Copper Did the Work, Free Cash Flow Did Not: Initiating Rio Tinto at Hold

Results vs. Consensus

Rio Tinto reports semi-annually, so the 19 February release is a full-year result covering the twelve months to 31 December 2025, and the Street's forecasting effort concentrates on full-year underlying earnings, underlying EBITDA and the final dividend rather than on a standalone fourth-quarter earnings per share. On the two lines that matter most the company came in modestly light. On revenue it came in ahead. Neither gap was the informational content of the day.

FY25 Scorecard

Income statement, year on year

Cash flow and balance sheet, year on year

Revenue

Consolidated sales revenue of US$57,638M was the only headline that beat, at 2.1% above the US$56.47bn consensus and 7.4% above the prior year. The composition matters more than the number. Iron Ore segmental revenue fell 8.3% to US$28,989M on a realised Pilbara price of US$90.0 per dry tonne, down 8% from US$97.4. Copper segmental revenue rose 48.0% to US$13,729M and Aluminium & Lithium rose 25.0% to US$17,056M. Iron Ore now supplies less than half of reportable-segment revenue: US$28,989M of a US$59,774M total, or 48.5%, down from 58.0% a year ago.

Two revenue lines deserve a flag because they are not operating income in the ordinary sense. Copper's realised price of 457 US cents per pound excludes provisional pricing adjustments, which added US$758M to revenue in 2025 against a US$92M drag in 2024, a US$850M year-on-year swing driven by a copper price that ended the year 44% above where it started. And Pilbara segmental revenue includes US$2.1bn of freight revenue, down from US$2.3bn. Neither is a problem. Both are reasons to treat the revenue beat as lower-grade evidence than the volume and unit-cost delivery underneath it.

Margins and EBITDA

Underlying EBITDA of US$25,363M rose 8.8% and landed 0.9% to 1.4% below consensus. The group underlying EBITDA margin was unchanged at 40%. Inside that flat number the segment mix moved sharply: Iron Ore margin fell to 52.4% from 53.7%, Copper margin rose to 53.7% from 37.1%, and Aluminium & Lithium rose to 26.8% from 26.0%. Copper is now, for the first time, a higher-margin business than iron ore at the segment EBITDA line.

The waterfall is the most useful disclosure in the release. Prices contributed nothing on net. Exchange rates added US$0.1bn. Volumes and mix added US$2.4bn, net of a US$0.6bn hit from four Pilbara cyclones in the first quarter. General inflation cost US$0.5bn, offset by US$0.1bn of easing energy prices. Operating cash unit costs added US$0.3bn on net, comprising a US$0.8bn improvement partly given back by US$0.4bn of temporary increases at Kennecott and Iron Ore Company of Canada. Exploration and evaluation added US$0.4bn, of which US$0.2bn was a gain on selling 30% of the Winu project. Non-cash costs and other subtracted US$0.6bn, including Simandou operating expenditure as the asset ramps and Arcadium integration costs.

Assessment: Strip the US$195M one-off gain on Winu and the EBITDA increase is US$1.85bn rather than US$2.05bn, which puts the underlying growth rate at 8.0% rather than 8.8%. That is still a good year given a flat price backdrop, and it is the first year in several where an investor can point to the portfolio construction rather than to the iron ore price as the reason earnings moved.

Earnings and the gap between underlying and reported

Underlying earnings of US$10,868M were flat against US$10,867M, a rounding-level outcome that hides US$2.0bn of EBITDA growth being consumed below the line. Depreciation and amortisation cost US$0.6bn more as Oyu Tolgoi ramped and Arcadium consolidated. Interest and finance items cost US$0.2bn more against the US$8.9bn increase in net debt. Tax on underlying earnings cost US$1.0bn more, driven by a larger contribution from Escondida at a higher underlying tax rate together with unrecognised deferred tax assets and disallowed costs across the group. Non-controlling interests took US$0.2bn more, which is what a ramping 66%-owned Oyu Tolgoi looks like on the attributable line.

Reported net earnings fell 13.7% to US$9,966M, and the gap to underlying earnings widened to US$0.9bn of exclusions from a US$0.7bn benefit in 2024. The largest single item is US$0.4bn of after-tax foreign exchange and derivative losses, itself the net of US$0.8bn of post-tax losses on intragroup balances against US$0.3bn of post-tax gains on external net debt. Impairments net of reversals cost US$0.2bn, mostly the Yarwun alumina refinery tailings facility. Closure-estimate changes at non-operating and fully impaired sites cost a further US$0.2bn.

Assessment: The exclusions are conventional and the company's underlying-earnings definition is applied consistently, with no items in the judgemental category in either year. But the direction of travel on the reconciliation is worth watching: a company with US$23.5bn of financing liabilities, a US$17.8bn closure provision and assets in Mongolia, Guinea, Argentina and Canada will keep generating currency and provision noise below the underlying line. Basic EPS at 613.7 cents is 8.3% below underlying EPS at 669.2 cents, and that gap is now structural rather than episodic.

Segment Performance

The reportable-segment structure was recast during 2025 following the organisational restructure announced on 27 August 2025, which cut the product groups from four to three. Iron Ore Company of Canada moved into Iron Ore, Rincon moved into the new Aluminium & Lithium group alongside Arcadium, and the remainder of the old Minerals group (Rio Tinto Iron & Titanium, Borates, Diamonds) now sits outside the reportable segments. Prior-year comparatives below are the company's recast figures.

Product group summary

Reportable-segment revenue exceeds consolidated sales revenue because segmental revenue includes Rio's proportionate share of equity accounted unit sales, principally Escondida. The reconciling item is US$5,237M in 2025 against US$4,048M in 2024. The group EBITDA margin of 40% is calculated on the wider US$62,875M revenue base that includes that share.

Production and price KPIs

The 2025 Pilbara unit cash cost of US$23.5 per wet tonne was US$0.5 per tonne above 2024, and improved from US$24.3 in the first half as shipments recovered from the cyclones. The 2026 guidance range assumes an Australian dollar of 0.67 against 0.64 realised in 2025, which the CFO said in Q&A takes the outcome toward the midpoint of the range before any productivity offset.

Iron Ore

The division did what a mature iron ore business does in a year when the index falls 6%: it held volume and gave back margin. Pilbara production was flat at 327.3 Mt on a 100% basis and shipments fell 1% to 326.2 Mt, an outcome the company frames as a recovery, and fairly so. Four cyclones in the first quarter cost US$0.7bn of EBITDA in total, US$0.6bn of volume and US$0.1bn of recovery costs, and the operation then set record mining rates from April and record shipments in the second half. Segmental revenue fell 8.3% to US$28,989M and EBITDA fell 10.5% to US$15,194M, with US$2.3bn of the decline attributable to lower realised prices across Pilbara and Iron Ore Company of Canada. Underlying ROCE fell to 39% from 48%, and free cash flow fell 31% to US$6,061M as capital expenditure rose 34% to US$4,422M on the replacement-mine program.

The product strategy reset is the structural change of the year. Rio combined the previous Pilbara Blend and SP10 products into a single blend at an average 60.8% iron content, down from 61.6%, with shipments of the new blend starting in July. SP10 volumes halved to 10% of second-half Pilbara shipments on a 100% basis, from 20% a year earlier. Portside sales in China fell to 23.2 Mt from 29.9 Mt, and portside inventory closed at 6.4 Mt from 7.1 Mt.

"I think we're changing our approach the way we think about portfolio because Simandou having been something that's coming is something that's arrived. And so, as we did the work last year on product strategy, we obviously had a pretty clear view around what the future mix would look like in terms of our own portfolio."
— Simon Trott, Chief Executive

Assessment: The volume story is real and the cost control is credible, but the segment is now carrying a 34% step-up in capital spend against a falling realised price, and the four Pilbara replacement mines that consume it do not add capacity, they preserve it. Four of five are ramping or under construction with first production in 2027, and the mid-term system capacity target of 345 to 360 Mtpa is unchanged. An investor is funding maintenance of the cash engine, not growth in it. Iron Ore free cash flow of US$6,061M against US$6,145M of group dividends paid in cash is the single cleanest way to see how much of this company still rests on the Pilbara.

Aluminium & Lithium

Aluminium had an excellent year and lithium did not. Segment EBITDA rose 28.8% to US$4,574M, of which the aluminium business contributed US$4,398M at a 13% ROCE and the lithium business US$176M. Bauxite set an annual production record at 62.4 Mt, up 6% after a 7% increase the prior year, and generated US$1,847M of EBITDA. The realised aluminium price including value-added product premiums rose 17% to US$3,318 per tonne, against an LME average up 9% to US$2,632.

The tariff line is the disclosure to read closely. Rio lost its 10% Section 232 exemption from March 2025 and incurred US$1,030M of gross tariff costs on 1,353 kt of US-destined shipments. The offset came through the Midwest premium, which averaged US$1,301 per tonne duty-paid for the year but US$1,731 in the second half against US$855 in the first.

"The US Midwest premium has adapted to levels fully compensating for the 50% tariff."
— Rio Tinto, 2025 full year results release

Lithium is the problem. The Arcadium acquisition closed on 6 March 2025 and the combined lithium business generated US$944M of revenue, US$176M of EBITDA and US$288M of depreciation and amortisation, which makes it loss-making at the earnings line in its first year. It consumed US$1,365M of capital expenditure and carries US$9,783M of operating assets, up from US$1,088M. On those numbers the lithium business earned an EBITDA return of 1.8% on its operating asset base.

Assessment: The aluminium franchise is the quiet compounder in this portfolio and is being underwritten by a Midwest premium that has, so far, passed the tariff through in full. That is a policy-dependent margin and it should be marked as such. Lithium is a 2028 story being paid for in 2025 and 2026: US$9.8bn of capital employed today, US$2.7bn of remaining capital across Rincon, Fénix 1B, Sal de Vida and Nemaska on a Rio share basis, and a 200 ktpa capacity target that does not arrive until 2028. Management is right that the price cycle is not the point for an asset with a 40-year life. Investors funding four years of build before the return still get to price the wait.

Copper

Copper was the year. Production rose 11% to 883 kt on a consolidated basis, driven by a 61% increase at Oyu Tolgoi as the conveyor to surface came fully online and open-pit grades improved, with better head grades and recoveries at Escondida on top. Mined gold rose 65% to 464 koz. Segmental revenue rose 48.0% to US$13,729M and EBITDA rose 114.4% to US$7,369M. C1 net unit costs fell 53% to 67 US cents per pound, below the 80 to 100 cent range guided at the December Capital Markets Day, helped by gold by-product credits at a rising gold price. Free cash flow rose to US$2,820M from US$526M, and ROCE rose to 14% from 6%.

Two qualifications. Kennecott refined production fell 31% on a planned 45-day smelter shutdown and geotechnical constraints that limit ore availability until higher-grade Slice 2 access in 2027, against a 2024 comparative flattered by inventory drawdown after the 2023 rebuild. And the US$758M positive provisional-pricing adjustment sits inside segment revenue.

"Kennecott is on track to deliver production increase by 40% to 50% over the next few years, as we outlined at CMD. Its operating performance is much improved, but the financials were impacted by the base effect of refining high intermediate product inventories in 2024."
— Peter Cunningham, Chief Financial Officer

The Oyu Tolgoi underground development project completed in the fourth quarter, fully invested against a US$7.06bn capital cost with nil remaining. The asset is guided to average around 500 kt of copper a year on a 100% basis from 2028 to 2036. Against that, Oyu Tolgoi LLC disclosed tax assessments of MNT 1.6 trillion, roughly US$440M, covering the 2021 and 2022 financial years.

Assessment: This is the segment that justifies the equity story, and it is also the segment where the 2026 guide goes backwards. Consolidated copper production is guided to 800 to 870 kt against 883 kt delivered, 5.4% lower at the midpoint, with the footnote that operated assets grow roughly 10% year on year. The decline sits in the Escondida equity share on lower grade. That distinction is fair and the market should respect it, but attributable copper tonnes are what flow to earnings, and they fall in 2026. Copper carried the year and copper is guided to pause.

Outside the reportable segments

Rio Tinto Iron & Titanium generated US$148M of EBITDA against US$609M, on weaker TiO2 demand and a 2024 comparative that included a US$0.2bn one-off insurance receipt. Borates generated US$210M and Diamonds a US$79M loss. Both Iron & Titanium and Borates are under strategic review, and production guidance for them was withdrawn while the process runs. Restructuring, project and one-off costs rose to US$606M from US$254M, concentrated in the first half on Arcadium integration, with second-half costs from cutting the Executive Committee from eleven to nine and reducing senior management roles by 22%. Simandou, still outside the reportable segments, recorded a US$96M EBITDA loss as operating expenditure ramps ahead of revenue.

Assessment: The below-the-segments block cost US$1,678M of EBITDA in 2025 against US$638M in 2024, a US$1.04bn swing that consumed a third of the US$3,163M improvement at the segment level. Some of it is genuinely non-recurring restructuring. Some of it, notably Simandou operating costs and the drift at Iron & Titanium, is the running cost of a portfolio in transition and will persist through 2026.

Key Topics & Management Commentary

Overall Management Tone: This was a composed, on-message call from a leadership team presenting its first full-year result together, and the posture was one of continuity with the December Capital Markets Day rather than of new information. Management was most confident on operational delivery and on the cost program, where it went beyond the guided figure without being pressed. It was most controlled on the Glencore process, where every answer routed back to a single valuation test, and least specific on capital release, where the language stayed at the level of options and systematic evaluation rather than timetables. The one place the tone shifted was safety, where the CEO opened the call on a fatality at Simandou and led with actions taken rather than with reassurance.

1. Walking away from Glencore, and what the process revealed about the bar

Talks with Glencore ended on 5 February, and this was the first public commentary since. Management framed it entirely as a failed valuation test rather than a strategic reconsideration, and returned to that framing on all five occasions the topic was raised in Q&A. The perimeter under discussion was the whole of Glencore, coal included, and the CEO declined every invitation to separate the coal question from the value question.

"We went under the hood with a singular focus on whether we could create value for shareholders. We considered what we could bring to the table and the extent to which we can generate incremental value across a combined portfolio. We had constructive discussions between the two teams. Ultimately, we concluded that we could not reach an agreement that would deliver value for Rio Tinto shareholders."
— Simon Trott, Chief Executive

Pressed on how the gap was measured, the CEO confirmed a bottom-up asset-by-asset valuation with synergies layered on top as one input rather than as the driver. Pressed separately on whether a re-rating of the combined entity carried weight, the answer was the sharpest line of the call.

"So valuation by its very definition is forward-looking. And so it completely flowed into our view of value. But strategic rationales don't pay the grocery bills. It's got to come back to cash accretion for Rio shareholders, and that's the lens we talk."
— Simon Trott, Chief Executive

Assessment: Discipline is the right answer and the framing was consistent under repeated pressure, which is more than can be said for many aborted processes. The investment consequence is that the growth case now has to be carried entirely by the organic pipeline and the cost program, on the timetable the company set in December. There is no transaction to wait for, and nothing management said suggests the same value test would clear on a different day at a different price.

2. The dividend is a policy on earnings, and 2025 shows what that means in a peak-capex year

Rio declared US$6.5bn of ordinary dividends for 2025, a 60% payout on underlying earnings and the tenth consecutive year at the top of the 40% to 60% band. The final dividend of 254.0 US cents was 12.9% above the prior year's 225.0 cents, with the interim at 148.0 cents against 177.0 cents. Cash dividends paid during the year were US$6,145M against free cash flow of US$4,025M.

"Our strong cash flow and balance sheet enable us to sustain a 60% payout ratio with a $6.5 billion ordinary dividend, making it the tenth consecutive year at the top end of the range."
— Simon Trott, Chief Executive, 2025 full year results release

The exchange on whether the band should be recalibrated was the most direct pushback of the call, and the CFO would not be moved past the existing policy.

"Well, I mean, I think I'd sort of push back as well and say that the business has kind of performed at a level to have the 60% payout range. I mean, that's what we've had. I think that's sort of just reflective of the cash flows, quality of assets. And the reality is now we're growing the business. That pie will grow. And so the absolute number in line with the growth of the earnings would increase as well."
— Peter Cunningham, Chief Financial Officer

Assessment: The 60% payout is safe in the sense that it is an earnings policy applied to a stable underlying earnings base, and net debt at 0.57x underlying EBITDA with 18% gearing gives ample room to keep paying it. It is not safe in the sense that a buyer of the shares for the 4.2% yield should understand that in 2025 the yield was funded from borrowing capacity, not from cash generated after investment. That is defensible during a two-year investment peak. It stops being defensible if capital release does not arrive and capex does not step down on schedule.

3. The productivity program: US$650M is the floor, not the number

The December Capital Markets Day set a US$650M annualised run rate of productivity benefits to be reached by the end of the first quarter of 2026, comprising US$370M already realised and US$280M to come. Management used this call to reframe that figure as a starting point, and volunteered a larger 2026 number without being asked for one.

"So on the $650 million, so that was a run rate that we announced at Capital Markets that we'd said we'd hit by the end of Q1. So what we're saying today is that our 2026 cash delivery will be materially above the $650 million, which was a run rate. And so that sizes it for 2026."
— Simon Trott, Chief Executive

"Now nothing has changed from the parameters that we set out at the CMD. We are pushing very hard on productivity improvements and cost reductions building on the initial $650 million already identified and secured. I would, therefore, expect the aggregate volume and cost improvements, net of headwinds, to be a material uplift on that number in 2026."
— Peter Cunningham, Chief Financial Officer

The second phase was described as larger in scale, multi-year, and running through 2027 and 2028, with named workstreams: contingency stockpiles and asset shut-sequencing in the Pilbara, underground equipment productivity and concentrator recoveries in copper, smelter stability and maintenance quality in aluminium, and a central operating-model reorganisation.

Assessment: This is the most investable thing management said. It is also the least verifiable, because no 2026 figure was put on it, and because "aggregate volume and cost improvements, net of headwinds" mixes a controllable cost program with a volume outcome that guidance already shows going the wrong way in copper and bauxite. The company earned some credit here in 2025: it delivered a 5% operating unit cost reduction and beat its own revised copper C1 guidance by a wide margin. We will treat the second phase as an option with real value and no quantification until a number appears.

4. 2026 volumes go sideways at best, and management said so plainly

The clearest forward statement on the call was not in the guidance table.

"Looking forward to 2026, volume growth will be more muted at around 3% across our managed operations, which will be offset by closures at Arvida, Diavik and the midyear curtailment at Yarwun, and an expected grade decline at Escondida."
— Peter Cunningham, Chief Financial Officer

The guidance table bears this out. Copper production is guided to 800 to 870 kt against 883 kt delivered, with a footnote that operated assets grow roughly 10% year on year, which locates the decline in the Escondida equity share. Bauxite is guided to 58 to 61 Mt against 62.4 Mt. Aluminium is guided to 3.25 to 3.45 Mt against 3.38 Mt. Iron ore total sales are guided to 343 to 366 Mt against 342 Mt, with 5 to 10 Mt of that from Simandou.

Assessment: 2025's earnings growth came from volume, and 2026's volume is guided flat to lower on an attributable basis before the cost program is applied. That places an unusual amount of weight on the productivity number nobody has quantified, and on commodity prices that management explicitly declined to forecast. A year in which volume subtracts and the price is a coin flip is a year where the earnings outcome is a cost-program outcome.

5. Capital expenditure stays at the peak for two more years

Rio's share of capital investment was US$11,403M in 2025 against US$9,493M, at the high end of the guided range. The 2026 split is up to US$3.0bn growth, roughly US$4.0bn sustaining, roughly US$3 to 4bn replacement and roughly US$0.2bn decarbonisation, for a total of up to US$11bn.

"Given this context, we see no change to our guidance of up to $11 billion for the next 2 years before stepping down to $10 billion thereafter."
— Peter Cunningham, Chief Financial Officer

Simandou is described as nearly two-thirds complete with US$2.1bn of Rio-share capital remaining. The Pilbara replacement program continues, and a final decision on the Amrun expansion at Weipa is expected later in 2026.

Assessment: Holding the guide is a positive in an industry where capital cost guidance rarely survives contact with execution, and Rio has now delivered Oyu Tolgoi underground, Western Range and Simandou first shipment without a headline overrun. The arithmetic is still unforgiving: hold capital investment near US$11bn and hold operating cash flow at the US$16,832M delivered in 2025, and free cash flow does not cover the dividend in 2026 either. Every year of the investment peak is a year the payout is part-funded from the balance sheet.

6. Capital release: named, sized, and not yet started

The US$5bn to US$10bn asset-proceeds target from the December Capital Markets Day was reaffirmed, with Rio Tinto Iron & Titanium and Borates under market testing and infrastructure monetisation and streaming named as further options. Production guidance for Iron & Titanium and Borates has been withdrawn while the process runs, which is itself a signal that the company expects them to move.

"I mean I suppose all of this comes down to the fact that we've got lots of options across our portfolio to release capital, and that's our focus. I mean, we've talked about the strategic reviews of borates and our RTIT, we're testing the market. We've got options around infrastructure. We do have options around streaming. But we're just going to work through these systematically and say what's the best option that we can undertake."
— Peter Cunningham, Chief Financial Officer

Asked whether a depressed mineral sands cycle argued for waiting, the CEO was explicit that the company is not a forced seller.

"We're going to do it patiently, yes. As I've said earlier, we are a long-term business. And similarly, I think the people that are interested in that or the borates business is going to look through the market as it stands. But we're going to be patient, as you say, we're not under any pressure. And so if we don't get the sort of value that we see in the business, we won't progress them."
— Simon Trott, Chief Executive

Assessment: Both statements are correct and together they remove the timetable. A patient seller of a cyclical asset in a weak part of its cycle, with no pressure to transact, is a seller who may not transact. This is the single largest swing factor in the 2026 and 2027 cash story: US$5bn to US$10bn of proceeds against a US$14.4bn net debt position and a dividend running above free cash flow. Until something is signed we assign it no value in the base case, and we would treat a completed sale at a credible price as the most likely upgrade catalyst for this equity.

7. Simandou arrives, and a fatality stops the site

Simandou achieved first ore shipment in December 2025, with the cargo landing in China in January 2026 and 2.3 Mt of crushed iron ore produced on a 100% SimFer basis during the year. The mine is 62% complete, the rail spur is mechanically complete and operational, and the port is 66% complete and ahead of plan. Commissioning of the shared rail-to-port infrastructure is expected around the end of the first quarter of 2026, followed by a 30-month ramp to full capacity.

The call opened on the death of a colleague at the mine site on the Saturday before the results. The CEO led with the actions rather than the reassurance.

"We've stopped all site works and construction activities. We started an independent investigation with both internal and external experts. And in addition, we will appoint an independent safety advisory panel."
— Simon Trott, Chief Executive

Assessment: Guidance of 5 to 10 Mt of Simandou sales in 2026 was maintained and the 60 Mtpa target reaffirmed, both while site works are stopped. That is a defensible position taken five days after the event and it is also the assumption most likely to move. The wider point is that Simandou converts from a capital-commitment story into an execution-and-ramp story in 2026, and the 30-month ramp means it does not reach full rate until roughly the end of 2028. The group all-injury frequency rate was 0.37 for 2025, consistent with 2024, so this is not a deteriorating aggregate safety record. It is a jurisdiction-specific execution risk that management acknowledged in exactly those terms.

8. Oyu Tolgoi completes, and Mongolia sends a tax bill

The Oyu Tolgoi underground development project completed during the fourth quarter against a US$7.06bn capital cost with nil remaining, and the second primary crusher finished ahead of plan in the third quarter. Production is guided to average around 500 kt of copper a year on a 100% basis from 2028 to 2036. Set against that, the release disclosed a new tax dispute.

"Oyu Tolgoi LLC confirms that it has received tax assessments amounting to MNT 1.6 trillion (approximately $440 million) in primary tax interest and penalties from the Mongolian Tax Authority. This assessment pertains to a tax audit covering the financial years 2021 and 2022."
— Rio Tinto, 2025 full year results release

The company disputes the assessment as inconsistent with the Oyu Tolgoi Investment Agreement. Separately, the Entrée licence transfer remains under discussion with the Government of Mongolia, and the mine plan retains flexibility to bring either Panel 1 or Panel 2 South into production first depending on that timing.

Assessment: US$440M is not financially material against a US$156bn market capitalisation, and Rio has settled Mongolian tax disputes before. The signal value is higher than the dollar value. The asset that supplies the growth in this equity story sits in a jurisdiction that has now assessed it twice, and where a licence transfer needed for optimal mine sequencing is still not done. That belongs in the discount rate, not in the earnings model.

9. Lithium: a US$9.8bn balance-sheet position earning US$176M

Arcadium closed on 6 March 2025 for US$7.6bn including US$0.7bn of acquired net debt, funded from a US$9bn bond issue that is the direct cause of the US$8.9bn increase in net debt. The combined lithium business produced 57 kt of lithium carbonate equivalent, of which 46 kt was attributable to Rio, on US$944M of revenue and US$176M of EBITDA against US$288M of depreciation. It consumed US$1,365M of capital expenditure and carries US$9,783M of operating assets.

Asked whether a lithium price that had roughly doubled since the December site visit changed the plan, the CEO declined to accelerate.

"And the lithium, just given the size of the industry and the rate of growth, we fully expect prices in lithium to be volatile, and we've certainly seen that over the last little while. But we've got to look through that at the long-term pricing because those assets, once we bring them into production, they are going to be in production for decades. And so it's not so much about next week, next month. It's about the years that follow."
— Simon Trott, Chief Executive

Assessment: The discipline is right and the refusal to chase a price spike into incremental capital commitments is exactly what an investor should want from a company that has just raised its net debt by 162%. The problem is that the discipline is being applied after the acquisition, not before it. As of today the lithium business is US$9.8bn of operating assets returning 1.8% at the EBITDA line, with the capacity target that justifies it not arriving until 2028. It is the clearest example in this portfolio of value that depends on execution across three more years of build.

10. Diversification did the job, and management knows which chart proves it

Iron ore's realised price fell 8% and its EBITDA fell 10.5%, yet group EBITDA rose 8.8%. That is the entire argument for the portfolio construction of the past decade, and management made it explicitly when asked about geopolitical risk.

"I think in the numbers today, you can see the real value of the diversified model, and it goes a little bit to your question as well, whilst iron ore prices were down, EBITDA has gone up because of greater contribution from copper as we ramp up and obviously, a strong contribution from aluminum as well."
— Simon Trott, Chief Executive

The forward commitment is a 3% compound annual growth rate in copper-equivalent production to 2030 and a 4% compound annual unit-cost improvement over the same period, with 85% of the exploration budget now directed at copper.

Assessment: This is the strongest structural argument in the equity and it is now evidenced rather than asserted. The qualification is that 2025 was the year the diversification paid because copper volume happened to arrive at the same moment the iron ore price fell. In 2026 the copper volume pauses. The portfolio is genuinely more balanced than it was; it is not yet balanced enough that a soft iron ore price and a flat copper year can both be absorbed.

11. Aluminium tariffs, Yarwun, and the footnotes that matter

Two items in the aluminium chain carry more forward significance than their 2025 dollar impact. Rio lost the 10% Section 232 tariff exemption it had held since 2018, incurring US$1,030M of gross tariff cost on 1,353 kt of US-destined shipments, against which the Midwest premium duty-paid rose to an average US$1,731 per tonne in the second half from US$855 in the first. And at Yarwun, Rio will curtail alumina production by 40% from October 2026 to extend the refinery's life to 2035 while it looks for a tailings solution, which drove a US$233M closure-estimate charge and an impairment in 2025.

Assessment: The tariff pass-through worked in 2025 and the disclosure is unusually clear about the mechanism. It is still a policy-contingent margin sitting inside the segment that delivered the second-largest EBITDA increase of the year. The Yarwun curtailment removes volume from mid-2026 and is one of the named offsets in the CFO's 3% volume comment. Neither is a thesis-changer. Both are reasons the 2026 aluminium result is unlikely to repeat 2025's 28.8% EBITDA increase.

Guidance & Outlook

Rio does not guide revenue or earnings. It guides volume, unit cost, capital spend and tax rate, and all four were reaffirmed at the levels set at the Capital Markets Day on 4 December 2025. That is the headline: nothing changed between December and February. The table below sets the 2026 guide against 2025 actuals rather than against a prior guide, because there is no prior guide to move.

Production guidance for Iron & Titanium and Borates has been withdrawn while the strategic reviews run, so the group's disclosed volume guidance now covers a smaller share of the portfolio than it did a year ago.

Implied 2026 shape: Iron ore sales grow 3.7% at the midpoint, and 60% of that increment is Simandou at 5 to 10 Mt on a 100% basis. Copper falls 5.4% at the midpoint on a consolidated basis while operated assets grow roughly 10%, which means the attributable decline is an Escondida grade effect. Bauxite falls 4.6%. Lithium grows 9.6% off a small base. Net of the Arvida and Diavik closures and the Yarwun curtailment, the CFO's stated managed-operations growth of roughly 3% is close to nil at the group attributable level. On costs, Pilbara guidance is 3.2% higher at the midpoint at a currency assumption 4.7% less favourable, and the CFO's Q&A comment was that the exchange rate alone takes you toward the midpoint of that range.

Street at: Every published estimate for 2026 was set against the December Capital Markets Day parameters, and management reaffirmed all of them. The result is that this release moves 2026 consensus only through the 2025 base effect, which is 1.4% to 2.1% lower on underlying earnings than the Street had modelled. Same-day sell-side commentary characterised the print as in line with no new reveals since December, and that is a fair reading of the guidance section.

Guidance style: Rio guided conservatively in 2025 and beat itself twice on the lines that carry the most operating leverage. Copper C1 net unit costs of 67 cents came in below the 80 to 100 cent range revised at the December Capital Markets Day, and the effective tax rate of 31.5% came in below the 33% guided, though the tax outcome was helped by re-recognising deferred tax assets in Australia and the United States in the second half rather than by operations. The Pilbara unit cost landed exactly in line at US$23.5. On that record the 2026 volume ranges are probably honest midpoints rather than sandbagged floors, and the cost ranges have a modest bias to the low end.

Analyst Q&A Highlights

What the failed merger process revealed about the valuation bar

This was the dominant topic in Q&A, raised in five separate exchanges from the opening question onward, and management held a single line throughout. The first questioner went straight at whether coal ownership had been the obstacle, and the answer refused the framing: the perimeter was the whole company, and the test was value.

Q: "So, do you feel comfortable owning coal? That would be your first question. What do you think you've learned from the discussions? What sort of synergies did you see from that sort of combination? Obviously, the value didn't work, but any other issues that kind of stopped the deal from happening?"
— Myles Allsop, UBS

A: "So you always learn through these processes. The constructive discussions, you learn, I guess, about your own business, you learn about others as well. And as I said in my presentation, we went deep, we went under the hood. We look rigorously and clinically and ultimately didn't get there on value. The discussions were for the full perimeter."
— Simon Trott, Chief Executive

Assessment: A clean, consistent answer that survived four re-tests, including a direct challenge on whether synergies had been undervalued relative to what the counterparty had said publicly the previous day. Management never conceded a second reason, which is either genuine discipline or a very well-rehearsed line. The behavioural evidence supports the former: the same value test is what stopped the deal and what is now slowing the asset sales.

Sizing the cost program beyond the announced run rate

A recurring line of questioning pressed for a number on the second phase of the productivity program, and specifically on how much can come out of the Pilbara. Management sized 2026 and declined to size 2027 and 2028, while confirming the program spans all businesses rather than iron ore alone.

Q: "And the second question is on cost cutting. You've put out a slide there, looking at the cost-cutting opportunities beyond the $650 million program. The Pilbara seems to be at the heart of it. Can you help us frame a little bit the opportunity there to quantify how much can be taken out of the business in terms of costs?"
— Alain Gabriel, Morgan Stanley

A: "I think the main point here, and Pete talked about it, we've gone systematically asset-by-asset looking at full potential with clear plans then around delivery, and it will be a multiyear program. And so we've sized it for 2026, but clearly, there's more to come in '27 and '28. And I should say it's across all businesses. So yes, iron ore, but it's across each of our businesses in the portfolio."
— Simon Trott, Chief Executive

Assessment: The refusal to quantify beyond 2026 is normal practice and not a dodge, but it leaves the largest single driver of the 2026 earnings outcome as an unquantified commitment. Note also what "sized it for 2026" actually means: materially above a US$650M run rate, with no upper bound and no split between volume and cost. That is a narrative, not a bridge.

Whether the Pilbara can close the unit-cost gap to its neighbours

The sharpest operational challenge of the call put Rio's Pilbara cost base directly against the guided targets of the two other major Pilbara producers and asked how the medium-term ambition gets delivered. Management pushed back first on comparability, then handed to the CFO for the mechanism.

Q: "Now I understand, obviously, your mine systems are quite different to theirs. But today, in terms of the next phase examples, you've talked about the Pilbara. So I guess, how can you better that $20 a tonne? And what level of betterment do you think we can expect? And sort of where can you end up in terms of where you sit versus your competitors?"
— Rahul Anand, Morgan Stanley

A: "Probably the first point I'd make is you've got to look at it on apples-to-apples. And people can flip between full unit costs and C1 costs. But the numbers that you're referring to for us anyway is about full unit cost, and so got to compare the same."
— Simon Trott, Chief Executive

Assessment: The apples-to-apples point is legitimate and frequently ignored in cross-company Pilbara comparisons. It is also a deflection from the question that was asked, which was about the path to the medium-term target rather than about the peer benchmark. The substantive answer came from the CFO and rested on the replacement mines carrying volume through the system, which is the same argument that justifies the 34% increase in iron ore capital spend. The cost target and the capital program are the same commitment viewed from two ends.

Whether a fatality changes the Simandou ramp

The most uncomfortable exchange of the call asked directly whether the safety record at the project threatens the volume target, and whether a sale to the partners was on the table. Management reaffirmed the target while conceding the operating practices need to adapt to the jurisdiction.

Q: "Firstly, on Simandou. It seems to have a high rate of fatalities for the time period. And obviously, you haven't changed your guidance for this year. But looking forward, like do you see a risk to kind of hitting that 60 million tonnes in a reasonable period of time unless the safety culture improves quite dramatically? If not, would you consider like portfolio adjustments, i.e., potential disposals of Simandou to your partners?"
— Ephrem Ravi, Citigroup

A: "And I think the team at Simandou have made enormous strides and the events of the weekend show we've got further to go. And so that's our real focus at the moment. And I think the work that they have done, we know we can get there. We've just got to put in place the blocks to make sure that we really can. It is a different jurisdiction in a different environment, and we need to adjust our operating practices to that, but we're confident of the 60 million tonnes that we've announced."
— Simon Trott, Chief Executive

Assessment: The disposal half of the question went unanswered, which is itself informative: the project is not for sale. Reaffirming 60 Mtpa days after stopping all site works is a statement about long-run capacity rather than about near-term schedule, and the two are being conflated. We would treat the 5 to 10 Mt of 2026 sales guidance as the number most exposed to the investigation and the safety review, and the 60 Mtpa nameplate as unaffected.

Whether the payout band still has a lower half

The most persistent pushback outside the merger topic asked why a policy band of 40% to 60% has any meaning when the company has paid the top of it for ten consecutive years, and whether falling capital spend argues for recalibrating higher. Management defended the existing framework twice and would not be drawn to a floor.

Q: "If I can push back a little bit. What's the point of having the low end of the range of 40%, if over the last 10 years, not every year has been an easy year, but you've never paid 40%."
— Alan Spence, BNP Paribas

A: "Well, I mean, I think I'd sort of push back as well and say that the business has kind of performed at a level to have the 60% payout range. I mean, that's what we've had. I think that's sort of just reflective of the cash flows, quality of assets. And the reality is now we're growing the business. That pie will grow. And so the absolute number in line with the growth of the earnings would increase as well. I think that's a pretty good place to be. It's growth and its returns."
— Peter Cunningham, Chief Financial Officer

Assessment: The exchange ended with a one-word non-answer to a follow-up asking whether the policy is now effectively a 60% minimum. That is the correct answer for a CFO to give and the wrong one for an investor to rely on. The band exists precisely so it can be used, and 2025 is the first year in a decade where free cash flow would have supported the lower half rather than the upper. The dividend is not at risk in 2026; the assumption that 60% is a floor is.

Geopolitical risk and the price of entering new jurisdictions

One exchange stepped back from the quarter entirely to ask how a company whose growth now sits in Mongolia and Guinea prices assets in jurisdictions it has historically avoided. The answer was unusually candid about the absence of a clean method.

Q: "So when you're valuing Glencore in that situation, how do you -- is it a much higher discount rate that you're using? How do you get comfort around assets in those types of regions?"
— Christopher LaFemina, Jefferies

A: "Look, it's an excellent question, and it's one that we spend a lot of time grappling with and thinking about it, and I'm not sure there's a perfect answer. You're right in the sense of, ultimately, it's got to come back to value. And so a higher discount rates, the way you think about the opportunity could clearly, in more challenging projects, whether they're more challenging because of the jurisdiction or more challenging because of technical aspects. The size of the prize has to be there to really step in and take on some of those challenges."
— Simon Trott, Chief Executive

Assessment: The most revealing answer of the call, and the one that best explains the merger outcome. A management team applying a jurisdiction premium to a portfolio spanning several higher-risk geographies will arrive at a lower valuation than a market that prices the same assets on spot cash flow. That is a defensible way to run a balance sheet. It also means this company is structurally unlikely to be the winning bidder for high-return, high-risk assets, and investors should stop modelling transformational acquisitions into the story.

Converting the copper option set into projects

A closing line of questioning contrasted Rio's copper growth pipeline with a peer's recently published multi-decade project list and asked how the longer-duration options progress. Management leaned on near-term delivery rather than on the pipeline, which is the honest framing given Resolution remains enjoined and Winu is only at feasibility.

Q: "How are you guys thinking about those longer-duration copper growth options that you may have in your portfolio, noting resolution currently is still in the ground and not being mined. And I'm sure you would like to have a project there. But can you give us a bit of a flavor as to how the copper JV is going as well with Codelco and how quickly that's progressing?"
— Robert Stein, Macquarie

A: "I guess the nice thing about today's results is we're growing today. And then we've got the 3% copper equivalent growth through to 2030. And so that's why we've tended to focus on the here and now because our growth is through this period. And we have the options then to extend that growth out into the 2030s. And so we'll come to market and update as those projects commence -- progress."
— Simon Trott, Chief Executive

Assessment: "We're growing today" is the right rebuttal for 2025 and a weaker one for 2026, when consolidated copper production is guided down 5.4% at the midpoint. On the options themselves: Resolution had oral arguments completed on 7 January 2026 with a decision anticipated during the year, La Granja drill results are expected in the first quarter, Winu's feasibility study concludes at the end of 2026, and the two Chilean joint ventures are expected to close in the first half. None of them contribute a tonne this decade. The 3% copper-equivalent compound growth rate to 2030 rests on assets already built.

What They're NOT Saying

1. No number on the second phase of the cost program. Management sized 2026 as "materially above" a US$650M run rate and confirmed the program extends through 2027 and 2028, but declined to quantify either the 2026 figure or the multi-year target. In a year where guided volumes are flat to down, that unquantified figure is the swing factor in the earnings outcome.
2. No timetable on the US$5bn to US$10bn of capital release. Iron & Titanium and Borates are at market testing, infrastructure and streaming are described as options, and the explicit framing is that the company is not a forced seller and will not transact below its own value. No expected completion window was given for anything.
3. No revision to the Simandou 2026 sales guidance after the fatality. All site works and construction activities were stopped, an independent investigation is running and an independent safety advisory panel is to be appointed, yet the 5 to 10 Mt sales guide and the end-of-first-quarter commissioning milestone were both left in place. Either the stoppage is expected to be very short, or the guidance has not yet been revisited.
4. No standalone second-half income statement. Rio reports on a full-year basis, and the release provides half-year splits only for Pilbara and aluminium realised prices, Pilbara unit costs and the US tariff cost. There is no disclosed half-on-half EBITDA or earnings, which makes the second-half exit rate a matter of inference rather than of disclosure in a year where the first half absorbed four cyclones.
5. No discussion of what a 60% payout costs when free cash flow is US$4bn. Management was asked twice whether the payout band should be recalibrated upward and answered both times on the framework. Neither question was about the downside, and management did not volunteer the observation that 2025 cash distributions were 153% of free cash flow.
6. Nothing on the Chinalco stake beyond confirming engagement. Asked directly whether discussions over the shareholding in Rio Tinto plc had progressed, the answer was that engagement continues and there is nothing to announce, delivered alongside a reference to the separate Brazilian aluminium transaction with the same counterparty.
7. No quantification of the Yarwun care-and-maintenance cost. The direct question on the annual cost of the curtailment received "low single digits" without a unit, and no long-term plan for the asset beyond the extension to 2035 while a tailings solution is sought.
8. No 2026 guidance for Iron & Titanium or Borates at all. Withdrawn while the strategic reviews run. Those two businesses generated US$358M of combined EBITDA in 2025, so the omission is not material to the earnings model, but it does mean the group's disclosed volume guidance now covers less of the portfolio.
9. No provision commentary on the Mongolian tax assessment. The US$440M assessment is disclosed and disputed, with no indication of whether anything has been provided against it or what the range of outcomes looks like.

Market Reaction

• Pre-print setup: RIO closed at $98.93 on 18 February, up 23.6% year to date against the S&P 500's +0.5%, up 57.6% over twelve months and up 15.5% over the trailing thirty days. The 52-week closing range into the print was $52.32 to $99.52, so the stock entered the result 0.6% below its 52-week closing high.
• Print-day session: Rio Tinto reports before the US open. The shares gapped down 3.8% to open at $95.21, traded a $94.09 to $96.45 range and closed at $96.34, down 2.6% or $2.59 on the day.
• Volume: 4.0M shares against a 4.6M thirty-day average, or 0.9x. There was no volume event.
• Peers on the same session: BHP +0.7%, Vale +0.6%, Freeport-McMoRan +0.0%, Alcoa -0.9%, against the S&P 500 at -0.3%. The diversified miners did not move with Rio, which makes this a stock-specific reaction rather than a sector one.

A miss on a stock priced for a beat. The gap between the print and the reaction is not explained by the size of the miss. Underlying earnings came in 1.5% to 2.1% light and underlying EBITDA 0.9% to 1.4% light, which on any ordinary day is a rounding error for a company of this scale. What made it matter is the setup: a stock that had added 57.6% in twelve months and sat within 0.6% of its 52-week closing high had already priced the copper ramp, the cost program and the December Capital Markets Day plan. In that position, in line with no new information is a sell signal, and that is exactly how the day traded.

The comparison two days earlier did not help. BHP had released its half-year result on 17 February with underlying EBITDA up 25% and underlying attributable profit up more than 20%. Rio's +8.8% EBITDA and flat underlying earnings, presented into that comparison, read as the weaker of the two diversified prints even though the two companies were reporting different periods against different bases. The peer moves on the day support this reading: BHP rose while Rio fell, which is not what happens when an iron ore price move drives the sector.

The cash statement is what a careful reader would have sold. Free cash flow down 27.5%, dividends paid at 153% of free cash flow, net debt up 162%, gearing doubled and ROCE down two points is a coherent set of reasons to take profits on a stock at 14.4x underlying earnings. Add the removal of the merger option two weeks earlier and the absence of any completed capital release, and the print left a holder with nothing new to underwrite the last twelve months of multiple expansion. The print-day move suggests the market landed in roughly the same place: a good business, fully valued, with the next catalyst some distance away.

Street Perspective

Debate: Is "in line" good enough after a 57.6% twelve-month run?

Bull view: The bull case on the Street is that reaffirming every December Capital Markets Day parameter in February is exactly what a management team in year one of a turnaround should do, and that the small consensus miss is noise against a 9% EBITDA increase delivered with no help from prices. On this view the cost program is a call option the market is not paying for, and the copper and aluminium franchises are still under-appreciated relative to the iron ore legacy.

Bear view: The bear camp contends that a stock which has moved 57.6% in twelve months needs new information to hold the multiple, and this release contained none. Guidance was set in December, volumes are flat to down in 2026, capital spend stays at the peak, and the one genuine catalyst of the past six months was removed when the merger talks ended. On this view the equity has already discounted the plan and the next twelve months are execution risk without upside surprise.

Our take: The bears have the better of it on a twelve-month view, and the bulls have the better of it on a three-year view. Nothing in this print damages the structural case. What it does is confirm that the payoff is 2027 and 2028, when Simandou ramps, the lithium projects commission and capital spend steps down to US$10bn, and that 2026 is the year an investor pays for the wait. That is a Hold, not a sell.

Debate: Is the 60% payout a floor or a ceiling?

Bull view: Ten consecutive years at the top of the band, a balance sheet at 0.57x net debt to underlying EBITDA and 18% gearing, and a company that has just told the market its cost base is coming down and its capital spend is coming down. The bull reading is that the payout ratio only goes up from here, and that the US$5bn to US$10bn capital-release program is the funding mechanism for a special distribution or a buyback once it lands.

Bear view: The bear camp points at the cash statement rather than the policy. Cash dividends were 153% of free cash flow in 2025 and would remain above it in 2026 on unchanged operating cash flow at up to US$11bn of capital investment. The 40% end of the band exists for a reason, and the year in which it becomes relevant is a year in which the iron ore price falls without a copper volume offset.

Our take: Both are right about different things. The dividend is not at risk, because the policy is applied to underlying earnings and underlying earnings are stable and supported by a balance sheet with capacity. But a holder buying the 4.2% yield should not extrapolate the 60% as a contractual floor, and should not assume a capital-release windfall arrives on a knowable date. On the numbers as they stand, the payout is a claim on the balance sheet until capital spend steps down in 2028.

Debate: Was walking away from the merger a display of discipline or a failure of ambition?

Bull view: The discipline reading dominates the sell-side commentary. Refusing a transformational transaction on a bottom-up valuation test, after going through the counterparty's full portfolio asset by asset, is the behaviour investors say they want and rarely get. It also frees the balance sheet and management attention for the cost program and the capital release, and it removes the risk of paying up for coal exposure and higher-risk jurisdictions.

Bear view: The skeptical view is that a company whose 2026 volume guidance goes sideways, whose copper growth is guided down and whose longer-duration copper options are all at least a decade from production has just declined the fastest available route to scale in exactly the commodity it says it wants. Rio spent six months and considerable senior attention on a process that produced nothing, while the peer set was compounding.

Our take: The discipline is real and the framing held up under five separate rounds of questioning, which is more than most aborted processes manage. But the practical consequence deserves to be stated plainly: this equity is now a pure organic-execution story with no M&A optionality, and the same valuation discipline that killed the merger is what makes the asset-sale program slow. Investors should price the business on the December plan and nothing else.

Model & Valuation Framework

We do not carry a published financial model on Rio Tinto at initiation. The framework below sets out the assumptions we would build from and the sensitivities that matter most, anchored on the 19 February closing price of $96.34.

Where the equity trades

Valuation impact: On a 6.5x to 7.0x enterprise value to underlying EBITDA range applied to an unchanged FY26 EBITDA of roughly US$25.4bn, and deducting net debt held flat at US$14.4bn, the implied equity value is US$150bn to US$163bn, or $92 to $100 per share. At $96.34 the shares sit almost exactly at the midpoint of that range, which is the arithmetic behind the Hold. The band widens meaningfully in either direction on two variables: a completed asset sale toward the upper end of the US$5bn to US$10bn target would fund de-leveraging worth roughly $3 to $6 per share of enterprise value transfer, while a repeat of 2025's cost delivery applied to the unquantified second phase could add US$1bn or more to EBITDA, worth roughly $4.00 to $4.30 per share at the same multiple. Neither is in our base case today.

What would change the rating. Upgrade triggers: a signed disposal of Iron & Titanium or Borates at a credible price; a quantified multi-year cost target with a bridge; or a de-rating that takes the free cash flow yield above 4% without a corresponding deterioration in the plan. Downgrade triggers: a Simandou schedule slip that pushes the 30-month ramp beyond 2028; capital spend guidance moving above US$11bn; or a copper outcome below the low end of the 800 kt guide.

Thesis Scorecard Post-Earnings

This is our first published view on Rio Tinto, so the scorecard establishes the pillars rather than grading a standing thesis. Each is stated in the form we will carry forward and score every half.

Overall: Thesis established. The structural bull case is stronger than we expected coming into the print, and the near-term cash case is weaker. Those two facts together are what produces a Hold rather than either of the extremes.

Action: Hold. We would add on a de-rating that lifts the free cash flow yield above 4%, or on a signed disposal that starts the capital-release program. We would not chase the shares within a few percent of the 52-week high on a plan that does not pay until 2027.

Bottom Line

Rio Tinto delivered the year its December strategy said it would deliver. Volume and cost did the work, prices contributed nothing on net, and the portfolio absorbed an 8% fall in the realised iron ore price while growing group EBITDA 8.8%. That is a genuine validation of a decade of diversification, and it came with a copper business that more than doubled its EBITDA and a bauxite operation that set a production record.

The same year produced US$4.0bn of free cash flow against US$6.1bn of dividends paid, a net debt position up 162%, and a 2026 guide in which the copper volumes that drove the result step backwards. The merger option that had been the market's swing factor for six months is gone, and the capital-release program that would replace it has no timetable. At $96.34 the shares carry 14.4x underlying earnings and a 2.6% free cash flow yield after a 57.6% twelve-month run.

We like this business and we like the direction of the plan. We are initiating at Hold because the plan pays in 2027 and 2028, and the price already assumes it will.