This is investment research, not personal financial advice.

First concentrate is not steady-state production

Core Lithium (ASX:CXO) rose 6.06% to A$0.35 on 8 September after its Finniss plant produced the first spodumene concentrate of the restart. The gain added about A$64.7 million to equity value and left the company worth A$1.133 billion at the close. The announcement arrived within six months of the March restart decision and inside management's September-quarter timetable (Core first concentrate 2026; ASX close).

That is a clean schedule win. It is also a narrow one.

Core did not report tonnes, recovery, throughput, product grade, impurities, plant availability or unit cost. Commissioning and optimisation continue, mill clutches are due to be replaced during a planned campaign, and work on the crushing circuits is still progressing. First shipment of newly produced concentrate remains scheduled for the December quarter. The market therefore rewarded proof that the upgraded plant can make concentrate, not proof that Finniss can make it repeatedly at a profitable cost.

The distinction matters because this is Finniss's second attempt at commercial production. The first run began in 2022, generated only one profitable statutory year, consumed repeated equity injections and ended with mining suspended in January 2024. Core later impaired A$119.6 million of Finniss assets. Today's response looks roughly proportionate to the milestone itself, but the post-move valuation already asks for more than a successful switch-on. It asks for a material part of the restart plan to work.

The rebuilt operation rests on one underground mine

Finniss sits about 90 kilometres south of Darwin. Its operating chain combines open-pit and underground mines, a dense-media-separation processing plant, road haulage and Darwin port. Existing infrastructure is the reason Core could move from a March final investment decision to concentrate in September. It had already tested the logistics with 20,000 tonnes of lithium fines and 5,000 tonnes of legacy concentrate shipped during the June quarter (Core Q4 2026).

The near-term feed comes from the Grants open pit. Core expects Grants to deliver about 784,000 tonnes of ore and 134,000 tonnes of SC5 product under an 11-month mining contract. That is a bridge, not the long-life mine. BP33 is supposed to provide first development ore in mid-2027 and reach steady state around mid-2028. It accounts for A$184 million of the A$208 million restart budget and 88% of ore feed in the first decade (Core FID 2026).

The production design is substantial for a company that had stopped generating revenue. At full rate the plant is meant to process 1.2 million tonnes a year and produce up to 214,000 tonnes of SC6-equivalent concentrate. The 20-year schedule contains 15.6 million tonnes of reserves at 1.27% Li2O and 17.9 million tonnes of plant feed. About 13% of scheduled material is inferred rather than reserve-classified (Core FID 2026; Core Reserve 2025).

Those numbers need two qualifications. First, 2.87 million tonnes of planned life-of-mine concentrate spread over 20 years averages 143,500 tonnes a year, well below the 214,000-tonne nameplate figure. Ramp-up, lower-rate years and later deposits fill the difference. Second, 15.6 million reserve tonnes represent roughly 13 years of feed at 1.2 million tonnes a year. The 20-year label cannot be read as 20 uninterrupted years at nameplate. Carlton, inferred material and future resource conversion become more important later in the schedule.

Blackbeard may add optionality, but it was still an exploration target of 7-10 million tonnes at 1.5-1.7% Li2O in the June material. It is not yet a resource and does not support the current production schedule. Core's business remains unusually concentrated: Grants must bridge the restart, then BP33 must carry the first decade.

The accounts remember the first restart

The financial record is less smooth than the six-month commissioning timeline. Figures below are filed values in A$ million, except production, cost and share count. FY2026 is the June quarterly cash-flow record rather than audited annual accounts. Capex is payments for plant, equipment and mine development; it excludes capitalised exploration. FY2023 operating cash flow includes a large customer prepayment and should not be treated as recurring owner cash.

Year Revenue NPAT Operating cash flow Plant/mine capex Cash Concentrate production Reported cash cost Shares on issue
FY2022 0.0 (7.5) (6.2) 51.4 135.2 0t n/a 1,732.6m
FY2023 50.6 10.8 90.8 150.6 152.8 18,274dmt A$1,155/t 1,858.5m
FY2024 189.5 (207.0) (77.9) 60.6 87.6 95,020dmt A$1,396/t 2,136.9m
FY2025 0.0 (23.4) (43.9) 11.8 23.5 0t n/a 2,143.0m
FY2026 quarterly record n/a n/a (21.1) not directly reported as one line 181.8 0t new production n/a 3,236.8m

Sources: Core 2022; Core 2023; Core 2024; Core 2025; Core Q4 2026. Parentheses denote outflows or losses. FY2024 revenue is the A$189.5 million net figure after A$50.5 million of quotation-period adjustments.

Core moved from construction in FY2022 to 18,274 dry tonnes of concentrate in FY2023, then to 95,020 dry tonnes in FY2024. Volume rose, yet cash cost deteriorated from A$1,155 to A$1,396 a tonne as realised concentrate prices fell to US$1,574 a tonne. The statutory FY2024 loss included the A$119.6 million impairment, a A$25.3 million onerous-contract charge and derecognition of A$65.3 million of deferred tax assets (Core 2024).

The FY2023 profit also overstates operating quality. A US$61.4 million Yahua prepayment inflated receipts; Core later delivered product and repaid US$18.6 million after provisional pricing moved against it. FY2024's accounts were subsequently re-presented to separate A$240.0 million of gross customer revenue from the A$50.5 million pricing adjustment. That history explains why tonnes, grade, realised price and cash conversion matter more than a photograph of first concentrate.

Dilution tells the same story from another angle. Shares on issue rose 87% from 1.733 billion at June 2022 to 3.237 billion at June 2026. Gross placements, purchase plans and option exercises raised roughly A$519 million across FY2022-FY2025, before the A$120 million restart equity raise completed in FY2026. Core has built and rebuilt Finniss with outside capital rather than retained owner cash.

Owner cash has not yet appeared

A mining developer can report positive operating cash flow before it has created distributable cash. Core's bridge is a good example. FY2023 NPAT of A$10.8 million became A$90.8 million of operating cash flow mainly because the Yahua prepayment created an A$83.6 million financial liability. Against that inflow, plant and mine capex was A$150.6 million and capitalised exploration was another A$15.0 million. Owner cash remained negative before the next expansion dollar was considered (Core 2023).

FY2024 was more direct. Operating cash outflow was A$77.9 million, plant and mine capex was A$60.6 million, and exploration absorbed A$26.6 million. The combined cash use was about A$165.2 million before financing and minor items. In FY2025, care and maintenance still consumed A$43.9 million of operating cash, with another A$19.4 million spent across mine assets and exploration (Core 2024; Core 2025).

The FY2026 quarterly record shows operating cash outflow of A$21.1 million and investing cash outflow of A$21.9 million. Equity and the first US$26 million convertible-note tranche supplied A$201.1 million of financing cash. Cash consequently rose to A$181.8 million. That is financing capacity, not owner earnings (Core Q4 2026).

For Finniss, the useful owner-cash test begins after commercial ramp-up:

receipts from saleable concentrate - mining - processing - transport - royalties - site overhead - sustaining capital - corporate costs - cash interest - tax

Core's A$762-per-tonne design cost is FOB, excludes royalties and applies around BP33 steady state. The commodity price assumption is US$1,500 per tonne CIF. Freight, marketing, royalties, corporate overhead, financing and tax sit between those two quoted numbers. Until actual product and cost disclosures fill that bridge, an EBITDA margin or project NPV cannot be read as cash available to equity holders.

Funding removed one risk and introduced three claims

The restart plan uses A$208 million for mine and plant work, A$69 million for working capital, corporate purposes and transaction costs, and A$30 million for growth and exploration. Core's March sources-and-uses table combined A$120 million of new equity, US$70 million of convertible notes, US$25 million of initial senior debt, A$41 million of forecast cash and A$11 million from legacy stockpiles. A second US$25 million senior tranche sat outside the core A$307 million sources (Core FID 2026).

At 30 June, Core reported A$181.8 million cash, A$193 million of restart capital remaining and A$320 million of cash, committed receipts and undrawn facilities. That produced A$127 million of stated headroom, excluding Grants cash flow. Cash alone was A$11 million below the remaining restart capital. In August the company received the deferred US$44 million, or about A$62 million, second convertible-note tranche after foreign-investment approval (Core Q4 2026; Core note funds 2026).

The funding package is adequate on management's budget. Its claims are not free. The US$70 million notes mature in six years, convert at A$0.252 a share and carry a 10% coupon before BP33 reaches steady state, then 8% before possible sustainability adjustments. The US$50 million senior facility matures in four years, costs at least 8.5% because of the SOFR floor, and begins monthly principal repayments 18 months after closing. Drawing it adds cash and an equal liability; it does not add project value (Core FID 2026).

Full note conversion at the FID exchange assumption would add roughly 390 million shares while removing the note claim. That is about 12% of the June 2026 share count. If commissioning or BP33 takes longer, three pressures arrive together: the higher note coupon persists, senior amortisation approaches, and Grants has less time to finance the transition.

Management did cut one structural cost. Core bought the crushing plant in March 2025 for A$19.5 million to end the remaining services contract. The accounting included a A$12.8 million gain from derecognising part of an onerous provision. Owning the crusher should reduce contractor margin and improve control, but the acquired circuit now has to demonstrate availability under Core's operating model (Core 2025).

Recovery is the operating moat test

Finniss does have defensible assets. It is close to Darwin, holds an established approvals and logistics chain, and has a high-grade resource base near existing plant. Those advantages compressed restart time. They do not guarantee low unit costs.

The previous plant reached about 65% recovery. The FID assumes 78% global recovery once the Reflux Classifier circuit is operating, together with better ore hygiene, lower dilution and improved liberation. Throughput is planned to rise about 20% to 1.2 million tonnes a year. First concentrate confirms that material moved through part of the circuit. It says nothing yet about whether that 13 percentage-point recovery gain is repeatable (Core funding 2026; Core first concentrate 2026).

The peer comparison sets a demanding benchmark. PLS Group produced 879,500 tonnes and sold 891,600 tonnes in FY2026 at A$569 per tonne FOB unit cost. Its June-quarter cost was A$616 per tonne, with a realised SC6-equivalent price of US$2,415 per tonne CIF. Core's A$762 design cost is 34% above PLS's full-year outcome and still excludes royalties (PLS 2026).

PLS is a much larger open-pit producer, so cost parity is not a sensible expectation for a smaller underground operation. The comparison instead shows that Finniss has no demonstrated cost moat. Its advantage must come from grade, compact infrastructure, disciplined recovery and low incremental capital. If recovery stalls near 65%, or underground dilution exceeds plan, fixed costs are spread across fewer payable tonnes and the A$762 figure recedes quickly.

The Australian Government's June outlook adds a second warning. It estimated 2026 spodumene prices at US$2,236 a tonne after a sharp rally, then forecast US$1,575 in 2027 and about US$1,300 from 2028. Demand was expected to grow around 11.5% a year through 2031, but supply near 10.0% kept the long-run price below today's spike (Australian Government 2026).

Core's US$1,500 FID assumption sits below the 2026 estimate and near the government's 2027 forecast, but above the US$1,300 longer-run projection. The restart does not require the 2026 spot price to persist. It does require the margin to survive normalisation.

A$0.35 prices more than the FID project value

A project-NPV bridge fits Finniss better than an earnings multiple because there is no current earnings base to capitalise. Core's FID reported a post-tax, unlevered NPV8 of A$837 million at US$1,500 per tonne SC6 CIF, A$0.70 per US dollar, 78% recovery, A$762 per tonne FOB cost before royalties and a 20-year schedule. It reported an unlevered post-tax internal rate of return of 76.5% and A$1.7 billion of undiscounted life-of-mine free cash flow (Core FID 2026).

At A$0.35 and 3.237 billion shares, the market capitalisation is A$1,132.881 million. Start with June cash of A$181.8 million and deduct the US$70 million notes at roughly A$99 million. Then allow A$20-A$70 million for subsequent restart spending, corporate overhead, fees, interest and working-capital needs before Grants receipts become dependable. That gives an illustrative net balance-sheet bridge of A$10-A$60 million. The undrawn senior facility is excluded because cash and debt offset on draw.

On that basis the share price implies roughly A$1.07-A$1.12 billion for Finniss and residual optionality, 28%-34% above the company's A$837 million FID NPV. This is not a direct like-for-like premium: the market may assign value to Blackbeard, other Finniss resources or a commodity price above US$1,500. It may also be overlooking financing costs and later dilution that the unlevered NPV does not deduct.

A useful author sensitivity starts with 2.87 million tonnes of life-of-mine concentrate. Applying the schedule, tax and discounting gives an estimated 1.5-1.7 million discounted payable tonnes. Each US$100-per-tonne change in long-run price is therefore worth roughly A$150-A$170 million of post-tax project NPV at the FID exchange rate. This estimate becomes less reliable near break-even because tax losses and operating responses are not linear.

The reverse bridge says A$0.35 is consistent with a long-run SC6 price around US$1,640-US$1,680 if Core otherwise delivers the FID ramp, cost and capital plan. Another route is a lower operating cost, near A$570-A$620 per tonne at US$1,500, or output materially above 214,000 tonnes without equivalent extra capital. The cost route would put Finniss near PLS's FY2026 outcome despite much smaller scale. The price route sits above the government's longer-run US$1,300 projection. Resource conversion can close some of the gap, but only after drilling and studies turn optionality into an economic schedule.

Four paths from commissioning to value

These ranges are author scenarios in Australian dollars per share on 3.237 billion current shares. They are not forecasts. The bridge starts with a risked Finniss project value, adds a conservative net balance-sheet amount, deducts financing and corporate leakage, then divides by issued shares. Conversion dilution is reflected more heavily in weaker cases.

Case Long-run price and operating state Equity-value range Value per current share
Severe downside US$900-US$1,100/t; recovery near 65%; output below 150ktpa; cost above A$950/t; additional capital A$65-A$260m A$0.02-A$0.08
Bear US$1,250-US$1,350/t; BP33 slips 12 months; 170-190ktpa; A$850-A$950/t cost A$390-A$710m A$0.12-A$0.22
Base US$1,450-US$1,550/t; mid-2028 steady state; near 214ktpa; A$730-A$800/t cost A$840-A$1,100m A$0.26-A$0.34
Bull US$1,900-US$2,100/t; 214-240ktpa; A$650-A$750/t cost; reserve conversion adds productive years A$1,520-A$2,070m A$0.47-A$0.64

The current A$0.35 sits just above the base range. The close therefore assumes more than a Finniss restart: either performance beyond the FID reference case or value from commodity price and resources that the current plan has not yet secured.

Price and recovery dominate the sensitivity. Holding the rest of the FID design broadly constant produces this approximate matrix:

Long-run SC6 CIF 65% recovery 72% recovery 78% recovery
US$1,200/t A$0.05-A$0.10 A$0.11-A$0.17 A$0.16-A$0.22
US$1,500/t A$0.15-A$0.21 A$0.23-A$0.29 A$0.27-A$0.33
US$1,900/t A$0.31-A$0.40 A$0.42-A$0.52 A$0.50-A$0.61

Author estimates, not company figures. The ranges vary output and unit cost with recovery and include financing/corporate deductions. They are directional because Core has not published annual project cash flows or a numeric NPV sensitivity table.

The plant can produce more value than this cautious bridge allows. The plant is upgraded, the crusher is owned, lithium prices are currently above the FID assumption, and Grants can generate cash before BP33 reaches ore. If those factors combine, first concentrate may become steady production quickly, the FID NPV may understate near-term cash, and the resource base may support more than the scheduled output.

The opposing case carries more historical evidence. The prior operation needed roughly 18 months to build recovery and output before suspension. The new plan concentrates capital and first-decade feed in BP33, while financing costs continue through ramp-up. A photograph confirms neither repeatability nor margin.

Three reporting dates decide the reaction

December 2026 is the first test. A shipment of newly produced concentrate would show that commissioning output passed product qualification and moved through the full logistics chain. The useful disclosure is not the shipment headline alone. It is tonnes, grade, moisture, impurities, realised price, recovery, plant availability and cash receipt.

Mid-2027 tests the bridge. Grants has a finite pit and contract life. BP33 first development ore must arrive before stockpiles and Grants production stop carrying plant utilisation. Decline metres, underground development spend and remaining capital should show whether the A$184 million allocation still holds.

Mid-2028 tests the economic claim. Core's design reaches 1.2 million tonnes a year, 214,000 tonnes of SC6-equivalent output, 78% recovery and A$762 per tonne FOB cost around steady state. A cost above A$850 after ramp-up, recovery stuck near 65%, or funding headroom below A$50 million before BP33 ore would each weaken the bridge behind today's valuation.

Operational evidence can move the ranges upward. Blackbeard assays and a maiden resource can extend the reserve clock. Consistent recovery above 72% would show the plant moving away from its first-run history. Customer acceptance of SC5 product would turn the September milestone into revenue rather than inventory.

Source notes: confidence and missing information

The primary trigger, four annual reports, June quarterly, restart decision and presentation, reserve filing, August funding receipt, government outlook, peer quarterly and independent ABC report were fetched and read. Verification is marked partial because the Finance API sidecar returned a stale A$0.38 market snapshot dated 21 August and no financial-facts series; the government PDF also returned an error to the automated source-liveness check after it had been downloaded successfully during research. The ASX company page and 8 September close were used for price, move, issued shares and the equity-value calculation instead.

Core has not published actual restart throughput, recovery, grade, availability or unit cost. It has not provided annual project cash flows, a numeric NPV sensitivity table, a current cash balance after August funding and subsequent restart spend, or the amount drawn under the senior facility after June. Those gaps explain why the valuation ranges are wide.

First concentrate removed the risk that the recommissioned plant could not produce any concentrate on the planned timetable. The 6.06% reaction reflects that progress. At A$0.35, though, the market is already pricing a project value above the FID NPV. The next disclosure has to move from a milestone photograph to a repeatable operating bridge: recovery, payable tonnes, cash cost and a shipment that funds BP33.

References