Aura Minerals Inc.

MSG's Q2 2026 cash cost rose 33% to US$3,852/GEO as output fell to about 7,000 GEO.

• Against an implied realised price of about US$4,450/GEO, MSG cash margin was roughly US$600/GEO, or about US$4m for the quarter. That is before sustaining capex and the heavy development spend.
• Root cause, on the evidence: a largely fixed underground cost base (three mechanised UG mines, a plant, site G&A) spread over very low ounces. Output was low while Aura diverted effort to primary development, infrastructure and the switch from top-down to bottom-up sequencing.
• The lever is ounces per fixed dollar. That means development metres, stope cycle time, UG equipment availability and dilution control, plus disciplined unit-cost management as volumes rebuild.
• Q3 output rose 85% QoQ to about 12,950 GEO (estimate: 7,000 × 1.85). That is still only ~52 koz annualised, against 80 koz in 2024 under AngloGold and the 2027 target of ~80 koz.

Reaching about 80 koz in 2027 from a Q3 run-rate of about 52 koz/yr (estimate) needs roughly +55% more ounces. The underground mines must deliver all of the following at once:

• the top-down to bottom-up sequencing change;
• accelerated primary development;
• longitudinal sublevel stoping at higher productivity.

Typical failure modes in such transitions are development-metre shortfalls, stope-cycle delays (drill, blast, muck, fill), dilution creep in narrow longitudinal stopes, and low UG fleet availability. The specific bottleneck (ventilation, hoisting/haulage, fill, development fleet) is not disclosed, which is itself a diagnostic gap. Management already deferred stronger performance to 2027, so further slippage would hit guidance credibility in the first full year as a US-listed issuer.

Management says MSG processing will rise from about 2 Mtpa to 3 Mtpa by year-end 2026. Back-calculation of what that capacity implies:

• 80 koz/yr at 91.8% recovery needs ~87 koz contained, about 2.8 t Au.
• At 3 Mtpa that is a feed grade of only ~0.93 g/t, versus a UG resource grade of ~3.1–3.4 g/t and a UG cut-off of 1.85 g/t.

So either the plant will run well below capacity on UG ore, or it will depend heavily on low-grade open-pit feed. The low-grade OP recovery assumption is 85.2% versus 91.8% for HG/MG, a 6.6-pt gap. This makes ore blending, stockpile management, P80 control, OEE and recovery management central to MSG economics. There is also a commissioning and ramp-up risk on the expanded circuit at year-end. Stated capacity figures conflict: 1.5 Mtpa on an older operations page versus ~2 Mtpa per management.

On 21 Oct 2026 Aura agreed to buy the contractor fleets, equipment leases and dedicated workforce at three operating mines for R$612m (≈US$111m est.). Moving from contractor to owner-mining transfers the following to Aura:

• equipment availability and maintenance strategy;
• work-order systems, spares and critical-spares inventory;
• workforce supervision and dispatch;
• payload and cycle-time management.

Transitions like this commonly see an availability dip and cost creep in the first 6–12 months if planned maintenance, backlog management and KPI systems are not built in from day one. The value case for insourcing (removing contractor margin) depends on matching or beating contractor productivity. Apoena alone was about 35 koz GEO in FY2025 and is relying on a higher-grade H2 sequence, so any load-and-haul disruption lands on guidance.

9M 2026 production was 253,131 GEO against guidance of 340–390k GEO. Q4 needs:

• 86,869 GEO for the low end;
• 111,869 GEO for the midpoint (+17% vs the record Q3 of 95,557);
• 136,869 GEO for the high end (+43%).

The low end looks achievable on the Q3 run-rate. The midpoint and top end depend on MSG continuing to ramp and on Apoena grades, with the fleet-ownership transition occurring in the same window. Aura also expects cash cost and AISC within guidance, supported by higher MSG production, so an MSG shortfall hits both volume and unit cost.

Q2 capex was US$84.3m. 2026 guidance includes higher sustaining and expansion capex for MSG integration and optimisation, and the R$612m fleet purchase comes on top. Aura paid about US$55m of dividends in Q1. S&P upgraded Aura to BB- but explicitly named sustained high capex, aggressive dividends, weaker operations and working-capital pressure as triggers that could weaken credit metrics. Liquidity is not stressed today. The risk is that an MSG ramp delay plus capex overrun erodes the credit trajectory and the dividend story.

MSG has been under Aura ownership only since December 2025. Aura itself flagged that the MSG and Era Dorada acquisitions required specific focus on integrating personnel and processes into its compliance programme, with site assessments planned during 2026. AngloGold-era operating systems (maintenance, planning, KPI reporting, procurement) are likely being replaced or harmonised during an active turnaround, which raises the risk of data and control gaps (inference).

The company-wide 2025 reserve update showed tonnes +14% (+1.59 Mt) but contained metal -7% (-57,077 GEO) and NSR -9%. That implies an average reserve grade decline of about 18% (0.93 ÷ 1.14), as a rough estimate. Lower-grade reserves raise sensitivity to dilution, grade control, cut-off optimisation and recovery. The research contains conflicting headlines claiming reserves 'more than doubled' to 7.22M GEO, so the like-for-like basis must be verified.

S&P cites extreme-weather exposure and downstream tailings dams as negative E&S factors. Risk factors cover permits, communities, Indigenous peoples, caves, fauna, flora and climate requirements in Brazil. No specific incident, dispute or permit delay at MSG was identified. A plant expansion to 3 Mtpa will, however, increase tailings throughput (inference).

Aura no longer files an MD&A after its TSX delisting. MSG-level tonnes, grade, recovery, unit costs and AISC are not publicly broken out in the material retrieved. Investors and lenders therefore have less visibility into the turnaround, which raises the value of credible, independently verified operating KPIs.