Deleveraging Arrives a Year Early, on Prices bp's Own Planning Deck Says Will Not Last

Results vs. Consensus

Q2 2026 Scorecard

bp does not guide to revenue and no bp-covering analyst writes to it, because roughly a fifth of the top line is a commodity-derivative gross-up that swings with price rather than with profitability. Published revenue estimates for the quarter spanned $57.9B to $61.8B against a reported $69,105M, which tells you about the estimates rather than about the company. The line the Street actually models is underlying replacement cost profit, and there the pre-print consensus was tightly clustered.

Year-Over-Year Comparisons

Quarter-Over-Quarter Comparisons

Quality of Beat

Revenue and the shape of the top line

Sales and other operating revenues of $69,105M rose 48.2% year over year and 32.2% sequentially, and neither figure should be read as a volume statement. Reported upstream production fell 4.3% year over year and 5.9% sequentially, and marketing sales of refined products of 2,624 mb/d were 3.5% below the 2,720 mb/d of a year earlier. The entire top-line increase is price, mix and trading gross-up. Refinery throughput of 1,467 mb/d was the one genuine volume gain, up 13.9% year over year against a second quarter of 2025 that carried a much heavier turnaround programme.

The useful read on the top line is what happened beneath it. Total revenues and other income of $70,114M sat $1,009M above sales and other operating revenues, and that spread narrowed from $1,116M in Q1 as gains on the sale of businesses and fixed assets fell to $15M from $102M. Disposal gains are not carrying this quarter's earnings, which is worth saying plainly given the volume of portfolio activity announced alongside it.

Margins and the realization problem, one quarter on

This is the line item we told readers to watch, and it improved materially. bp's average liquids realization of $85.14/bbl represents 82.0% of the $103.85/bbl Brent marker. Last quarter that capture ratio was 74.5%; in the second quarter of 2025 it was 88.6%. Measured against that prior-year baseline, the deficit narrowed from 14.1 points to 6.6 points, so slightly more than half of the disconnect closed in a single quarter. That is close to the test we set in April.

It is also a better result than the raw numbers suggest, because capture ratios mechanically deteriorate as the marker rises: production-sharing and technical-service entitlement volumes shrink at higher prices, and price caps bind harder. bp held 82.0% capture into a marker that rose 28.0% sequentially. Both upstream segments participated. Oil production and operations realized $84.10/bbl in liquids against $59.75/bbl in Q1, and gas and low carbon energy realized $94.09/bbl against $67.17/bbl.

The regional split is where the residual problem sits. European liquids realizations of $97.49/bbl and rest-of-world realizations of $95.39/bbl are close to marker. US realizations of $76.03/bbl against a West Texas Intermediate marker of $93.11/bbl are not. Natural gas tells the same story in reverse: the US realized $1.72/mcf against a Henry Hub first-of-month index of $2.90/mmBtu while Europe realized $17.25/mcf, and the group average of $5.89/mcf rose only 5.9% year over year because the two moved in opposite directions. bp's American gas book is capturing a fraction of an already weak marker, and nothing in this quarter's disclosure explains why.

Earnings per ADS: the number that matters and the two that do not

Three profit figures are available for the quarter and only one of them is analytically useful. Profit attributable to bp shareholders on an IFRS basis was $3,911M, up 1.8% sequentially, because Q1 carried a $4,159M inventory holding gain and Q2 carried an $870M inventory holding loss, a $5.0B swing that has nothing to do with operations. RC profit attributable to shareholders of $4,628M strips the inventory effect and rises 599% sequentially off a Q1 base that had been crushed by $2,536M of post-tax adjusting items. Underlying RC profit of $5,732M, or $2.22 per ADS, strips both and is the figure the company, the Street and this report use.

The bridge from the underlying number to the reported one is worth naming: $1,104M of post-tax net adverse adjusting items, including roughly $800M of post-tax net impairments concentrated in the transition businesses inside gas and low carbon energy, offset by a favourable $1.0B pre-tax fair value accounting effect. bp has now taken impairments in the low carbon portfolio in consecutive quarters while telling investors it is reducing capital in exactly that place, which is a coherent sequence but not a costless one. The underlying effective tax rate of 34% was two points above Q1's 32% on geographic mix, while the headline effective tax rate on profit before taxation was 45%.

Segment Performance

Underlying RC Profit Before Interest and Tax by Segment

Customers & Products by Business

Upstream Production and Realizations

Downstream Operating Metrics

Customers and Products: Still the Largest Segment, but No Longer the Whole Story

Customers and products earned $4,954M of underlying RC profit before interest and tax, 48.0% of the group total, up from $1,533M a year earlier. Within it, refining and trading earned $3,183M against $477M in the second quarter of 2025 and $2,194M last quarter. The refining indicator margin averaged $29.6/bbl, two and a half times its year-ago level, and refining sits at the exact intersection of the physical dislocation and bp's asset base.

The more interesting movement was on the customers side. Convenience and mobility earned $1,771M against $1,009M last quarter and $1,056M a year ago, an increase management attributed to seasonally higher volumes, higher fuels margins, a stronger Castrol result and a slightly higher midstream contribution, partly offset by lower earnings from bioenergy. Castrol itself earned $429M, up 75.1% year over year, which is a notable data point three quarters into a sale process that values 65% of that business at roughly $6B.

"Within customers, profit benefited from seasonally higher volumes, higher fuel margins, a stronger Castrol performance and a slightly higher midstream contribution, partly offset by lower earnings from bioenergy. Within products, profit benefited from significantly stronger realized refining margins and a slightly higher oil trading contribution, partly offset by higher planned turnaround and maintenance activity as well as the impacts of the third-party event at Whiting in April."
— Kate Thomson, Chief Financial Officer

Assessment: the segment's share of group underlying profit fell to 48.0% from 51.1% last quarter even as its absolute earnings rose 54.7%, which is what a broadening earnings base looks like. Refining and trading's share of the group fell 410 basis points to 30.9%. Investors should still discount this line heavily, because bp continues to disclose refining and trading as a single number and the oil trading contribution inside it was described only as "slightly higher." One quarter of $3,183M against $477M a year earlier is not a run rate.

Oil Production and Operations: The Lag Unwinds

Underlying RC profit before interest and tax of $3,581M rose 80.8% sequentially and 58.3% year over year, and the driver is exactly the reversal we flagged in April. Liquids realizations of $84.10/bbl compare with $59.75/bbl last quarter and $59.74/bbl a year ago, a near-identical base that makes the point cleanly: bp earned 40.8% more per barrel of liquids than it did a year ago on 10.6% fewer liquids barrels.

Volumes went the other way. Reported production of 1,436 mboe/d fell 5.4% year over year and 6.8% sequentially on Gulf of America seasonal maintenance, the divestment of Culzean, a reduced equity interest in Pan American Energy after the 50% to 40% step-down completed in June, and Middle East disruption. bp's underlying production measure, which strips portfolio effects, was 1.8% higher in the quarter and 3.8% higher across the half, driven by bpx Energy and major project ramp-ups. Exploration write-offs of $478M, almost entirely the exit from Bay du Nord in Canada, were the largest single drag on the segment.

"In Oil Production and Operations, segment underlying operating profit increased by around $1.6 billion. This reflected higher liquids realizations, including the impact of price lags, production mix benefit and higher income from equity accounted entities. These positive factors were partly offset by higher exploration write-offs, mainly related to exiting Bay du Nord and lower production due to seasonal maintenance in the Gulf of America."
— Kate Thomson, Chief Financial Officer

Assessment: the timing hypothesis won. Roughly $900M of price-lag drag was disclosed across the two upstream segments in Q1 and the majority of it has now converted, which validates the bull framing of that debate and removes the most acute version of the marker-to-realization bear point. What has not been resolved is the US gap, where a $76.03/bbl realization against a $93.11/bbl West Texas Intermediate marker cannot be a lag effect three months later. The structural residual is smaller than we feared in April and larger than zero.

Gas and Low Carbon Energy: Europe Carries It

Underlying RC profit before interest and tax of $2,122M rose 58.8% sequentially and 45.1% year over year, and the composition is unusual. Liquids realizations of $94.09/bbl did most of the work alongside a lower depreciation charge, while the gas marketing and trading result was described as broadly flat sequentially and slightly lower year over year. This is a gas segment whose quarter was made by oil and by European gas prices, not by its trading book.

Production of 765 mboe/d fell 2.1% year over year on base decline partly offset by major project ramp-up. Depreciation of $1,139M fell 19.0% from $1,407M a year earlier, which flatters the comparison, and the segment absorbed $558M of adverse adjusting items, the bulk of the group's impairment charge on transition businesses.

"In Gas and Low Carbon Energy, segment underlying operating profit increased by around $800 million, reflecting higher realizations, including the impact of price lags with gas marketing and trading broadly flat compared with the first quarter."
— Kate Thomson, Chief Financial Officer

Assessment: a strong number with a soft interior. The gas trading book, which is supposed to be one of the crown jewels, produced an average result in a quarter when bp's own European gas realization nearly doubled sequentially, from $8.76/mcf to $17.25/mcf. The marker behind that realization moved far less: NBP averaged 112.4 p/therm against 100.9 p/therm, up 11.4%. The distinction is worth holding onto, because management's explanation is that gas volatility has been benign relative to oil and products, and the marker supports them. A trading book earns on dislocation, not on a high realized price. The segment's quarter was therefore made by equity gas rather than by the book, and the narrower criticism that survives is that the crown jewel added nothing to a quarter its own commodity handed it. Meanwhile the low carbon side is where the impairments keep landing.

Other Businesses and Corporate: The Drag Widens

The underlying RC loss before interest and tax widened to $345M from $272M last quarter and $38M a year ago, which management attributed to the impacts of the Ventures divestments and one-off corporate items. bp reached agreement in July to sell the majority of its direct bp Ventures investments and intends to close bp Ventures entirely. Full-year guidance for the segment charge remains around $1.0B, and the first half has already consumed $617M of it.

Assessment: immaterial in a $10.3B quarter, but the trajectory is the wrong way and the full-year guide now requires the second half to run at roughly $383M against a first half of $617M. That is a tightening, not a loosening, and it sits alongside a reorganization whose restructuring charge has still not been quantified.

Key Topics & Management Commentary

Overall Management Tone: markedly more self-critical than any bp call of the past two years, and deliberately so. Management led with what went wrong (safety, reliability, costs not reaching the bottom line) before presenting a quarter that beat consensus by 14.6%, and framed the five priorities as a diagnosis rather than an announcement. Where the posture weakened was on distributions and on the restructuring cost of the reorganization, both of which drew direct questions and both of which were answered with a commitment to communicate later. Compared with April, the tone shifted from defending a plan inherited from a previous chief executive to owning a diagnosis, which is a genuine change and not merely a change of speaker.

1. A New Chief Executive's First Full Quarter

Meg O'Neill became chief executive on 1 April 2026 and marked her hundredth day shortly before this print. Her first full quarter produced a strategic framing built around five priorities: strengthening the balance sheet, simplifying the portfolio, investing with discipline, driving operational excellence, and hardwiring accountability. The framing that carried the call was "get fit to grow," and its substance is that growth has to be earned before it is funded.

"Going forward, every part of the company needs to earn its place, generating cash, improving returns and strengthening the whole. We need to improve the quality of our earnings and cash generation and unlock more value for shareholders."
— Meg O'Neill, Chief Executive Officer

What distinguishes this from the usual first-quarter-of-a-new-chief-executive reset is the willingness to name the failures directly and in the results announcement rather than in a private meeting.

"We have not delivered consistently enough across our operations. We have written off too much shareholder value, and we face a challenge of liabilities and costs that means our resilience to a low price environment is insufficient, exacerbated by a portfolio that is too stretched and too complex."
— Meg O'Neill, Chief Executive Officer

Assessment: credible and unusually specific, but note what the five priorities are not. Four of the five are inward-facing operational disciplines and the fifth is culture. There is no distribution policy, no production target, no return-on-capital commitment beyond the existing 2027 targets, and no capital markets day announced. A diagnosis is worth something; this one is worth more than most because it is accompanied by a cost bridge and a portfolio framework rather than adjectives. It is still a diagnosis.

2. The Price Environment Is the Quarter

Brent averaged $103.85/bbl in the second quarter against $81.13/bbl in the first and $67.88/bbl a year ago. bp's refining indicator margin averaged $29.6/bbl against $11.9/bbl a year ago. For context on how far outside normal this is, bp's own 2026 planning assumptions, disclosed in the same slide pack, are $72.9/bbl Brent and $11.2/bbl refining indicator margin in 2024 real terms. The quarter was earned at a Brent price 42% above the company's planning deck and a refining margin 164% above it.

Applying bp's published full-year rules of thumb to that gap gives a sense of scale. At $340M of pre-tax profit per $1/bbl of Brent and $450M per $1/bbl of refining indicator margin (the post-Gelsenkirchen sensitivity), a single quarter at $30.95/bbl of Brent and $18.4/bbl of refining margin above the deck is worth roughly $2.6B and $2.1B of pre-tax profit respectively.

Assessment: this is not a criticism of bp, which does not set Brent. It is a statement about what an investor is buying. The company earned $10.3B of underlying pre-interest profit in a quarter that its own planning process does not expect to repeat, and the disclosure that quantifies this most cleanly is bp's own price-adjusted free cash flow restatement discussed below.

3. The Realization Gap Closes by Roughly Half

In April we set an explicit test: if the second quarter had a comparable Brent average and the marker-to-realization gap did not close by roughly half, the disconnect was structural rather than mechanical. Brent was not comparable, it rose 28.0%, which makes the test harder rather than easier. The capture ratio nonetheless improved from 74.5% to 82.0%, recovering 7.5 of the 14.1 points of deficit against the prior-year baseline of 88.6%.

bp had pre-announced the mechanism on 14 July: realizations were expected to add $0.5B to $0.7B in gas and low carbon energy and $1.8B to $2.1B in oil production and operations, "including the impact of the price lags on bp's production in the Gulf of America and the UAE." The gas segment delivered roughly $786M and the oil segment roughly $1,600M, both at or above the top of their ranges.

Assessment: the bull side of April's debate was right on the main point. Gulf of America one-month lags and UAE two-month lags did roll through and did convert. The residual is regional and specific: the US at 81.7% capture against West Texas Intermediate is the outlier, and US gas at $1.72/mcf against a $2.90/mmBtu Henry Hub index is worse. We are downgrading the marker-to-realization bear point from materializing to contained, and moving the remaining question from "is it structural" to "why is the American book so much weaker than the European one."

4. Refining: The Indicator Margin Doubles Again and the Realized Number Is Still Not Disclosed

The refining indicator margin averaged $29.6/bbl, up 75.1% sequentially and 148.7% year over year. Products, refining and trading earned $3,183M. Throughput of 1,467 mb/d fell 3.9% sequentially on higher planned turnaround activity and a third-party event at the Whiting refinery in April, and refining availability fell 160 basis points to 94.7%.

In April, management told investors the gap between the indicator margin and bp's realized refining margin could exceed $5/bbl if conditions persisted, and promised to "give as much color as we can once we get to the trading statement for the second quarter." The trading statement gave a dollar range for the sequential earnings effect. It did not give a realized margin, a gap, or a per-barrel figure, and neither did the results. Investors are still left inferring the realized number from a segment line that also contains an undisclosed oil trading result.

The one structural change is the sensitivity itself. Following the completion of the Gelsenkirchen sale on 31 July, bp cut its refining rule of thumb to $450M of pre-tax profit per $1/bbl from $550M, an 18.2% reduction in the group's leverage to refining margins.

Assessment: a promised disclosure was not delivered, and it is the disclosure that would make the largest single profit line in the company modellable. bp has now cut its own exposure to refining margins by 18.2% at the exact moment those margins are at a cyclical extreme, which is either good portfolio discipline or unfortunate timing depending on where margins go from here. On our reading it is the former, because Gelsenkirchen was a structurally high-cost asset, but the sale does mean the second half earns less per dollar of margin than the first half did.

5. The Balance Sheet Inflects

This is the most important thing that happened in the quarter. Financial obligations, bp's own composite of net debt, hybrid bonds and securities, lease liabilities and Gulf of America settlement liabilities, fell to $53.6B from $60.5B, a $6.9B reduction in three months. The components: net debt down $3,058M to $22,251M, hybrid bonds down $3.0B to $13.0B, the Gulf of America settlement liability net of deferred tax assets down $0.9B to $5.0B, and lease liabilities flat at $13.3B.

The hybrid redemption was the commitment we asked management to execute and it executed on schedule: BP Capital Markets exercised its option to redeem €2.5B of hybrid bonds on 22 June, a $2.9B cash payment, alongside $1,129M of pre-tax Gulf of America consent-decree payments. Gearing fell to 22.6% from 24.7%, and gearing including leases to 31.8% from 33.4%.

"This quarter, stronger earnings converted into stronger cash generation, helping us to reduce financial obligations by around $7 billion. Now rather than follow the cash flow statement line by line, I want to walk through the quarter's sources and uses of cash, showing how cash generated by the business flowed through to net debt and financial obligations."
— Kate Thomson, Chief Financial Officer

Assessment: the plan we called credible-but-unproven in April is now proven for one quarter. The balance-sheet inflection bull pillar is doing exactly what the thesis said it would, and it is doing it faster than the schedule. This is the strongest single argument for owning the shares and it is why conviction rises even though the rating does not.

6. The $39 to $41 Billion Guide, and What Sits Underneath It

Management now expects financial obligations of roughly $39B to $41B by the end of 2026, from $53.6B today, and says this brings the $14B to $18B net debt target forward into 2026 from the end-2027 schedule.

"On the stated price assumptions, we expect to see financial obligations reducing to around $39 billion to $41 billion by the end of 2026. This would mean delivering our $14 billion to $18 billion net debt target ahead of plan, including our plan to repay $1 billion of perpetual hybrid securities in the third quarter."
— Kate Thomson, Chief Financial Officer

The phrase doing the work is "on the stated price assumptions." Those assumptions, disclosed on the same slide, are second-half Brent of $80/bbl, Henry Hub of $3/mmBtu and a refining indicator margin of $28/bbl, alongside $7B to $8B of divestment proceeds, a $2B to $3B working capital release and $7B to $7.5B of capital expenditure. The refining margin assumption is the striking one: $28/bbl for the second half against a full-year planning assumption of $11.2/bbl, and against a first-half actual of $23.3/bbl. bp is assuming refining margins go up from here.

The sensitivity is calculable from bp's own rules of thumb. Half-year sensitivities are half the published annual figures: $170M per $1/bbl of Brent, $200M per $1.00/mmBtu of Henry Hub, and $225M per $1/bbl of refining indicator margin post-Gelsenkirchen. Against the planning deck, the assumed second half carries $7.1/bbl of extra Brent (roughly $1.2B), $16.8/bbl of extra refining margin (roughly $3.8B), and $1.20/mmBtu less Henry Hub (roughly negative $0.2B). Net, roughly $4.7B of second-half pre-tax profit is attributable to prices above bp's own deck.

Management said the quiet part out loud immediately afterward.

"But I want to be clear that at that level, there would still be more to do. We will continue reducing financial obligations beyond 2026 with organic cash generation and further expected divestment proceeds."
— Kate Thomson, Chief Financial Officer

Assessment: the guide is achievable and the largest components of it are not price-dependent. Roughly $7B to $8B of the $12.6B to $14.6B reduction comes from divestments and $2B to $3B from working capital, so the arithmetic does not actually require the dislocation to persist. What it does require is Castrol closing, which is a single regulatory event carrying roughly $6B. The price assumptions are the difference between comfortably hitting the range and scraping it, not between hitting and missing. That is a better answer than we expected in April, and it is why this pillar stays on track.

7. Costs: The Best New Disclosure of the Print, and What It Concedes

In April management deferred the restructuring question to a second-quarter "deep dive into costs generally." That deep dive was delivered and it is the most useful new disclosure bp has produced in several quarters. Production and manufacturing expenses plus distribution and administration expenses are now split into variable costs, which are activity-linked and belong alongside the gross profit they generate, and underlying operating expenditure, which is the structural cost base. First-half variable costs rose 47% while the related gross profit rose 50%. First-half underlying operating expenditure was $10,702M against $10,761M a year earlier, down 0.5%.

The concession embedded in the disclosure is significant. Cumulative structural cost reductions since the programme began stand at $3.5B ($0.8B in 2024, $2.0B in 2025, $0.8B in the first half of 2026) against a $4B to $5B target, or $6.5B to $7.5B including the Castrol and Gelsenkirchen transactions. And the bridge from 2025 underlying operating expenditure of $21.9B to roughly $18B in 2027 shows inflation adding $1.0B, environment $0.2B and activity $0.6B, against $3.0B of structural cost reductions, roughly $1.5B from divesting non-integrated businesses and roughly $1.5B from further high-grading. Half of the cost reduction is selling businesses.

"Here, we are disappointed that underlying operating expenditure is not coming down quickly enough. Since the start of the program, we have delivered $3.5 billion of structural cost reductions, but the benefits are not yet sufficiently visible in earnings and cash flow. The actions taken so far have not been sufficient to overcome inflation, some acquired costs and the complexity of our cost base."
— Kate Thomson, Chief Financial Officer

Assessment: bp deserves credit for a disclosure that makes its own record look worse. On the numbers, $3.0B of genuine structural reduction over two years is offset by $1.8B of inflation, environment and activity headwinds, leaving roughly $1.2B of net efficiency on a $21.9B base, or about 5.5%. The rest of the journey to $18B is portfolio arithmetic that also removes the earnings those businesses generate. The restructuring charge for the upstream and downstream reorganization, deferred from April to this quarter, was still not quantified.

8. Portfolio: Five Processes Running at Once

The quarter's transaction list is long. Completed: the Gelsenkirchen refinery sale to Klesch on 31 July, and the reduction of the Pan American Energy Group shareholding from 50% to 40% in June. Agreed: the Austrian mobility, convenience and bp pulse businesses to volenergy AG; Bay du Nord to Equinor; a 42% interest in BP Energy Company of Kirkuk to ConocoPhillips and a further 15% to TPAO; a 5% interest in Browse to GS Energy; and the majority of direct bp Ventures investments, after which bp Ventures closes. Launched: marketing processes for the UK North Sea business and for Archaea Energy. Pending: the 65% Castrol sale to Stonepeak at roughly $6B, targeted to complete by the end of 2026.

"We are taking an objective view asset by asset, business by business, looking at cash generation, returns, capital efficiency and strategic fit."
— Meg O'Neill, Chief Executive Officer

Assessment: the pace is real and the framework presented alongside it, a two-axis chart of free cash flow against returns by asset over three years, is more analytical rigour than bp has shown publicly in years. Two cautions. First, management conceded the chart's axes are historic and therefore penalise anything recently invested in. Second, an organization running five concurrent marketing processes while executing a segment reorganization and searching for a chair is an organization with a lot of leadership bandwidth committed to transactions rather than operations, in the same quarter its plant reliability fell 330 basis points.

9. Safety and Reliability Went Backwards

Tier 1 and Tier 2 process safety events rose to 18 in the quarter from 7 in the first quarter and 5 in the second quarter of 2025, taking the first half to 25 against 15 a year earlier. A Castrol colleague died following an incident at the Gemlik blending plant in Turkiye in April. bp-operated upstream plant reliability fell to 92.4% from 95.7%, and refining availability to 94.7% from 96.3%.

"On process safety, we saw an increase of events in the first half of 2026 when compared with the same period in 2025, including an increase in Tier 1 events. Nothing is more important than the safety of our people."
— Meg O'Neill, Chief Executive Officer

The operational causes were specific: the Glen Lyon floating production vessel offline for a couple of months and trips at ETAP in the North Sea, an extended turnaround with restart problems in Indonesia, and the third-party event at Whiting.

Assessment: a 2.6 times sequential increase in process safety events is not a rounding error and it is not explained by the individual operational failures management listed, which were availability events rather than process safety events. bp disclosed the number in a table and discussed the category qualitatively without ever addressing the figure itself. For a company whose equity story has twice been reset by a safety failure, that is the wrong emphasis, and it is the single item in this print that most warrants escalation rather than monitoring.

10. Distributions: The Minimum Dividend and Still No Buyback Condition

bp raised the quarterly dividend 4.1% to 8.660 cents per ordinary share, equivalent to $0.5196 per ADS, which is the policy minimum and nothing more. Buybacks have now been suspended since February 2026. In April, two direct questions on the restart produced no condition. In August, a direct question produced the closest thing to a numeric threshold bp has yet offered.

"So we fully understand that shareholders are keen to have increasing TSR and we're fully committed to that. But we've got work to do on the balance sheet, $40 billion of total liabilities is still too much. We are still not going to be able to offer that resilience through the cycle that we need to be able to. So we've got a bit of work to do."
— Meg O'Neill, Chief Executive Officer

There is a mechanical consequence of the suspension that is easy to miss. Shares in issue rose to 15,546,713 thousand ordinary shares at 30 June from 15,496,882 thousand at 31 March, an ADS-equivalent count of 2,591,118 thousand against 2,582,813 thousand. The share count went up 0.3% sequentially and is down only 0.3% year over year.

Assessment: "$40 billion of total liabilities is still too much" is the first quantitative anchor management has put on this question, and read against a guide of $39B to $41B by the end of 2026 it says the distribution conversation is a 2027 event at the earliest, on management's own arithmetic. That is consistent with what we assumed in April and it is the primary reason the rating does not move. An equity holder is being asked to wait through another five quarters of balance-sheet optimization while the share count drifts up.

11. Capital Expenditure Up, Divestment Proceeds Down

Full-year capital expenditure guidance rose to $13.5B to $14.0B from $13B to $13.5B, and full-year divestment and other proceeds guidance fell to $8B to $9B from $9B to $10B. Both moves are unhelpful to the deleveraging arithmetic and together represent a $1.5B swing, and both have the same cause: bp deferred a planned farm-down of Paleogene assets in the Gulf of America rather than accept the price on offer.

"We now see full year CapEx in the range of $13.5 billion to $14 billion, reflecting our decision to delay asset farm-downs to capture better value."
— Kate Thomson, Chief Financial Officer

First-half capital expenditure of $6,376M was 8.7% below the $6,984M of a year earlier, which means the raised full-year range requires a materially heavier second half.

Assessment: we take management at its word that this is discipline rather than slippage, because refusing a bid is the behaviour the new framework promises and because Kaskida and Tiber-Guadalupe are genuinely long-dated assets worth partnering carefully. The honest counterpoint is that capital discipline announced as a priority in the same presentation that raises capital expenditure guidance and lowers proceeds guidance is a harder sell, and the market took it that way.

12. Governance: A Chair Gone, Three Directors Gone, a Search Open

The board that bp reported with in August is not the board it reported with in February. Carol Howle stepped down as executive director and interim chief executive on 31 March. Meg O'Neill was appointed on 1 April. Melody Meyer, Karen Richardson and Simon Henry each stepped down as non-executive directors on 23 April. Albert Manifold ceased to serve as a non-executive director and chair on 26 May, and Ian Tyler was appointed interim chair. The permanent chair search remains open and management declined to comment on it.

Assessment: a chief executive four months into the job, an interim chair, three non-executive directors departed in a single day, and a chair search running concurrently with a segment reorganization, five divestment processes and a cost programme. Management's answer, that the board is in place and supportive, is what any chief executive would say and is not falsifiable. This is the reason the execution and leadership risk point escalates this quarter despite everything else going right.

13. Impairment Assumptions Move Down While Prices Move Up

bp revised its value-in-use impairment testing assumption for Brent to $80/bbl in real 2024 terms, from the $82.80/bbl it set in the first quarter, and raised its Henry Hub assumption to $3.34/mmBtu from $3.00/mmBtu. The Brent assumption continues to carry the same condition it carried in April.

"With reference to the Brent price, this estimate assumes that the ongoing supply disruptions resulting from geopolitical instability in the Middle East resolve before the year end 2026."
— bp p.l.c. Group results, second quarter and first half 2026

No material impairment or reversal arose in the quarter from the change to these assumptions. The roughly $800M of post-tax net impairments that did arise related primarily to transition businesses in gas and low carbon energy, which is a portfolio decision rather than a price decision.

Assessment: this is the disclosure that most cleanly contradicts the bullish reading of the quarter. bp earned its result at $103.85/bbl Brent and simultaneously lowered the Brent price it uses to test whether its assets are worth their carrying value, while restating that it expects the disruption to end this year. Management is telling investors, in the accounting notes, not to capitalize this environment. The market did the arithmetic on the day.

Guidance & Outlook

Full Year 2026

Third Quarter 2026

The third-quarter guide is the softest part of the print and explains a good deal of the share-price reaction. Throughput falls roughly 9.3% at the midpoint on the Gelsenkirchen divestment, customers is guided "significantly lower" on a lower midstream result and the lagged impact of higher base oil costs at Castrol, production drifts down at the midpoint on continued Middle East disruption and an estimated 40 mboe/d allowance for Gulf of America weather, and cash tax rises roughly $1B. The only line guided up is the hybrid repayment.

Implied second-half shape: to reach financial obligations of $39B to $41B from $53.6B, bp needs roughly $12.6B to $14.6B of reduction across two quarters, against second-half assumptions of $7B to $8B of divestment proceeds, a $2B to $3B working capital release and $7B to $7.5B of capital expenditure. The divestment component depends overwhelmingly on Castrol completing.

Street at: consensus entering the print was clustered around $5.0B of underlying RC profit for the second quarter and was too low by 14.6%. On second-half numbers the Street has not had time to reset to a $28/bbl refining assumption, and estimate revisions were running upward around the print.

Guidance style: bp pre-announces. The 14 July trading statement pre-flagged production, segment realization deltas, net debt, the impairment charge and the effective tax rate, and every one landed inside its band. Read the trading statement, not the consensus.

Analyst Q&A Highlights

bp splits its results call into a pre-recorded video of prepared remarks released with the print and a live call that is Q&A only. The live call ran 45 minutes with a one-question-per-person convention, and management took questions from the phone queue and from an online queue read by the moderator.

What Weight Class bp Is Actually Competing In

The opening question was an identity question rather than a numbers question, and it went straight at a phrase from the prepared remarks. The premise was that bp is the smallest of the traditional supermajor peer group and that several peers produce multiples of its volumes, so the request was for management to say plainly whether it aspires to that group at all. The answer declined the label outright, which is a meaningful departure from a decade of bp positioning.

Q: "So Meg, you've made some comments around getting fitter to grow and building that platform for growth. And then the other comment was around competing in the right weight class. So the question really is what weight class do you think bp is in? Because if I look at the traditional supermajor peer group on the key metrics, you're the smallest of the bunch and many of them produce multiples of what you produce. So, do you want to be in that weight class? And is that a fair assessment of where you want to be?"
— Biraj Borkhataria, RBC Capital Markets

A: "Look, I think it's prudent for us not to try to label ourselves as a supermajor. And in fact, that's one of the things that I'm trying to reinforce with the comments around weight class. You look at our numbers; we produce about 2.2 million barrels of oil equivalent a day. Our refining capacity is about 1.5 million barrels a day. And so, sizable on both fronts. But the reality is we need to make sure we're competing with players at our size and we need to make sure in each part of the business - upstream, downstream, trading - that we are competing to win. We want to be in the best basins, but we can't be in every basin."
— Meg O'Neill, Chief Executive Officer

Assessment: this is the most consequential sentence on the call. A chief executive publicly declining the supermajor label is signalling that portfolio decisions will not be defended on scale grounds, which is precisely what makes marketing the North Sea and Archaea coherent rather than defensive. It also lowers the bar management will be judged against on production growth, and investors should notice that both effects are intended.

Why Structural Cost Reductions Are Not Reaching the Bottom Line

The most substantive exchange of the call concerned the gap between a cost programme that reports cumulative delivery of $3.5B and a cost base that has barely moved. The question asked for a diagnosis rather than a restatement of the target, and it got one, along with the first hard number for where the absolute cost base is going.

Q: "I wonder, have you been able to diagnose why the improvements you have made have not managed to flow through to the bottom line? You made a comment about that in your prepared remarks. And when do you think is it reasonable for us to start seeing these operating cost changes and the changes you're making to the organisation actually flow through to the bottom line?"
— Joshua Stone, UBS

A: "The teams are working really hard across the company to drive our cost base down to get us competitive, as competitive as we can be. But frankly, it's not moving fast enough to be able to deliver that outcome all the way to the bottom line. And that's what our shareholders care about. ... the disclosure that we're giving you today is line of sight to the material reduction in our absolute cost base by the end of next year, getting to around $18 billion compared to $22 billion at 2025. ... Just to be clear, we expect to have delivered $5.8 billion of structural cost reductions compared to the original target of $4 to 5 billion. But as I say, what matters is what comes through to the bottom line in terms of earnings and cash flow."
— Kate Thomson, Chief Financial Officer

Assessment: an unusually candid answer that reframes the target from a gross savings number to an absolute cost base, which is the harder and more honest metric. The tension management has now created for itself is that beating a $4B to $5B savings target with $5.8B while the cost base falls by roughly $4B tells investors most of the savings are being consumed. The chief executive added that the benefits start reaching earnings in 2027, which means the cost story is a 2027 story, the distribution story is a 2027 story, and the reorganization benefits are a 2027 story.

Whether bp Needs Hybrid Capital at All

A pointed challenge argued that the scale of prospective free cash flow, including disposals and the working capital reversal, is large enough that bp could theoretically eliminate its hybrid stack entirely and stop worrying about the credit rating. The chief executive framed the capital structure question first, and the chief financial officer then gave the schedule.

Q: "...which is that the scale of the free cash flow, including disposals, including the working capital reversal that you could potentially generate, not just in 2026, the second half, but also in 2027, starts to put some fairly big questions over what you do with the capital structure, how low you take it. And dare I tempt Kate to maybe answer this, why do you need any hybrid bonds and worry about the credit rating? Because you could theoretically wipe them out."
— Doug Leggate, Wolfe Research

A: "As I think about your specific question on hybrids, we've stated today there's $1 billion of hybrids that will be naturally redeemed in the third quarter. We redeemed $2.9 billion in 2Q. And there's another $1.4 billion that we have already told you we're going to allow to move off the books when they redeem in 2Q next year. So the most cost-effective way to remove hybrids is to wait until they mature. Buying them back in advance of that can be a very cost-ineffective approach and not necessarily the most value accretive for shareholders."
— Kate Thomson, Chief Financial Officer

Assessment: the answer is technically correct and strategically evasive in equal measure. Waiting for call dates is the cheapest way to retire hybrids, and it also conveniently defers the harder question of what the target capital structure is. The chief executive's framing beforehand was more revealing: the concern is that "there's too much cash going to liability holders," which is a statement about the direction of travel and not about a destination. The schedule now visible is $2.9B redeemed, $1.0B in the third quarter and $1.4B in the second quarter of 2027, which against the $13.0B of hybrid bonds and securities standing at the quarter end leaves roughly $10.6B outstanding thereafter.

When Shareholders Get a Distribution Framework

A recurring line of questioning on the call observed that all five stated priorities are inward-facing and asked what shareholders get, framed either as a cash-flow payout ratio or as an indication on restarting buybacks. This is the third consecutive quarter the question has been asked and the first time an answer contained a number of any kind.

Q: "Thanks for laying out the five priorities. I noticed that they're very inward focused, getting the performance up to speed. What can you tell us already at this point? And apologies if you think that's unfair to sort of say to shareholders, look, this is the landing point in terms of returns to shareholders. Maybe you want to frame it as CFFO payout or some sort of indication how you're thinking around restarting the buybacks or giving shareholders more than the 4% minimum dividend per ordinary share increase."
— Christopher Kuplent, Bank of America

A: "Kate and I are doing a tremendous amount of work on the financial frame to make sure we've got laser like clarity on what a good frame for bp at this point in time looks like. Again, we need to make sure the balance sheet is positioned well. We need to understand the financial liabilities that we want to carry. ... But we've got work to do on the balance sheet, $40 billion of total liabilities is still too much. We are still not going to be able to offer that resilience through the cycle that we need to be able to. So we've got a bit of work to do. We know the market is very interested in hearing this. And as you know, as soon as we've got our views ready, we will be communicating with you."
— Meg O'Neill, Chief Executive Officer

Assessment: partial progress on a question that produced nothing in April. "$40 billion of total liabilities is still too much" is an implicit threshold, and set against a year-end guide of $39B to $41B it says the answer arrives no earlier than the fourth-quarter results and probably at a capital markets event in 2027. Investors should read the framing as sequencing rather than reluctance, but the practical effect on a holder is the same: another five quarters of waiting.

What Went Wrong Operationally This Quarter

A question submitted online went directly at the two reliability numbers rather than at the earnings, asking specifically what failed and what is being done about it. The response listed the individual events and then, more usefully, reframed the problem as a systems question rather than an incident question.

Q: "Upstream plant reliability fell to 92.4% due to some operational issues in the North Sea and Indonesia and refining availability dropped to 94.7%. Can you discuss specifically what went wrong operationally this quarter in both upstream and downstream? And what are you doing to address these issues?"
— Kim Fustier, HSBC (question submitted online)

A: "And look, it's been a disappointment. After four or five quarters of what I would call really quite strong reliability in both upstream and downstream, we have seen a drop off. North Sea, it was a couple of different issues. One at the Glen Lyon FPSO that took that facility offline for a couple of months and then some trips in ETAP. Indonesia, we had a turnaround that ended up being extended and had some operational issues when we were trying to restart. And in refining, we had a third-party event at the Whiting refinery that caused a bit of downtime in April. ... we need to step back and ask ourselves, do we have the right framework for our teams all around the world to deliver strong operational performance?"
— Meg O'Neill, Chief Executive Officer

Assessment: the honest answer to a question management could have deflected, and the systems framing is the right instinct. What is missing is any acknowledgement of the process safety event count, which more than doubled sequentially and is a different category of problem from an extended turnaround. A question about availability was answered on availability; nobody asked about the 18, and management did not volunteer it.

Why Capital Expenditure Rose Without a Volume Offset

The clearest challenge to the capital discipline narrative pointed out that guidance moved up with nothing visible bought for it. Both executives answered, and the two answers together are the fullest statement bp has made on its Paleogene strategy.

Q: "I wanted to ask about the increase in capital expenditures for this year. Seems like capex moved higher, but there wasn't any associated increase in production levels or downstream levels or nothing we could really discern. So wondering what you're getting for that higher capex, and if you think capex could trend higher over the next few years as you look to strengthen the earnings potential of the company."
— Jason Gabelman, TD Cowen

A: "When we had built the plan last year and when we initially put out guidance, we were assuming a farm down of some of our Paleogene assets. And we've deferred that because we want to make sure we're getting good value for bp shareholders on assets that are going to be part of our portfolio for the next 30 years. So we're being extremely disciplined in our divestment processes to get that fair value. That means we're picking up more of the capital onto our books. But net net, it is the right decision for our shareholders."
— Meg O'Neill, Chief Executive Officer

Assessment: a good answer to a fair question. Deferring a farm-down because the price is wrong is exactly the behaviour the new framework promises, and Kaskida and Tiber-Guadalupe are described elsewhere on the call as commercializing over 600 million barrels between them, which is worth partnering carefully. The chief financial officer added that she sees no reason for capital expenditure to trend up, which is the commitment to hold management to. The awkward fact remains that a print emphasising capital discipline raised capital expenditure and cut proceeds in the same breath.

Whether the Divestment Target Still Governs the Portfolio

A follow-up asked whether the extent of portfolio pruning implies going beyond the $20B divestment target set for 2027, and how bp is tracking against it. The answer effectively retired the target as a governing metric, which is a change investors should register.

Q: "I just wanted to follow up on one of the earlier comments about asset sales and the extent to which the portfolio has too many assets in it at the moment. I wondered whether that points to potential of kind of going beyond the $20 billion target that you have set by the end of 2027, not necessarily by the end of 2027, but into the future. And just on that, have you got any update on how you are tracking against that target at the moment, particularly considering the slight reduction in the divestment proceeds guidance for the full year this year?"
— Fergus Neve, Rothschild & Co Redburn

A: "So, we delivered $5.3 billion last year in terms of proceeds. We're guiding on $8 to 9 billion this year. What I would say to be really clear is we've never held the $20 billion as a key target of ours. It was put into the market to demonstrate one main lever that we could utilise to drive our deal, drive our deleveraging and increase our financial resilience. I'm far more focused on getting our balance sheet to where it needs to be than the number of divestments we make. ... We're not selling assets at any price to hit a divestment target. ... But by the end of this year, we should be at about $13 to 14 billion."
— Kate Thomson, Chief Financial Officer

Assessment: the substance is right and the framing is revisionist. A $20B target announced to the market and tracked by the market for eighteen months was not "never held as a key target," and saying so from the podium invites investors to discount the next target the same way. The disclosure that cumulative proceeds reach roughly $13B to $14B by year-end against a $20B 2027 goal is useful and new. The principle that assets will not be dumped to hit a number is the correct one.

The Impairment Track Record and Whether It Stops

The last substantive question picked up the phrase about writing off too much value from the results announcement and asked what actually changes, distinguishing between sanctioning durable new projects and testing the existing portfolio down to lower prices. The answer conceded the record without qualification.

Q: "There was a statement in the SEA about writing off too much value. I wonder if you could talk a little bit about that ... there's really two ways to address it, right? There's making sure that the new projects you're sanctioning are durable at a low price, but you've also got a portfolio review going on. And I suspect that ... you're testing assets down to lower prices ... your expectation of when you'll be able to sort of confidently say that the portfolio that's under you and the assets that are on the balance sheet are durable at a lower price. And we can stop that cycle of writing off."
— Stephen Richardson, Evercore ISI

A: "Look, if we look back over the history, we have had too many impairments. And the reality is those are shareholder dollars that were not used effectively, really wasted. And we need to start by being upfront and acknowledging that we have made some decisions in the past that did not deliver the outcomes we expected. ... So your comment around new projects must be durable. That is absolutely how we're thinking about our future investment decisions. Now that said, we still have a bit of work to do as we look at the portfolio."
— Meg O'Neill, Chief Executive Officer

Assessment: "really wasted" is not language a chief executive uses about her own company's history without intending it to be quoted, and it is the clearest signal that further write-downs are coming as the portfolio review works through. Read alongside the roughly $800M of post-tax impairments taken this quarter on transition businesses, investors should expect the reported-to-underlying gap to stay wide through 2026 rather than narrow.

What They're NOT Saying

1. A buyback restart condition, for the third consecutive quarter: the closest thing offered is that "$40 billion of total liabilities is still too much." No net debt level, no coverage ratio, no date, and no capital markets event scheduled at which the financial frame will be presented.
2. The realized refining margin: promised as second-quarter trading-statement color in April, delivered as neither a per-barrel figure nor a gap to the indicator margin. The largest profit line in the company remains unmodellable from the outside.
3. The oil trading contribution, still a single line: products, refining and trading is disclosed as $3,183M with the trading component described only as "slightly higher" sequentially. bp will tell you trading adds roughly 4 percentage points to return on capital employed over six years, but not what it earned this quarter.
4. The restructuring charge for the reorganization: deferred in April to a second-quarter cost deep dive. The deep dive arrived, was genuinely useful, and did not contain the number. A full upstream and downstream reorganization has now been announced twice with no price attached.
5. Any sensitivity around the $39 to $41 billion guide: the second-half assumptions of $80/bbl Brent and a $28/bbl refining indicator margin are disclosed. What financial obligations look like at bp's own $72.9/bbl and $11.2/bbl planning deck is not.
6. The 18 Tier 1 and Tier 2 process safety events: disclosed in the operating metrics table, referenced only as an "increase of events" in the prepared remarks, never quantified aloud, and never asked about. The event count more than doubled sequentially.
7. What the North Sea sale does to production, reserves or the 2027 targets: the intention is a full divestment. No volumes, no reserve quantum, no proceeds range, and no statement on whether the primary targets are rebased on completion.
8. The Castrol completion date, beyond "by the end of 2026": a single regulatory approval carrying roughly $6B, on which the entire year-end obligations guide depends, still has no narrowed timeline eight months after announcement.
9. Anything about the chair search: asked directly and answered with an appeal for patience. bp has had an interim chair since 26 May and lost three non-executive directors on 23 April.
10. Why US realizations are so much weaker than European ones: $76.03/bbl of liquids against a $93.11/bbl West Texas Intermediate marker and $1.72/mcf of gas against a $2.90/mmBtu Henry Hub index. Neither gap is a price lag at this point in the cycle and neither was addressed.
11. The Lightsource balance sheet position: asked directly, deferred to "when we have signed a transaction." The only figure offered was the roughly $3B of debt acquired in the fourth quarter of 2024.

Market Reaction

• Pre-print setup: the ADS closed at $44.26 on 3 August, up 27.4% year to date against the S&P 500's 11.0%, up 36.2% over the trailing twelve months and up 18.3% over the trailing thirty days. The 52-week closing range entering the print was $31.75 to $47.63, so the shares arrived 7.1% below their own 52-week closing high after a very strong month.
• Trading-statement run-up: the ADS closed up 1.4% at $41.40 on the 14 July trading-statement session on volume of 10.2 million, and ran from $39.20 on 10 July to $44.26 on 3 August, a gain of 12.9% into the print.
• Reaction session (4 August, a before-open report): the ADS opened at $43.22, a 2.3% gap down, traded a $42.19 to $43.31 range, and closed at $42.44, down 4.11% or $1.82.
• Volume: 13.3 million ADS against a 30-day average of 9.6 million, or 1.4 times normal. Unlike April, this print did bring incremental participation, and it was participation on the sell side.
• Peer reaction, same session: Shell closed down 1.36%, TotalEnergies down 2.31%, Chevron down 1.44% and Exxon Mobil down 0.71%, against the S&P 500 up 1.8%. bp fell roughly three times the average of the four peers and 5.9 percentage points more than the index.

This was a bp day, not a sector day, and it is the mirror image of April. Every major was lower on a session when the index rose 1.8%, which is a sector signal, but bp's incremental 2.7 points of decline against the peer average is its own. A company that beat consensus underlying profit by 14.6%, cut net debt by $3.1B, pulled its net debt target forward by a year and raised the dividend lost 4.1% of its market value on 1.4 times normal volume.

Three things explain it. The first is that there was no new information in the numbers: the 14 July trading statement had pre-announced the production range, the segment realization deltas, the net debt exit and the impairment size, and the shares had already appreciated 12.9% between that statement and the print. What was left to trade on was the third-quarter guide and the strategic framing, and the third-quarter guide is soft on throughput, customers and cash tax.

The second is composition of the ask. Capital expenditure guidance rose $0.5B, divestment proceeds guidance fell $1.0B, and the distribution question was answered with a commitment to communicate later. A shareholder who has watched buybacks stay suspended since February, and the share count drift up 0.3% sequentially, heard a company spending more, selling less this year and still not paying.

The third is that the reset itself is a risk as well as an opportunity. A chief executive who says the portfolio is too stretched, the cost base too complex and the impairment record indefensible is describing a company that requires work, and she is doing it four months in, with an interim chair, alongside five live sale processes. Some holders will pay up for that clarity. On 4 August, more of them took the price the rally had given them.

Street Perspective

Debate: Is the De-Rating an Opportunity or a Peak-Earnings Trap?

Bull view: the bull case being made on the Street is that the shares now trade on roughly 8.6 times trailing underlying earnings per ADS with a 4.9% dividend yield, against a business whose net debt target has just been pulled forward a full year. Numbers are going up (estimate revisions around the print ran upward) while the multiple went down, which is the classic setup for a re-rating once the distribution framework lands.

Bear view: the bear camp contends the trailing multiple is a mirage produced by a $103.85/bbl Brent and a $29.6/bbl refining margin, and that the correct denominator is the mid-cycle one. bp's own price-adjusted disclosure shows first-half free cash flow of $14.4B becoming roughly $10B on the planning deck. Applying that same haircut to the trailing earnings per ADS produces a multiple around 11 times, which is where the shares traded before the run.

Our take: the bear framing is arithmetically correct and should set the multiple. A cyclical trading at 8.6 times peak earnings is not cheap; it is a cyclical at peak earnings. What the bull case gets right is that the balance sheet is a different question from the earnings, and a company retiring $6.9B of obligations a quarter is genuinely worth more each quarter regardless of what the multiple says. We would not buy the trailing multiple. We would buy the obligation reduction, once there is a distribution policy attached to it.

Debate: Does Castrol Close, and Does the Year-End Guide Hold?

Bull view: the sequencing has never been more visible. Gelsenkirchen completed on 31 July, the Austrian retail sale is agreed, Bay du Nord is agreed, Kirkuk partners are agreed, and Castrol carries roughly $6B against a $12.6B to $14.6B second-half obligation reduction. Add a $2B to $3B working capital release that management has already sized and organic cash generation at any plausible price, and the $39B to $41B range is comfortable rather than heroic.

Bear view: the skeptics point out that the same guide already absorbed a $1.0B cut to full-year proceeds and a $0.5B increase to capital expenditure in the space of one quarter, and that every remaining lever except working capital requires a counterparty or a regulator. Castrol has been pending since 24 December 2025 with no narrowing of "by the end of 2026," and a slip of one quarter moves roughly $6B out of the range and the net debt milestone back into 2027.

Our take: the bull has the better of it on the base case and the bear has the better of it on the variance. bp will very likely land inside or near the range, but the distribution of outcomes is dominated by a single regulatory approval rather than by anything management controls. That is precisely why we will not pay for the year-end target before it is banked, and it is the same conclusion we reached in April with a materially better balance sheet behind it.

Debate: Is the Reset Genuine, or a Relabelling of the Existing Plan?

Bull view: a growing consensus view is that this is the real thing. The chief executive declined the supermajor label, published a portfolio chart that scores bp's own assets on cash and returns, called the impairment record a waste of shareholder money, launched processes on the North Sea and Archaea within her first four months, and delivered a cost disclosure that makes bp's record look worse rather than better. None of that is what a caretaker does.

Bear view: the bear case is that the underlying targets did not change. The net debt range, the structural cost target, the return on capital target and the free cash flow growth target are all the same numbers set at the February 2025 capital markets update. A reorganization was announced by the previous management and the reporting segments do not even change until 1 January 2027. Five priorities and a new vocabulary are not a strategy reset.

Our take: both are describing the same thing from different distances. The targets are unchanged because they were credible targets; changing them in quarter one would have destroyed rather than created credibility. What has changed is the standard of candour and the willingness to act on the portfolio, and those are the leading indicators. The lagging indicator is whether the cost base reaches $18B and whether the impairment cycle stops, and neither is testable before 2027.

Debate: Is the Upstream Shrinking Faster Than It Is Being Rebuilt?

Bull view: the bull argument is that reported production understates the business. Underlying production rose 1.8% in oil production and operations this quarter and 3.8% across the half, eight of ten major projects due between 2025 and 2027 have started up including Atlantis in July, two further final investment decisions were taken in the quarter, and Bumerangue holds 8 billion barrels of liquids in place with an appraisal campaign starting around year-end.

Bear view: the bear camp counts the other direction. Culzean divested, Pan American Energy cut from 50% to 40%, Bay du Nord exited, the entire UK North Sea business marketed for a full divestment, and Archaea marketed. Reported production fell 4.3% year over year, full-year guidance implies 2,180 to 2,270 mboe/d against 2025's 2,312, and management said plainly that 2026 is "a quiet year" with nothing major expected to reach a final investment decision.

Our take: the bear is right about the next two years and the bull may be right about the decade. bp is deliberately shrinking a stretched portfolio into a smaller, higher-return one, and the transition period has lower volumes by construction. The risk we are watching is not the shrinkage; it is that reserve replacement was 90% headline and roughly 76% excluding price effects in 2025, and selling the North Sea does not improve that ratio. This is why the upstream re-loading pillar stays at risk rather than on track.

Model Framework & Valuation

We initiated in April with a driver framework rather than a point estimate. The table below carries it forward, marking what this quarter changed.

Where the Shares Sit

Valuation view: the shares are cheaper than they were in April on every trailing measure and no cheaper on any normalized one. That is the whole argument. At $42.44 an investor pays 8.6 times a trailing figure that contains two quarters earned at Brent prices bp's own impairment note says will not persist, and roughly 11 times the same figure adjusted using bp's own price-adjustment methodology. The 4.90% dividend yield is real, covered many times over this year, and would still be covered at the planning deck. What an investor does not get is any claim on the incremental cash, which for at least five more quarters goes to liability holders by explicit management design. The equity story starts when the balance-sheet story finishes, and management has now told us roughly when that is.

Thesis Scorecard Post-Earnings

The pillars below are the ones established in April, graded against what this quarter showed. They are the same seven, in the same order, and they will be the same seven next quarter.

Overall: the thesis strengthened. Five of the seven pillars moved in our favour, including both of the two that matter most to the medium-term equity case. The balance-sheet inflection is no longer a plan, and the realization disconnect is no longer the acute problem it appeared in April. What did not improve is the price the earnings were made at, the absence of a distribution framework, and the operational and governance noise around a company doing five things at once.

Action: hold. Conviction rises from 5 to 6. Neither upgrade trigger set in April was met: net debt is $22.3B rather than inside $20B, no buyback restart condition was disclosed, and one quarter of realization catch-up is not the two consecutive quarters we asked for. Both triggers are now plausibly reachable inside two quarters, which is the first time that has been true. We would rather pay a higher price for a company that has banked Castrol and named a distribution policy than pay 8.6 times peak earnings for the possibility of both.