Everything Went Right and Price Still Did All the Work: Maintaining Hold on BHP

Results vs. Consensus

BHP reports semi-annually, so the FY26 result covers the twelve months to 30 June 2026 and the Street's forecasting effort concentrates on full-year underlying attributable profit and the dividend rather than on quarterly revenue and earnings per share. Both of those lines beat. Neither was the informational content of the day.

FY26 Scorecard

The volume lines were pre-released. BHP published its Q4 FY26 operational review on 16 July 2026, five weeks before the result, and that document already carried FY26 production, FY27 production guidance, a net-debt indication of roughly US$9bn, an expected Jansen impairment of roughly US$2,300M, negative EBITDA of roughly US$150M each at Jansen and Western Australia Nickel, and the statement that every asset was expected to land within FY26 unit-cost guidance. What 18 August added was margins, the realised-exchange-rate cost outturn, the final dividend, FY27 unit-cost ranges, the FY27 to FY31 capital path, and the rebased medium-term cost ceilings.

Year-Over-Year Comparison (FY26 vs. FY25)

Free cash flow is disclosed by BHP only to one decimal place in US$bn; the +83% change is the company's stated figure and does not recompute exactly from the rounded levels shown. The gap between +30% underlying and +9% statutory is the exceptional line. FY26 carries a US$3.4bn exceptional loss made up of the US$2.3bn Jansen impairment and US$1.1bn of Samarco dam-failure impacts, against US$1.1bn in FY25 (US$0.9bn Samarco and US$0.2bn on the Western Australia Nickel suspension). Non-controlling interests took US$3,193M, up 50.3%, which is Escondida's minorities participating in the copper year.

Half-on-Half Comparison (H2 FY26 derived vs. H1 FY26)

BHP does not publish a standalone second half. The figures below are derived by subtracting the reported December half from the full year, which is the only way to see how the year actually shaped.

The derived second half cross-checks cleanly against the company's own disclosure. BHP states that the US$5.0bn final dividend represents a 72% payout ratio, and US$5,000M over the derived H2 underlying attributable profit of US$7,002M is 71.4%. The second half was the better half on every line that matters, and it is the half that repaired the balance sheet.

Revenue assessment. Group revenue of US$58,760M is a record, and the composition is the interesting part. BHP's own EBITDA waterfall separates the effect: changes in sales prices contributed US$7,710M, price-linked costs took back US$399M, and the net US$7,311M price impact exceeds the entire US$6,969M increase in Underlying EBITDA. Volumes were a US$1,167M drag, controllable cash costs a US$1,191M benefit, and other costs (currency, inflation, fuel) a US$1,586M drag. Strip price out of this year and the earnings base went backwards. That is not a criticism of management, who cannot set the copper price; it is a statement about what an investor is buying at 17 times underlying earnings.

Margin assessment. The margin story is genuinely good and deserves to be separated from the price story. Escondida cut unit costs 10% to US$1.07/lb while its concentrator feed grade fell to 0.90% from 1.02%, on record material mined and record concentrator throughput. Copper South Australia cut unit costs 73% to US$0.32/lb, helped by US$2.3bn of by-product revenue and a 20-year production record at Olympic Dam. Group by-product revenue across the copper assets reached US$4.5bn, up 45%, on 571 koz of gold, 17.9 Moz of silver and 3.6 kt of uranium. On BHP's stated basis, which excludes third-party product sales, the group Underlying EBITDA margin is close to 59%; on total reported revenue it is 56.1%. Copper's margin of 70% and iron ore's 61% are the numbers that matter, and both are excellent.

Earnings and cash assessment. Underlying ROCE of 26.1% is 550bp better and is the single cleanest measure of the year. Escondida's asset ROCE was 53% and WAIO's 41%. Free cash flow of US$9.8bn against a US$226bn market capitalisation is a 4.3% free cash flow yield, which is respectable rather than cheap for a business whose largest earnings driver is guided to produce 12% less next year. The statutory line is where the year's mistake sits: US$2.3bn written off against a project that is 84% built at Stage 1 and 16% built at Stage 2.

Segment Performance

Group Underlying EBITDA by Segment

Segment shares are stated on BHP's basis, which apportions across the three producing segments and excludes the negative Group and unallocated line; the four Underlying EBITDA figures sum to the group total of US$32,947M. Group and unallocated carries the Potash loss of US$326M, the Western Australia Nickel loss of US$255M and functions.

Asset-Level Detail

The NSWEC line above is the segment-reporting figure, which consolidates BHP's share of the equity-accounted Newcastle Coal Infrastructure Group; the statutory NSWEC result is US$0.23bn against US$0.16bn in FY25. Antamina is equity accounted, so its revenue and EBITDA appear in the asset table on BHP's share basis and are adjusted out of the statutory Copper segment result (US$18,932M of total Copper Underlying EBITDA less a US$745M equity-accounting adjustment gives the reported US$18,187M).

Copper

Copper revenue rose US$6,501M to US$29,031M and Underlying EBITDA rose US$5,861M to US$18,187M, a record, at a 70% margin and a 26% segment ROCE against 17% a year ago. It is the first year copper has been the majority of BHP's earnings on a full-year basis. Production of 1,952.8 kt was down 3.2% and landed inside the 1,900 to 2,000 kt guide, held up by record material mined and record concentrator throughput at Escondida against a feed grade of 0.90% (FY25: 1.02%), a 27% jump at Antamina to a financial-year record 151.5 kt, and 320.7 kt at Copper South Australia. Pampa Norte fell 21% to 212.6 kt as Spence worked deeper into hypogene ore.

The composition of that EBITDA increase is the point. Price impacts net of price-linked costs contributed US$6,693M, which is 114% of the segment's total US$5,861M increase. Volumes subtracted US$1,415M, controllable cash costs added US$749M, inflation and adverse currency took US$550M, and other movements added US$384M. Copper's realised price of US$5.74/lb was 35% above FY25's US$4.25/lb.

"Our copper business generated a record $18 billion of EBITDA, 54% of the group total, at a margin of 70%. Our operational performance, combined with a $4.5 billion contribution from by-products, saw unit costs improve by 10% at Escondida and over 70% at Copper SA, an impressive result."
— Vandita Pant, Chief Financial Officer

Assessment: The operating performance inside copper is the best thing about this company and it is not in dispute. Record throughput against a falling grade at the world's largest copper mine, a 20-year production record at Olympic Dam, and US$4.5bn of by-product revenue that turned Copper South Australia into a US$0.32/lb producer are all genuine. What the segment cannot do is grow earnings without the price. It did not this year, and FY27 guidance of 1,650 to 1,800 kt says it will not next year either.

Iron Ore

Iron ore revenue rose US$964M to US$23,883M on a realised price of US$84.56/wmt, up 3%, and record group production of 264.7 Mt. Underlying EBITDA barely moved, up US$133M to US$14,529M, and segment ROCE fell to 55% from 64%. WAIO produced a record 256.9 Mt (291.2 Mt on a 100% basis), with record material mined up 6%, South Flank above annual nameplate capacity, and record shipments after the Car Dumper 3 rebuild completed in Q1 FY26. Samarco rose 25% to 7.8 Mt.

The segment's waterfall is the cleanest illustration of the group's problem: a US$452M net price benefit and a US$168M controllable cost benefit against a US$536M hit from inflation and adverse currency produced a US$133M increase on a US$14.4bn base. WAIO's unit cost rose 6% to US$19.66/t at a realised AUD/USD of 0.68 against 0.65 a year earlier, though at guidance exchange rates it printed US$18.87/t, inside the US$18.25 to US$19.75/t range.

"It is worth repeating, WAIO produces around $10 per ton more free cash flow than its closest peer, and our goal is to increase this margin further."
— Brandon Craig, Chief Executive Officer

Assessment: Iron ore is doing exactly what a mature annuity is supposed to do, which is generate roughly US$14.7bn of asset EBITDA on US$3.0bn of capital expenditure and hand the difference to the copper build. What it is no longer doing is getting cheaper in US dollars. A 6% unit-cost increase and a medium-term ceiling rebased to below US$19.00/t from below US$17.50/t are mostly the Australian dollar, not mismanagement, but the earnings consequence is the same either way.

Coal

Coal revenue rose US$544M to US$5,590M and Underlying EBITDA rose US$259M to US$832M, a 45% increase from a low base and a 15% margin. Segment ROCE was 0%, up from -1%. BMA produced 18.6 Mt, up 3%, with the highest stripping volumes in five years, and lifted raw coal inventory roughly 30%. Realised steelmaking coal was US$210.21/t, up 8%, after prices rebounded 28% in the second half. NSWEC produced 16.4 Mt, up 9% and above the top of its 14 to 16 Mt guidance range, on increased bypass coal and lower strip ratios as the asset runs toward its FY30 closure.

Assessment: BMA generated US$3,876M of revenue and US$161M of Underlying EBIT. After depreciation, this business earns approximately nothing on US$6.4bn of net operating assets, and the medium-term plan asks for 43 to 45 Mtpa on a 100% basis at below US$120/t to change that. Management explicitly ruled out selling it. That is a decision to spend management attention and capital on the lowest-returning segment in the portfolio at a moment when the copper pipeline is competing for both.

Potash and Other Assets

Potash recorded a US$326M Underlying EBITDA loss against US$284M in FY25, on US$1,814M of capital expenditure, and absorbed the US$2.3bn impairment. Jansen Stage 1 is 84% complete with first production still targeted for mid-CY27 and total estimated expenditure of US$8.4bn. Stage 2 is 16% complete, targeted for late-FY31, and its estimate moved to US$6.9bn from US$4.9bn. Potash spot prices rose 23% in FY26 to US$342/t Vancouver FOB.

Western Australia Nickel narrowed its loss to US$255M from US$589M in temporary suspension. BHP will review the suspension decision by February 2027 and is assessing a potential divestment alongside restart and closure.

Assessment: Jansen remains a good asset inside a poorly estimated project. A world-class, low-cost, 60-year potash mine delivering roughly US$1bn of EBITDA per stage at above 60% margins is worth building. The problem is that BHP has now revised the capital estimate on this asset three times in thirteen months, and the cumulative revision is large enough that the accountants required a US$2.3bn write-down of the investment to date.

Unit Costs: Guidance Basis vs. Realised Basis

BHP sets unit-cost guidance at a fixed exchange rate and reports actuals at the realised rate, so the two columns answer different questions. FY26 guidance was set at AUD/USD 0.65 and USD/CLP 940. FY26 realised was AUD/USD 0.68 and USD/CLP 920. FY27 and medium-term guidance are set at AUD/USD 0.70 and USD/CLP 890.

Read the middle column, not the fourth. At guidance exchange rates every asset met its range, and Copper South Australia beat the bottom of its range outright. Read the realised column and BMA printed above the top of its FY26 range and WAIO landed in the upper half. Both statements are true. Which one matters depends on whether an investor thinks the Australian dollar is BHP's problem or the market's.

Key Topics & Management Commentary

Overall Management Tone: Confident and forward-leaning, with a noticeable shift in register from the February presentation. Where the prior half was framed as the payoff from a decade of executed strategy, this one is framed as a starting point, with an explicit three-priority agenda and a stated intention to add pace. Management addressed the fatality first and at length, without deflection. The one place the presentation was less convincing was on project capital: the Jansen reset and the impairment were acknowledged as a problem to fix rather than explained, and the medium-term cost and capital rebases were presented as inputs changing rather than as guidance moving.

1. A New Chief Executive, and What Changed in the Framing

Brandon Craig succeeded Mike Henry on 1 July 2026, the first day of FY27, which makes FY26 his predecessor's last year and his first result. He opened by crediting the inheritance rather than distancing himself from it, then set out three priorities that are more explicit than anything in the February presentation: accelerate performance, deliver programmatic growth, strengthen foundations.

"Let me begin by saying that BHP is in great shape. Mike has left us a strong platform with a Tier One portfolio, an operating system that continues to improve, and a balance sheet that allows us to invest through the cycle. I intend to build on this strong foundation by adding even greater pace, improving our operational practices, and focusing on our customers and markets."
— Brandon Craig, Chief Executive Officer

He also put a number on where he thinks the market is wrong, which his predecessor generally did not do:

"While investors understand the quality of BHP's portfolio, we believe there is more value to be recognized in the growth we have ahead of us. For example, many in the investment community model growth of around 2% per year through to 2035, well below our estimates of 3%-4%."
— Brandon Craig, Chief Executive Officer

Assessment: A new chief executive naming the specific gap between his growth number and the Street's is a useful commitment device, because it is falsifiable. It also tells us where the next twelve months of investor communication will be aimed. The risk in the framing is that the gap between 2% and 3% to 4% compound growth is worth far less to a shareholder than the difference between building projects on budget and building them 41% over, and this result contained one of each.

2. Copper Past Half the Company, at a Record Margin

Copper contributed 54% of producing-segment Underlying EBITDA in FY26, up from 45% in FY25, on a record US$18.2bn at a 70% margin, and generated US$6.9bn of free cash flow. On a full-year basis this is the first time copper has been the majority. BHP attributes it to portfolio actions taken over a decade rather than to the price.

"We are the world's largest copper producer, and for the second consecutive year, we produced around 2 million tons of copper, which contributed more than half of our annual EBITDA for the first time."
— Brandon Craig, Chief Executive Officer

Assessment: The mix shift is real and it is the reason to own BHP over its diversified peers. But 54% from 45% in a single year is mostly the copper price moving 35% while iron ore moved 3%. The volume path is what makes the mix shift durable, and the volume path has an air pocket in FY27. Judge this pillar on the FY29 to FY35 copper tonnage, not on this year's earnings share.

3. Price Supplied More Than All of the Earnings Growth

BHP's own principal-factors waterfall is the most useful disclosure in the result, and it says the following about the US$6,969M increase in Underlying EBITDA.

At the December half the same ratio was 96%. It has deteriorated. Within iron ore, price contributed US$452M against a total segment increase of US$133M, meaning the segment's operating performance was a net drag of US$319M once currency, inflation and fuel are counted. The Australian dollar alone cost the group US$798M and inflation another US$675M, at Australian inflation of 4.0% against 2.4% in FY25.

"The earnings waterfall reflects a familiar theme. While higher prices were a clear tailwind, up 35% for copper and 3% for iron ore, our operational performance and cost discipline enabled us to fully capture that benefit."
— Vandita Pant, Chief Financial Officer

Assessment: "Fully capture that benefit" is the right description of what BHP did and it is not the same claim as generating growth. A reliable low-cost operator converts a price move into margin better than a peer does, and that is worth a premium multiple. It is not worth pretending that a 27% earnings increase in a year when volumes fell 3% and unit costs rose in US dollars at three of five major assets is anything other than a price year.

4. The Provisionally Priced Revenue Nobody Discussed

Inside the revenue line sits a number that appears in the financial statements and nowhere in the presentation. Other revenue predominantly relating to provisionally priced sales was US$1,265M in FY26, against US$24M in FY25 and US$283M in FY24. That is the mark-to-market adjustment between the forward price used when revenue is first recognised and the final index price, on sales where the quotation period had not closed.

The exposure carried into FY27 is disclosed in the financial-instruments note: 423 kt of copper was provisionally priced at 30 June 2026 (FY25: 419 kt), with final prices set during the first half of FY27. BHP states that a 10% change in the realised copper price on those volumes moves profit after taxation by US$371M, against US$268M a year earlier. Trade receivables rose to US$3,807M from US$3,081M.

Assessment: US$1,241M of the year-on-year increase in revenue, roughly 17% of it, is a price remeasurement rather than a settled sale, and it drops straight through to Underlying EBITDA. This is not an accounting concern; the treatment is standard and the disclosure is complete. It is a quality-of-earnings observation: the same mechanism that added US$1.2bn on the way up subtracts on the way down, and US$371M of post-tax profit is still live on 423 kt of unpriced copper as the FY27 first half runs. Management raised none of this on the call.

5. FY27 Is a Designed Volume Trough

Group copper guidance for FY27 is 1,650 to 1,800 kt against 1,952.8 kt delivered, an 11.7% decline at the midpoint. Escondida carries most of it, guided to 1,000 to 1,100 kt from 1,261.2 kt on a concentrator feed grade falling to roughly 0.70% from 0.90%. Copper South Australia is guided to 290 to 320 kt from 320.7 kt, reflecting a planned anode-inventory build ahead of the six-yearly smelter campaign maintenance in H1 FY28 and the resumption of the Carrapateena conveyor after an unplanned belt failure in July 2026. Antamina steps down to 120 to 140 kt on planned lower feed grades.

The arithmetic is straightforward. A 228 kt decline at the guide midpoint is roughly 502 Mlb, worth about US$2.9bn of revenue at FY26's realised US$5.74/lb. Because BHP guides unit costs per pound higher rather than lower (Escondida to US$1.20 to US$1.50/lb from US$1.07), absolute cash costs are broadly held while the denominator shrinks, so most of that revenue decline drops through to earnings. Against BHP's own disclosed sensitivity of US$39M of Underlying EBITDA per US cent per pound on the copper price, offsetting it takes roughly 75 US cents per pound on the realised price, taking FY26's US$5.74/lb to about US$6.50/lb.

The counterweight is that copper spot has already moved. BHP notes spot prices averaged 26% higher in FY26, "with the second half of FY2026 experiencing increases of nearly 40 per cent as copper moved to >US$13,000/t (US$5.90/lb)."

Assessment: This is the central FY27 question and it is genuinely two-sided. The volume decline is planned, disclosed, grade-driven and temporary; medium-term Escondida guidance holds at 900 to 1,000 ktpa for FY28 to FY31 and the growth pipeline sits behind it. But a shareholder buying today owns a fiscal year in which the operating base contracts and the earnings line depends entirely on whether copper holds above roughly US$6.50/lb. That is a price bet dressed as a growth story, and it is priced at 17 times underlying earnings.

6. Jansen Stage 2 at US$6.9bn, and a US$2.3bn Impairment

In June 2026 BHP completed a detailed review of Stage 2 and confirmed a total investment estimate of US$6.9bn including contingencies, up US$2.0bn from the US$4.9bn approved in October 2023. First production moved to late-FY31. The consequence appeared in this result: a non-cash impairment charge of US$2.3bn, before and after tax, taken against the investment to date across the whole Jansen project.

"After exceptional items, including a $2.3 billion non-cash impairment relating to Jansen, our total attributable profit increased by 9% to almost $10 billion."
— Vandita Pant, Chief Financial Officer

The company's own framing of the cause is unusually direct: "Given the higher forecast capital intensity for the Jansen project (including Stages 1 and 2 and potential future expansions), we recognised an impairment charge of US$2.3 bn (before and after tax) in FY26 in relation to our investment to date in the Jansen project." Note the scope of that sentence. The write-down is not against Stage 2 alone; it reflects capital intensity across Stage 1, Stage 2 and potential future expansions.

"Jansen has highlighted that we need to further improve our major project performance, and we are very focused on this. Embedding BOS and technology into projects is at the center of our approach, and I am excited to see how much further we can improve our project execution, just as we have seen across operational performance."
— Brandon Craig, Chief Executive Officer

Assessment: This is the third capital revision on Jansen in thirteen months. In July 2025 Stage 1 carried a preliminary US$7.0bn to US$7.4bn range; in January 2026 it was reset to US$8.4bn definitive; in June 2026 Stage 2 went from US$4.9bn to US$6.9bn. The combined project is now estimated at US$15.3bn against the US$11.9bn to US$12.3bn implied by the estimates standing in mid-2025. Craig's response is the right one in substance, and his stated track record of roughly 80% of projects delivered on time and on budget over the past decade is a real defence. But the bear point we raised in February, that BHP's project capital estimation is unreliable, has now been confirmed by BHP's own auditors in the form of a write-down. Investors should apply a haircut to the Escondida New Concentrator and Copper South Australia expansion capital estimates on the strength of it.

7. The Dividend: 99 US Cents, and Back Inside Free Cash Flow

The board determined a fully franked final dividend of 99 US cents per share, US$5.0bn, a 72% payout of the derived second-half underlying attributable profit. Combined with the 73 US cent interim, the full-year distribution is 172 US cents, US$8.7bn, a 66% payout and the highest in four years. Payment is 23 September 2026; the NYSE ex-dividend date is 4 September 2026, and each ADS carries two ordinary shares.

"This strong operating performance, along with proceeds from the Antamina silver streaming transaction, has enabled the board to determine a dividend for the June half of $0.99 per share. This takes our full-year dividend to $8.7 billion."
— Brandon Craig, Chief Executive Officer

The comparison that matters is against free cash flow rather than against the payout policy. The interim dividend equalled 127% of the December half's US$2,906M of free cash flow. The final dividend is roughly 72% of the derived second half's free cash flow, and the full-year distribution is 89% of the year's US$9.8bn. Net debt fell US$4,230M over the twelve months while that dividend was paid.

Assessment: This resolves the second of the two bear points we raised in February. The payout is back inside the cash the business generated, on a much larger cash number, with the balance sheet strengthening at the same time. The qualifier is in Craig's own sentence: the dividend was enabled by operating performance "along with proceeds from the Antamina silver streaming transaction". A US$4.3bn one-time receipt sits behind a US$8.7bn distribution. That is a legitimate use of monetisation proceeds, and it is also a reason not to extrapolate a 66% payout into a year with lower volumes and higher capital spending.

8. A Balance Sheet Below Its Own Target Range

Net debt of US$8,694M is below the unchanged US$10bn to US$20bn target range, at 0.3 times Underlying EBITDA and a 13.4% gearing ratio against 19.8% a year earlier and 20.9% at the December half. The bridge is US$21,778M of operating cash flow plus US$4.3bn of Antamina streaming proceeds and roughly US$0.8bn of asset sales, against US$10,257M of capital and exploration expenditure, US$9.1bn of dividend payments (US$6.8bn to BHP shareholders and US$2.4bn to non-controlling interests) and US$2.0bn of Samarco settlement obligations. Gross debt rose US$2.6bn to US$27.1bn while cash rose to US$18.5bn from US$11.9bn. Credit ratings were unchanged at A1 stable from Moody's and A stable from Fitch.

Assessment: Being below the bottom of your own leverage target with a US$11bn annual capital programme ahead is the definition of optionality, and it is worth something. It also creates a question management did not answer: whether the intention is to run the balance sheet back up into the range to fund the copper build, to hold it here, or to distribute the difference. The FY27 capital path implies the first, but nobody said so.

9. Capital Expenditure Rebased to Roughly US$11bn a Year Through FY31

FY26 capital and exploration expenditure was US$10,257M, up 5%. FY27 is guided to roughly US$11.0bn, unchanged from February. What changed is the out-years: FY28 is now also roughly US$11bn, and roughly US$11bn on average each year between FY29 and FY31, against roughly US$10bn per annum guided for the FY28 to FY30 window in February. BHP attributes the increase to Jansen Stage 2 and to higher foreign exchange rates, partially offset by capital efficiency.

"We now expect CapEx of around $11 billion per year on average over the medium term. This is in nominal terms at constant FX rates. This reflects updated estimates for Jansen Stage 2 and changes in FX, partly offset by improved capital productivity across the rest of the portfolio. More than half of our growth spend will go to copper, and approximately two-thirds if we include investment in our non-operated joint ventures."
— Vandita Pant, Chief Financial Officer

Components in the table above are disclosed to US$0.1bn and do not always sum to the stated total, which is as reported.

Assessment: The mix is right: growth capital rises from US$2.8bn to US$3.8bn while baseline sustaining falls, and more than half of the growth spend goes to copper. The level is what changed, by roughly US$1bn a year for three years, or roughly US$3bn cumulative, against February's guidance. The company's answer to whether it can afford this is that the copper programme is self-funding at consensus prices, which is a claim worth testing at FY27's guided volumes rather than at FY26's.

10. WAIO: Record Volumes, a Changed Cost Methodology, and a Higher Target

WAIO delivered its seventh consecutive year as the world's lowest-cost major iron ore producer and a record 291.2 Mt on a 100% basis. Reported unit costs rose 6% to US$19.66/t on a realised AUD/USD of 0.68 against 0.65, with the Singapore 10ppm Gasoil benchmark up roughly 30% across FY26 and up 60% in the second half against the first, on roughly 630 ML of diesel (BHP share).

Underneath the headline, the C1 measure was restated. BHP now includes inventory movements in C1, "aligning the methodology with competitor reporting." On the new basis FY26 C1 is US$16.40/t against a restated FY25 of US$16.16/t, an increase of 1.5%. On the previous methodology, the one in force when the December half reported C1 of US$17.66/t, FY26 C1 would be US$18.72/t against US$17.29/t, an increase of 8.3%. BHP discloses both, in a footnote.

"WAIO achieved record production and shipments. It was another year in which we delivered a real cost decline, with C1 unit costs up just 1%."
— Vandita Pant, Chief Financial Officer

The medium-term unit-cost ceiling also moved, to below US$19.00/t from below US$17.50/t, alongside a change in the guidance exchange rate from AUD/USD 0.65 to 0.70. Sustained production above 305 Mtpa (100% basis) is still targeted from Q4 FY28, with BHP now specifying that both the volume and the sub-US$19.00/t cost will be achieved in FY29. Ministers North was approved for execution in June 2026 at roughly US$0.9bn (100% basis) for roughly 20 Mtpa, and the Western Ridge Crusher reaches first production in H1 FY27.

Assessment: The "up just 1%" figure is accurate on the basis BHP now uses and misleading against the basis it used six months ago. The honest summary is that WAIO's costs rose in US dollars, most of the rise is the Australian dollar and diesel, the asset remains comfortably the lowest-cost major producer, and the medium-term ceiling moved up 8.6% against a currency assumption that moved 7.7%. Most of the target increase is the currency rebase. Changing the definition of your headline cost metric in the same result in which that metric rises is a choice that invites the question, and no analyst was present to ask it.

11. Copper South Australia Becomes the Second Engine

Copper South Australia produced 320.7 kt, up 2%, with a 20-year copper production record at Olympic Dam, record material mined and milled at Carrapateena, and higher feed grades at Prominent Hill. Unit costs fell 73% to US$0.32/lb on US$2.3bn of by-product revenue, and Underlying EBITDA rose 65% to US$3,203M with roughly US$1bn of free cash flow after capital project investment.

"At Copper South Australia, we delivered the highest production at Olympic Dam in 20 years, and our unit cost declined to just $0.32 per pound, benefiting from $2.3 billion of byproduct revenue. As a result, we set new records for EBITDA and free cash flow."
— Brandon Craig, Chief Executive Officer

The development path is now specific. Phase 1 targets roughly 500 ktpa of copper (roughly 770 ktpa copper equivalent) at capital intensities of US$18,000 to US$23,000 per tonne of copper equivalent, with a second phase to up to 650 ktpa (roughly 1 Mtpa copper equivalent). Prominent Hill's expansion completes in H2 FY27 and extends mine life to the mid-2040s; Carrapateena's block cave ramp-up starts FY30; the Olympic Dam Smelter and Refinery Expansion targets a CY27 investment decision with a design and supply contract already awarded; the mine and concentrator expansion targets CY29; Oak Dam targets CY27. A site tour is promised later this calendar year.

Assessment: The 73% unit-cost decline is the single most eye-catching number in the result and it needs a caveat: it is by-product credits at record gold and uranium prices, and FY27 guidance of US$0.30 to US$0.80/lb is set on gold at US$4,300/oz and uranium at US$80/lb. Strip the by-products and this is a mid-cost asset. That said, the polymetallic economics are real, the production record is operational, and Copper South Australia is now the growth option that does not require a new country, a new permit regime or a joint-venture partner. It is the highest-quality item in the pipeline.

12. Escondida New Concentrator: Bigger, More Expensive, Better Returns

BHP approved roughly US$0.5bn (BHP share) of pre-commitment funding in August 2026 for the Escondida New Concentrator, ahead of a final investment decision in CY27 to CY28 and potential first production in CY31 to CY32. The environmental permit application was submitted in March 2026, meeting the commitment made in February. The scope grew: a 50 Mtpa concentrator against 45 Mtpa previously, delivering 230 to 270 ktpa of copper against 220 to 260 ktpa. So did the estimate.

"As we move towards that point, the project continues to improve. The optimized scope is expected to deliver higher throughput, production, and returns, with the updated capital estimate reflecting the larger scale and design improvements."
— Brandon Craig, Chief Executive Officer

Assessment: This is the good version of a capital increase: more capacity, better returns, worse capital intensity, and a stated improvement in project economics. The IRR moving from 13% to 16% to 16% to 18% at an unchanged copper deck is the number that justifies the extra spend. The reason to note it anyway is that a capital estimate moving up 7% to 23% between a preliminary and a pre-FID stage is exactly the pattern that produced the Jansen write-down, and this project has not reached FID.

13. Active Capital Management: US$6.3bn Executed, US$3.5bn to Go

Both February commitments landed, with a wrinkle. The Antamina silver stream completed and delivered US$4.3bn of cash in FY26. The WAIO inland power transaction did not complete in H2 FY26 as guided; instead, the December 2025 agreement was terminated and replaced with a new transaction, with Global Infrastructure Partners providing US$2.0bn in August 2026 for a 49% stake in a partnership. BHP retains full operational control of WAIO including the inland power infrastructure, and the new agreement does not affect ownership of any WAIO assets. Roughly US$1bn of cash came in during the year from previously announced non-core asset sales, and BHP now sees potential to unlock up to a further roughly US$3.5bn.

"This includes the potential to unlock around $10 billion in undervalued capital, with $6.3 billion of this already executed in the last nine months. On top of this, we received almost $1 billion in cash during the year from previously announced non-core asset sales."
— Vandita Pant, Chief Financial Officer

Assessment: The pillar holds. US$6.3bn executed against an "up to US$10bn" ambition inside nine months is a good strike rate, the structures transferred no ownership of copper or iron ore, and the residual US$3.5bn keeps the option alive. The wrinkle deserves a note: a signed December agreement was terminated and re-cut with the same commercial effect under a new structure, and the proceeds landed in FY27 rather than FY26. Nothing in the disclosure suggests a problem, and nothing in it explains what changed.

14. BMA Is Not for Sale

Craig used the presentation to close down a persistent Street question about the steelmaking coal business, in the plainest language in the transcript.

"Having spent a fair bit of time there as well, I know what that business is capable of and firmly believe its best years are in front of it. We have a plan to improve performance and returns, and that plan doesn't include selling the business. Our ongoing focus on strengthening the supply chain and cost productivity is expected to lift production and reduce costs over the medium term. We believe there is potential to invest further if fiscal conditions improve."
— Brandon Craig, Chief Executive Officer

BMA earned US$702M of Underlying EBITDA and US$161M of Underlying EBIT on US$3,876M of revenue and US$6,421M of net operating assets in FY26. The medium-term plan is 43 to 45 Mtpa (100% basis) at below US$120/t, from 37.3 Mt at US$134.05/t. BHP is explicit that the Queensland royalty regime "is not supportive of long-term capital investment in steelmaking coal assets in Queensland."

Assessment: Ruling out a sale is a real decision with a real cost. BMA generates a segment ROCE of zero and is now competing for management attention and incremental capital against a copper pipeline that BHP itself describes as the best in the sector. The conditional in the last sentence, "if fiscal conditions improve," is the honest version: this is an option on Queensland tax policy, held on US$6.4bn of net operating assets, in a portfolio whose entire investment case is the copper mix shift.

15. The Operating System, the Productivity Target, and the Technology Run-Rate

BHP put quantitative targets on the productivity programme for the first time. The Operational Excellence Index score has risen to 52 from 36 since 2020, with roughly 7,000 initiatives actioned and more than US$5bn of first-year cost savings. Craig set a company-wide target of 65, described as beyond external world-leading benchmarks. Technology initiatives were running at almost US$500M of incremental annual EBITDA by the end of 2026, with a target above US$650M by the end of 2027.

"We assess BOS maturity through an Operational Excellence Index, or OEI, and since 2020, we have increased our score from 36 to 52. Over this period, we have actioned around 7,000 initiatives right across the business, and in their first year of implementation, these delivered cost savings of more than $5 billion. There is still so much more we can do. To capture this value, I have tasked our teams to deliver significant sustained improvement towards a company-wide mark of 65."
— Brandon Craig, Chief Executive Officer

Assessment: An incremental US$150M of technology EBITDA inside twelve months is roughly 0.5% of group Underlying EBITDA, which is worth having and is not a thesis. The more interesting number is the OEI target, because it is the first internally-scored operating metric BHP has published with a destination attached. It is also unauditable from outside, which is why the useful test remains the unit-cost lines at guidance exchange rates rather than the index.

16. The Commodity Outlook Management Is Underwriting

BHP's demand case is unchanged in direction and slightly firmer in specifics. Copper demand grows from roughly 34 Mtpa today to more than 50 Mtpa by CY2050, with data-centre-related copper demand growing roughly sixfold between 2024 and 2050 to around 3 Mtpa; CY2026 demand growth of around 2.8% against 2.1% in CY2025, trimmed by the Middle East conflict. Chinese real steel production plateaus around 1 Btpa for the rest of the decade, with India structurally shifting to net iron ore imports, possibly above 80 Mt by 2030, against seaborne supply additions including Simandou. Potash prices rose 23% in FY26 to US$342/t Vancouver FOB and the rally is expected to be "tempered in CY27."

"As this slide shows, at spot prices, we expect to generate around $50 billion in attributable free cash flow over the next five years. That is after funding our investment in growth. Even under a sustained multi-year downside commodity price scenario, we still expect to generate around $15 billion of free cash flow over the period."
— Vandita Pant, Chief Financial Officer

Assessment: The downside stress test is the disclosure worth keeping. In February the same exercise produced roughly US$60bn at spot and roughly US$10bn in a prolonged downside; it is now roughly US$50bn at spot and roughly US$15bn in the downside case. The spot figure fell by roughly US$10bn while spot prices rose, which is the capital increase showing up in the five-year cash number, and the downside figure improved by roughly US$5bn, which is the balance sheet and the cost base. The two moves together are a fair summary of the year: more resilient, and more expensive to grow.

Guidance & Outlook

FY26 Delivered vs. Guided

The adjusted effective tax rate is BHP's global adjusted corporate tax rate of 36.5%, rising to 42.9% once royalties are included, on total payments to governments of US$12.4bn.

FY27 Guidance

FY27 unit-cost ranges are set at AUD/USD 0.70 and USD/CLP 890, against FY26 realised rates of 0.68 and 920, and Copper South Australia's range assumes gold at US$4,300/oz and uranium at US$80/lb against FY26 guidance assumptions of US$2,900/oz and US$70/lb. WAIO and BMA ranges are subject to the diesel benchmark, with every US$10/bbl move worth roughly US$0.15/t at WAIO and roughly US$1.10/t at BMA.

Medium-Term Guidance: Rebased

How much of the cost rebase is currency. The Australian dollar assumption moved from 0.65 to 0.70, which raises the US-dollar cost of an Australian-dollar cost base by 7.7%. WAIO's ceiling moved 8.6% and BMA's 9.1%. Most, though not all, of the increase in the two Australian assets' medium-term cost targets is therefore the currency rebase rather than a deterioration in the underlying plan. The Chilean assets moved the other way, with Escondida's medium-term range cut US$0.20/lb against a peso assumption that moved against it.

Guidance style. BHP's pattern is consistent and worth stating plainly, because it changes how the ranges should be read. Volume guidance is set conservatively and delivered inside the range or above it: eight of the ten FY26 production lines landed inside the range and the other two landed above it. Unit-cost guidance is set in wide ranges at a fixed exchange rate and then steered to an end of the range during the year rather than reset; FY26 saw the Escondida range narrowed to US$1.00 to US$1.20/lb from the US$1.20 to US$1.50/lb standing at the December half, and by July every copper asset was steered to the bottom end. The place the pattern does not hold is project capital, where estimates have moved materially and repeatedly on Jansen and, at a smaller scale, on the Escondida New Concentrator.

What is not guided. There is no revenue, EBITDA or earnings-per-share guidance, which is normal for the sector. There is also no stated payout-ratio intention for FY27, no plan for what happens to the balance sheet now that net debt sits below the bottom of its own target range, no identification of which assets make up the residual roughly US$3.5bn of unlockable capital, and no framework for what BHP would do if copper retraces toward the level that made FY26's earnings.

Analyst Q&A

BHP's full-year result is delivered as a recorded presentation of prepared remarks by the Chief Executive Officer and Chief Financial Officer. There is no analyst question-and-answer segment, and none was published, for the second consecutive result. The section that follows carries the weight that Q&A would normally carry.

What They're NOT Saying

With no analyst question segment, the omissions are the whole of the adversarial content. Nothing below was asked and dodged. All of it went unaddressed in a presentation management controlled end to end.

1. That price supplied more than 100% of the earnings growth. The waterfall is published in the results announcement and shows a US$7,311M net price impact against a US$6,969M total increase. The presentation's version is that operational performance "enabled us to fully capture that benefit," which is true and is a different sentence. Nobody put the two numbers side by side.
2. The US$1,265M of provisionally priced revenue. Disclosed in the revenue note, against US$24M in FY25, and never mentioned. Nor was the 423 kt of copper still unpriced at 30 June 2026, whose final price is set during the FY27 first half and which carries a US$371M post-tax swing per 10% price move.
3. That the C1 methodology changed in the year the metric rose. "C1 unit costs up just 1%" is the new basis, which includes inventory movements. The previous basis, in force at the December half, gives 8.3%. Both appear in the results announcement; only one appeared in the presentation.
4. What happens to the balance sheet from here. Net debt of US$8.7bn sits below the bottom of an unchanged US$10bn to US$20bn target range while the company guides to roughly US$11bn of annual capital spend through FY31. Whether the intention is to re-lever into the range, hold, or distribute the gap is the most consequential open capital-allocation question, and it was not addressed.
5. Which assets make up the remaining roughly US$3.5bn of unlockable capital. The same gap we flagged in February, now smaller. Two executed transactions were a by-product stream and an infrastructure partnership. Neither template obviously generalises to another US$3.5bn without touching something closer to the core.
6. Why the December WAIO power agreement was terminated and re-cut. The results announcement records that BHP completed a transaction with Global Infrastructure Partners in August 2026 through a new UK structure with the same commercial effect as the agreement announced on 9 December 2025, and that the December agreement has been terminated. The proceeds slipped from H2 FY26 into August 2026. No explanation was offered for the restructuring.
7. The China Mineral Resources Group. BHP's iron ore commentary runs to Chinese steel output, Indian import growth, Simandou and cost support. Neither the results announcement nor the Annual Report mentions the centralised Chinese buying entity by name, while sell-side commentary on the June-quarter review treats a long-term contract with it, denominated in renminbi, as a fact of the year. This was on our February watch list and remains undisclosed by the company.
8. What the FY27 payout ratio should be. FY26 ran at 66%, above the 50% policy minimum, with the June-half at 72% and explicitly enabled in part by the Antamina proceeds. There is no statement of intent for a year with lower copper volumes, higher unit costs and higher capital spending.
9. Whether the Escondida New Concentrator estimate is final. The capital range moved to US$5.4 to US$6.3bn from US$4.4 to US$5.9bn before reaching a final investment decision that is still one to two years away, on a project whose scope also grew. Given what happened at Jansen, the sensitivity of this estimate to further scope change is a question, and it was framed only as an improvement.
10. The by-product dependence inside the Copper South Australia cost number. A 73% unit-cost reduction to US$0.32/lb was attributed to by-product credits and inventory movements, and FY27 guidance is struck on gold at US$4,300/oz and uranium at US$80/lb. What the asset's cost position looks like at a materially lower gold price was not addressed.
11. The BMA return on capital. Management ruled out selling the business and committed to invest further "if fiscal conditions improve." The Underlying EBIT of US$161M on US$6,421M of net operating assets, and the segment ROCE of 0%, are in the results announcement and were not discussed.

Market Reaction

• Pre-print setup. The ADS line closed at US$88.37 on 17 August 2026, up 46.4% year to date against the S&P 500's 13.1%, up 65.4% over twelve months, up 9.6% over the trailing 30 days, and 4.4% below its 52-week closing high of US$93.15. The 52-week closing range entering the print was US$52.14 to US$93.15. On the local line, BHP.AX closed at A$62.20.
• After hours. BHP released at 8:30am AEST on 18 August, roughly 6:30pm ET on 17 August, after the US close. The ADS traded up 3.32% to US$91.30 in the after-hours session, and opened the following morning at US$88.65, up 0.3%.
• Reaction session. On 18 August the ADS traded US$88.24 to US$90.46 and closed at US$89.09, up 0.8% or US$0.72, on 3.3M shares against a 30-day average of 2.5M, or 1.4 times normal volume. Roughly three-quarters of the overnight gain was surrendered inside the session.
• Local line and peers. BHP.AX closed at A$63.85 on 18 August, up 2.65%; the gap to the ADS move is currency. Over the same session the S&P 500 fell 0.7%, Rio Tinto fell 0.5%, Vale fell 0.5% and Fortescue's local line fell 0.4%.

The shape of the reaction is more informative than its size. An after-hours print of US$91.30 says the first read of the headline numbers was strongly positive: a 30% increase in underlying profit, a dividend well ahead of published broker estimates, and a balance sheet below its own leverage floor. The fade to US$89.09 by the close says that once the market worked through the FY27 guidance ranges, the unit-cost rebases and the medium-term capital path, it revised the enthusiasm down without reversing it.

That is the correct reading of this result. The dividend surprise was real, and the surprise was arithmetic rather than strategic: the payout ratio came in at 66% against roughly 63% expected, on profit that itself beat by 4.3%. Neither is a re-rating event. What the session had to price against it was a fiscal year in which the largest earnings driver produces 12% less, four of five major assets cost more per unit, capital spending rises roughly 7%, and the out-year capital path moved up by roughly US$1bn a year.

A 0.8% gain against a tape down 0.7% and a peer group down 0.4% to 0.5% is genuine relative outperformance and says the result did real work. It is not the reaction of a market that has just been given a reason to pay a higher multiple. The stock finished the session 4.4% below the 52-week closing high it set before the print.

Street Perspective

Debate: Is FY27's copper step-down a timing issue or the start of a decline?

Bull view: The decline is planned, grade-driven and disclosed, and the medium-term Escondida range of 900 to 1,000 ktpa for FY28 to FY31 is unchanged. Behind it sits a pipeline that BHP says takes attributable copper to roughly 2 Mtpa (roughly 2.5 Mtpa copper equivalent) by FY35, an increase of roughly 40% on current levels, and copper prices have already moved well above the FY26 realised average. FY27 is a trough to be bought, not a trend.

Bear view: A trough is still a trough, and this one runs through a year in which unit costs rise at four of five major assets and capital spending rises. Group copper equivalent output falls roughly 6% year on year against prior market expectations of roughly 4%, and one desk has the group copper decline at 12%. The stock is being asked to hold a 65% twelve-month gain through it.

Our take: The bulls are right about the shape and the bears are right about the timing. Nothing in this result damages the FY29-plus copper case; the Escondida permit was submitted, the concentrator scope improved, Copper South Australia set records and Vicuña cleared its Argentine approvals. But an investor who buys today owns FY27 first, and FY27 is a year whose earnings depend on the copper price doing roughly 75 US cents per pound of work to stand still.

Debate: Does the Jansen write-down change how the growth pipeline should be valued?

Bull view: Jansen is one project in a portfolio, it remains a genuine Tier One asset, roughly 80% of BHP's projects over the past decade came in on time and on budget, and the impairment is non-cash. The new chief executive has named project delivery as a personal priority and the fix is the same operating system that has worked on the assets.

Bear view: Three capital revisions in thirteen months on the only major project actually under construction, totalling roughly US$3bn against the mid-2025 estimates, is a pattern rather than an incident. The write-down was taken against capital intensity across Stages 1, 2 and potential future expansions, which is a statement about the whole asset. The Escondida New Concentrator estimate has already moved up before FID.

Our take: The bears have this one. The correct response is not to sell the growth story but to apply a haircut to every pre-FID capital estimate in it, including Escondida's US$5.4bn to US$6.3bn, Copper South Australia's US$18,000 to US$23,000 per tonne of copper equivalent, and Cerro Colorado's improved US$1.4bn to US$2.0bn. A pipeline valued at management's capital numbers is worth materially more than the same pipeline valued at management's capital numbers plus the observed revision rate.

Debate: Is the 66% payout the new normal or a cycle-peak gesture?

Bull view: Net debt sits below the bottom of the target range, free cash flow rose 83%, roughly US$3.5bn of further capital unlocks are identified, and BHP has returned more than US$115bn since the framework was introduced in 2016. A payout at or above 66% is affordable and the board has demonstrated it will pay above the minimum when the balance sheet allows.

Bear view: The final dividend was explicitly enabled in part by a US$4.3bn one-time streaming receipt, FY27 brings lower volumes and roughly US$0.7bn more capital spending, and the policy floor is 50%. Extrapolating 66% into a lower-volume year is extrapolating a monetisation, not a distribution policy.

Our take: Somewhere between, and closer to the bears for modelling purposes. Plan around the 50% minimum plus discretion, not around 66%, until the FY27 first half shows whether the board treats a lower-volume year as a reason to normalise. The 3.9% trailing yield at the current price is the right starting point and the wrong forward assumption.

Debate: Does the balance sheet below target imply buybacks?

Bull view: Net debt of US$8.7bn against a US$10bn to US$20bn range, gearing of 13.4% and 0.3 times leverage is roughly US$1bn to US$11bn of unused capacity. With the copper programme self-funding at consensus prices, the obvious use of the gap is a return of capital.

Bear view: The gap is the funding buffer for a capital programme that just moved up by roughly US$1bn a year for four years, and management said nothing about buybacks. Reading a distribution into silence at a company that has just written off US$2.3bn on a project overrun is optimistic.

Our take: The bears again. Nothing in the presentation, the results announcement or the Annual Report signals a buyback, and the FY27 to FY31 capital path plus the Samarco obligations account for the headroom without one. Treat the sub-range net debt as pre-funding for the copper build, and treat any buyback as upside that has not been signalled.

Model & Valuation Framework

Aardvark Labs does not currently publish a financial model for BHP. The table below sets the drivers we carry into FY27, revised from the working assumptions set at initiation in February.

Where the Stock Trades

FY26 underlying basic earnings per ordinary share of 260.0 US cents is US$5.20 per ADS, each ADS carrying two ordinary shares. Against the 18 August close of US$89.09 that is 17.1 times FY26 underlying earnings. On the statutory basis, 193.6 US cents per share is US$3.87 per ADS and the multiple is 23.0 times, inflated by the Jansen impairment. On a weighted average of approximately 5,079M ordinary shares, market capitalisation is approximately US$226bn and enterprise value approximately US$235bn including net debt, or 7.1 times FY26 Underlying EBITDA.

At the February note the same measures on a trailing-twelve-month basis were roughly 16.7 times underlying, 18.4 times statutory and 6.6 times annualised first-half Underlying EBITDA at US$74.29. The stock has risen 19.9% since, and the underlying multiple has expanded by roughly half a turn because underlying earnings grew alongside it. This is not a stock that has re-rated; it is a stock that has compounded earnings into a broadly unchanged multiple.

Yield. Declared FY26 dividends of 172.0 US cents per ordinary share are US$3.44 per ADS, a trailing yield of 3.9% at the reaction-day close, with the final 99 US cents payable 23 September 2026 and the NYSE ex-dividend date 4 September 2026. On a 55% payout applied to lower FY27 volumes we would plan around something closer to 3%. Free cash flow of US$9.8bn against a US$226bn market capitalisation is a 4.3% free cash flow yield.

Valuation impact. Nothing here moves our estimate of intrinsic value enough to change the rating. The FY26 result was better than we assumed on cash, costs and the balance sheet, and worse than we assumed on project capital and on the FY27 starting point. Those largely offset. What has changed is the price: at US$74.29 in February we wanted the mid-US$60s before paying up, and the stock went to US$89.09 instead. We would become buyers on a retracement into the low US$70s per ADS, on an FY27 first half that holds unit costs inside the guided ranges at realised exchange rates, or on an explicit capital-allocation statement that puts the sub-range net debt to work in copper rather than leaving it as an unexplained buffer.

Thesis Scorecard Post-Earnings

The six pillars below are the ones established at initiation in February and carried in the standing thesis. They are graded against what this result revealed, not re-derived.

Overall: Unchanged on balance. Two bull pillars strengthened on evidence, one held; one bear point was substantially neutralised, and two escalated. The thesis we set in February survives this result intact, which is a different statement from saying the risk/reward improved.

Action: Hold. The upgrade path is unchanged in substance and repriced in level: a retracement into the low US$70s per ADS, an FY27 first half that holds unit costs inside the guided ranges at realised exchange rates, or an explicit statement of what the sub-range balance sheet is for.

Bottom Line

BHP's FY26 is the year the company did almost everything an investor could ask of an operator. Every major asset met its unit-cost guidance at the exchange rates the guidance was set on. WAIO set production and shipment records and held its cost lead for a seventh year. Escondida cut unit costs 10% while its feed grade fell. Copper South Australia set records at Olympic Dam and generated roughly US$1bn of free cash flow. Free cash flow rose 83%, net debt fell below the bottom of the company's own target range, and the dividend came back inside the cash the business produced while still beating what the Street had modelled.

None of that changes the arithmetic underneath. BHP's own waterfall says the price of copper and iron ore contributed US$7,311M against a total earnings increase of US$6,969M, and US$1,265M of the revenue that produced it is a period-end mark on sales whose prices had not settled. In February that ratio was 96% and we called it the central risk of owning this stock. It is now 105%, and FY27 guidance tells us that next year the operating base contracts: copper volumes down 11.7% at the midpoint, unit costs guided higher at four of five major assets, capital spending up roughly 7% and the out-year path up roughly US$1bn a year.

The one thing management controls that went visibly wrong went wrong in the way we said it might. Jansen Stage 2 moved 41% and the auditors required US$2.3bn of it back. The new chief executive named project delivery as a personal priority, which is the right answer, and inherited a genuine track record of roughly 80% on-time-and-on-budget delivery to build from. But the pipeline that constitutes this company's investment case is a series of pre-FID capital estimates, and this year the market learned what those estimates are worth.

The market read it about right. An after-hours print of US$91.30 on the headline numbers, a close at US$89.09 once the guidance ranges were digested, and 0.8% of relative outperformance against a tape and a peer group that were both down. That is a good result being priced as a good result, in a stock that has already gained 65% in twelve months and trades at 17.1 times underlying earnings with the Street's own average target sitting below the price. Maintaining Hold. We would rather own this business than any of its diversified peers, and we would rather own it at a lower price than this one.