every project on the disclosed list advanced or entered service, capacity reached 8.1 Bcf/d of processing and over 800 mbpd of de-ethanization, and the market paid for it by bidding the units above the parent on the print. The distributions cover MPC's standalone capital programme and its dividend, which is what allows all of the refining cash flow to be returned rather than retained.

Bear view: The bear camp contends the growth is bought rather than grown. Two of five disclosed physical throughput metrics fell year over year, the release attributes the increase to rates, equity affiliates and acquisitions without splitting them, and first-half segment EBITDA of $3,376M annualizes to $6,752M against $6,750M for all of FY2025, which is flat. Into that, MPLX is adding $500M of 2026 capital and committing to 12.5% distribution growth in both 2026 and 2027.

Our take: The bull case is now ahead on evidence and the bear case is ahead on arithmetic, and the arithmetic resolves within two quarters. To deliver even 5% full-year growth, the second half needs $3,712M against the $3,389M it earned in the second half of 2025, roughly 9.5%. The second quarter's 8.3% and the August start of Harmon Creek III make that reachable; nothing about the first half makes it comfortable. We are moving this pillar off its risk flag on the strength of the inflection and the project delivery, and we will move it back if the third quarter does not compound on it. Separately, we repeat the request from May: no consolidated multiple computed off this income statement means much when $25,640M of the $32,816M of debt and $400M of the quarter's net income belong to a partnership MPC does not wholly own. An MPC-standalone cash flow bridge remains the largest gap in an otherwise improving disclosure package.

Model Implications

Aardvark Labs does not yet maintain a published financial model for MPC. Rather than present deltas against a model that does not exist, the table below carries forward the driver set we established in May and marks where this quarter moved each one, so the third quarter can be graded against a stated starting point.

The Valuation Bracket

At the reaction close of $312.61, MPC's market capitalization is roughly $88.5B and enterprise value roughly $113.5B against $32,816M of consolidated debt and $7,768M of cash. That enterprise value is 9.5x FY2025 adjusted EBITDA of $11,956M, and 5.1x the first half of 2026 annualized at $22,446M. At the Q1 print the same trailing measure stood at 8.9x.

The same bracket on earnings: first-half adjusted EPS of $19.22 doubled is $38.44, which puts the shares at roughly 8.1x. FY2025 adjusted EPS of $10.70 puts them at roughly 29.2x. Neither number is a forecast and neither is the answer. The distance between 8.1x and 29.2x is the entire investment debate expressed as a multiple, and no amount of operating excellence collapses it. What collapses it is a view on where the refining margin settles, which is the one thing management will not publish.

Valuation impact: The business is worth more than it was in May, because the second bull pillar inflected, the yield question resolved favourably, and part of the outage hardened. The shares are also 20% more expensive. We do not think those two facts have netted to a rating change in either direction.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation in May, scored against this quarter's print and call. They are unchanged in wording, which is the point: the scorecard is only useful if it tracks the same thesis quarter after quarter.

Overall: thesis strengthened, on the bull side only. All three bull pillars are confirmed by this quarter's disclosure, one of them moving off a risk flag, and the crude-and-yield pillar is now supported by filing data rather than by management assertion. Neither of the two governance and cycle risks improved, and the cost risk we had marked contained is now emerging. The net is a better business than we described in May with the same two unanswered questions attached to it.

Action: Hold. First, the scorecard on ourselves. Since the May initiation MPC has returned 20.0% against 6.6% for the S&P 500, so the standing Hold has underperformed by 13.4 points in a single quarter. A rating that costs that much in three months has to be re-argued from scratch rather than rolled forward, and this section is that argument. We set three upgrade conditions in May and required two. Mid-Continent recovery above the FY2025 average alongside continued regional leadership was met emphatically. A stated repurchase framework was not met. Evidence that the crude and logistics advantage persists at narrower light-heavy differentials was not tested, because differentials widened rather than narrowed; the advantage broadened instead, which is favourable but is not the test we specified. On a strict reading that is one condition met and one deferred. The share price has meanwhile risen 20% since the May note, and a 22% earnings beat produced three basis points of excess return against the index. We will upgrade on a stated capital-return framework, on evidence the crude edge holds through a normalized differential, or on a price that reflects something closer to the mid-cycle management itself concedes exists. We would not chase a stock that has doubled into a quarter management has already told us will be weaker.

Bottom Line

Marathon Petroleum earned $17.73 a share in three months, more than it earned in all of 2024, and did it without a single non-GAAP adjustment. Capture reached 112%, all three regions cleared $20 a barrel of segment EBITDA, the Gulf Coast ran at 100% utilization, midstream finally grew, and the yield question we left open in May was answered in the company's favour by the filings rather than by the call. Two of the three commitments management made in May on volume were delivered to within a rounding error. This is a very good company having a very good quarter, and nothing in the print argues otherwise.

The market's answer was three basis points. The shares had already risen 20% between July 2 and July 21, setting their record close a fortnight before anyone saw these numbers. What the print did was confirm a macro that was fully discounted, and the refining group's flat-to-lower session on the same day says the confirmation was worth nothing incremental even to a company that outperformed its peers on every disclosed operating metric.

That leaves the same two questions we asked in May, both still unanswered and one of them now harder to avoid. What does this business earn when four to nine million barrels a day of capacity comes back, on management's own admission that the current environment is an enhanced mid-cycle rather than a new normal? And on what basis does a company holding $7.8 billion of cash against a stated need of $1 billion decide how fast to retire stock at a doubled share price? The second quarter answered neither, and it did not need to: the numbers were good enough that nobody had to. The third quarter, which management has now pre-guided lower in three separate voices, is where both questions get asked properly.