Mining in Australia
Industry Landscape and Outlook

Rio's SP10 at 58% Fe went to seven cargoes in Q4 2024, up from one or two. SP10 is the reject fraction from the same pits as PB Fines and you only sell seven cargoes of it when you cannot blend enough decent material to fill your flagship orders.

Hamersley Basin BIF started at about 30% Fe. Magnetite and hematite interbanded with chert. Meteoric water dissolved the silica over time and left a collapsed porous mass at 55-65% Fe in the top 100-200 metres. The enrichment was patchy because it followed fracture networks and permeability breaks within the BIF. Where the water found pathways, enrichment went deep. Where it did not, you get pods of 45-55% material sitting in the middle of what the geological model says should be 60%+ ore.

The first generation of Pilbara mining took the obvious stuff. Surface mapping and shallow RC drilling identified the fattest enrichment zones and the mines went in there. Forty years later the rich bits are substantially drawn down and what is left is more variable, harder to predict, and harder to blend into a product that meets spec. Grade control drill spacing has to tighten as enrichment becomes less predictable, which costs money per tonne that did not need to be spent when the orebody was more uniform. The blending problem is real: you are trying to combine parcels of ore from different faces and benches to hit a target Fe and keep alumina and silica within contract penalties, and the variance in the input feed is increasing. PB Fines at 60.8% is the aggregate outcome of that blending getting harder.

Rhodes Ridge is the big replacement deposit. Rio and Mitsui, 6.8 billion tonnes, 5.3 billion at 62.2% Fe. Rio also put AUD 2.8 billion into Brockman Syncline 1. Committed iron ore projects AUD 9.6 billion in 2025, up from AUD 6.3 billion per the Department of Industry data. BHP's Yandi ramping down, MinRes put Koolyanobbing into C&M in early 2025.

But Rhodes Ridge has native title overlay and the WA heritage framework has been dysfunctional since the 2023 Act was pulled. Getting a major greenfield Pilbara mine from FID to first ore takes five to seven years minimum. Rail spurs needed for deposits outside the existing network, and the existing network was built by and for the three incumbents. Rio's feasibility study mentions leveraging existing infrastructure. In the Pilbara that means negotiating track access with whoever owns the railway, which is usually yourself but not always, and port allocation at facilities running close to capacity.

This all connects and I am going to run it together rather than pretend these are separate topics.

DRI needs 67%+ Fe ore with low gangue. Pilbara hematite-goethite averages 60% and the impurity profile is getting worse. Rio acknowledged the problem in its climate report. HYFOR, the Fortescue-Primetals pilot in Linz, uses hydrogen in a fluidised bed to reduce fine ore without pelletising. The pitch is that it can handle Pilbara fines. The reality is that goethite-rich Pilbara ore loses water when heated and the loss on ignition changes gas dynamics inside the reactor in ways that make it behave very differently from the magnetite concentrates that commercial DRI plants were built to process. Iron Bridge, Fortescue's magnetite project meant to supply the high-grade feedstock for exactly this application, has had serious commissioning trouble. HYFOR at Pilbara scale is years away at best.

Rio started blending Pilbara ore with Simandou ore at Chinese ports from 2026. Simandou at 65-67% Fe, first ore November 2025, RBC estimating 7% of seaborne supply by 2028 at ramp. The blended product holds spec. The unblended Pilbara product increasingly does not.

CMRG fits here because Simandou gives CMRG a second lever. The first lever is consolidated buying power: 20 billion yuan, State Council backing, over 50% of Chinese steel mill procurement since 2022. Per Reuters, September 2025 CMRG blacklisted BHP's Jimblebar Fines, then a second product. BHP wanted dollar settlement plus 15%. CMRG said no. Got a USD 1/tonne freight discount from Rio on certain vessels. The PM commented.

The second lever is quality. Simandou ore is DRI-compatible. Pilbara ore is increasingly not. As green steel routes scale, CMRG can redirect volume toward Simandou for the mills investing in DRI-EAF and use Pilbara ore for legacy blast furnace demand, which is the lower-margin segment. Both levers get stronger as Simandou ramps and Pilbara grades decline.

AUD 10.2 billion in 2024-25 per the Australia Institute's ATO analysis. Mining roughly half. BHP about AUD 600 million, Rio about AUD 400 million. Cumulative claims over two decades exceed AUD 60 billion, though that number relies on ATO data with sectoral classification ambiguities that the ATO itself flags, so the precision some commentators attach to it is unwarranted.

FTCS refunds 48.8 cents per liter on off-road diesel. OECD and IEA call it a fossil fuel subsidy. 95% of excise goes to general revenue, no road fund. Fortescue's CEO Otranto wants a AUD 50 million cap. Matt Kean called it “insane.” King rejected change.

The MRRT campaign in 2014 was about AUD 22 million in six weeks of TV and it helped bring down a PM. Different fiscal environment now. No surplus. Cost-of-living politics. Greens and teals in positions they did not hold.

BHP wrote Nickel West to zero. Kwinana, Kalgoorlie, Mt Keith, Leinster gone. Over 150,000 tonnes national production to around 60,000 between late 2023 and mid-2024. Indonesia banned ore exports in 2020, attracted Chinese smelting capital, expanded to 1.58 million tonnes in 2022 per INSG. Prices USD 50,000 at the 2022 squeeze to USD 16,400.

Industry wanted a green premium. Nobody paid one. Indonesian RKEF nickel from coal-fired furnaces with documented damage in Sulawesi was cheaper per tonne. The critical minerals strategy assumes buyers will pay for supply chain security. Nickel says they will not.

Queensland met coal: thermal by-product revenue from CHPP absorbs fixed costs. Thermal weakens, effective met coal cost rises. Progressive royalty hit 40%+ at 2022 peaks, investment slowed.

MFP down 3.3% in 2024-25 per ABS. Fifth straight decline. 130-plus autonomous trucks at Rio. Grades falling means more rock per tonne, autonomous trucks do not change that.

Rio cut Yarwun alumina 40% from October 2026 on energy costs.

80% toward 60% across 800 million tonnes per year is tens of billions less cash. That cash funded exploration, processing, fleet replacement across the sector. Geology, CMRG, and the diesel rebate are all moving against Pilbara profitability on different timescales and none of them look like reversing.