Perseus Mining Increases Nyanzaga GOLD Project ORE Reserves to 4.0 Moz
20 FEBRUARY 2026
NEWS RELEASE
1
PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT ORE
RESERVES TO 4.0 Moz
OVERVIEW
Perseus Mining Limited (ASX/TSX: PRU) is pleased to announce an updated Ore Reserve of 4.0 Moz for the
Nyanzaga Gold Project ( NGP) in Tanzania (Table 1). This represents a 73% increase to the 2.3 Moz Ore
Reserve reported in April 2025 as part of the updated NGP Feasibility Study (FS).
Mine life for the Nyanzaga Gold Project is extended to 16 years, including 14 years of production at greater
than 200 koz/a. An increased Perseus Ore Reserve for Nyanzaga is based on a further cutback to a large scale,
open-pit mining operation as outlined in the NGP FS.
Table 1: Nyanzaga Gold Project Ore Reserves1,2,3,4,5,6
PROJECT
PROVED PROBABLE PROVED AND PROBABLE
QUANTITY GRADE GOLD QUANTITY GRADE GOLD QUANTITY GRADE GOLD
Mt g/t gold ’000 oz Mt g/t gold ’000 oz Mt g/t gold ’000 oz
Tusker - - - 86.1 1.41 3,900 86.1 1.41 3,900
Kilimani - - - 4.8 0.94 146 4.8 0.94 146
Total - - - 90.9 1.38 4,046 90.9 1.38 4,046
Notes for Table 1:
1 Based on November 2025 Mineral Resource estimate
2 Based on February 2026 Ore Reserve estimate
3 Pit designs are based on US$2,300/oz gold metal price
4 Variable gold grade cut-offs for each material type, ranging from 0.3 g/t to 0.4 g/t
5 Inferred Mineral Resource is considered as waste for optimisation purposes
6 Rounding of numbers to appropriate precision may have resulted in apparent inconsistencies
This increase is underpinned by approximately 82,700 m of reverse circulation and diamond drilling across
the Tusker and Kilimani deposit areas completed since Perseus acquired the Project in May 2024. This drilling
was aimed at infill and extension of the mineralisation, and the conversion of Mineral Resources from the
Inferred to Indicated category. In parallel, Perseus has continued to refine the underlying technical
assessments via various studies supporting the NGP Ore Reserve update. The contributions of these activities
to the updated Ore Reserve are presented in Figure 1.
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
2
Figure 1: Nyanzaga Gold Project Ore Reserve Changes – April 2025 to February 2026
Key metrics associated with the NGP include the following:
Production & Cost Metrics (100% basis):
Total gold production over a 16-year is currently estimated to be 3.5 Moz based on a JORC 2012
Probable Ore Reserve of 90.9 Mt @ 1.38 g/t gold for 4.0 Moz.
Gold production exceeds 200,000 ounces of gold per annum from FY28 to FY41.
Applying Perseus’s assumed long-term gold price of US$3,000/oz, the NGP’s investment metrics are updated
to:
Over the life of the mine, the estimated average All-In Site Cost (AISC) is US$1,621/oz.
Undiscounted free cashflow post-tax of US$2,651 million
Net Present Value (NPV10%) of US$864 million post-tax.
Internal Rate of Return (IRR) of 28.8% post-tax
Note: Perseus owns an 80% contributing interest in the NGP together with the Government of Tanzania who owns a 20% non-contributing interest.
This announcement was approved for release by the Managing Director & CEO, Craig Jones.
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
3
PROJECT DETAILS
LOCATION
The NGP is located in north-western Tanzania, south of Lake Victoria within the Sengerema District of the
Mwanza Region (Figure 2). The NGP is approximately 60 km southwest of Mwanza (Tanzania’s second largest
city) and 7 km southeast of Ngoma, with the total area of the Project being 16,085 ha.
Figure 2: NGP location plan
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
4
MINERAL RESOURCES
The NGP Mineral Resource estimate consists of two discrete but adjacent deposits: Tusker and Kilimani.
Following approximately 82,700 m of drilling across both deposit areas , Perseus has prepared a revised
Mineral Resource estimate for the NGP. The combined M&I Mineral Resource for the NGP is estimated at
110.4 Mt grading 1.33 g/t Au, containing 4.7 Moz of gold. A further 6.5 Mt of material grading 1.6 g/t Au,
containing 343 koz of gold is classified as Inferred Mineral Resources (Table 2).
As a result of pit design parameters such as ramp configuration and batter arrangements , there are
approximately 500 oz (0.01%) of ore reported in the Ore Reserve that is not included in the Mineral Resource.
This is not considered material, and no adjustment has been made to the figures reported to account for this.
Table 2: NGP Mineral Resources1,2,3,4,5,6
PROJECT
MEASURED RESOURCES INDICATED RESOURCES MEASURED & INDICATED
RESOURCES INFERRED RESOURCES
QUANTITY GRADE GOLD QUANTITY GRADE GOLD QUANTITY GRADE GOLD QUANTITY GRADE GOLD
Mt g/t
gold
’000
oz Mt g/t
gold
’000
oz Mt g/t
gold
’000
oz Mt g/t
gold
’000
oz
Tusker - - - 100.7 1.37 4,436 100.7 1.37 4,436 5.5 1.8 312
Kilimani - - - 9.7 0.90 279 9.7 0.90 279 1.0 1.0 31
Total - - - 110.4 1.33 4,715 110.4 1.33 4,715 6.5 1.6 343
Notes for Table 2:
1 Based on November 2025 Mineral Resource estimate
2 0.25 g/t gold cut-off applied to in-situ open pit material
3 In-situ open pit resources are constrained to US$2,700/oz pit shells
4 Rounding of numbers to appropriate precision may result in summary inconsistencies
5 Mineral Resources are inclusive of Ore Reserves with the exception of approx. 500 oz reported as Ore Reserves but excluded from Mineral Resources
6 Mineral Resources are reported on a 100% basis
MINERAL RESOURCE MATERIAL INFORMATION
Mineral Tenement and Land Tenure Status
The NGP consists of one Special Mining Lease (SML), six granted prospecting licences, and one prospecting
licence application (Table 3). Through its subsidiary, Nyanzaga Mining Company Limited (NMCL), Perseus
holds an 80% interest in Sotta Mining Corporation (SMCL). The SML was granted on 13 December 2021 to
SMCL for an initial period of 15 years, with extensions permissible under application before the final year.
The Treasury Registrar holds the 20% free carried interest of the Government of Tanzania in accordance with
the Mining Act.
Table 3: NGP tenement summary
Geology
The NGP is situated on the north-eastern flank of the Sukumaland Greenstone Belt. Rock sequences comprise
Nyanzian-aged sequences of mafic volcanics in the far south -west and banded- iron-formation, tuffs,
mudstone, sandstone and epiclastics of general volcanogenic origin and massive sulfide lenses and volcanics
of rhyodacitic to andesitic composition. The sequences are tightly folded and strike generally west -
northwest.
TENEMENT HOLDER STATUS APPLICATION DATE GRANT DATE EXPIRY DATE AREA (km2)
SML 653/2021 SMCL (100%) Active 10/10/2017 13/12/2021 12/12/2036 23.36
PL 11873/2022 SMCL (100%) Active 1/02/2022 29/03/2022 28/03/2026 17.03
PL 11874/2022 SMCL (100%) Active 1/02/2022 29/03/2022 28/03/2026 21.22
PL 12428/2023 SMCL (100%) Active 6/07/2023 24/07/2023 23/07/2027 42.78
PL 12429/2023 SMCL (100%) Active 6/07/2023 24/07/2023 23/07/2027 4.20
PL 12430/2023 SMCL (100%) Active 6/07/2023 24/07/2023 23/07/2027 1.37
PL 11186/2018 Perseus Tanzania Limited (100%) Active 14/12/2016 26/10/2018 25/10/2027 18.21
PL 11961/2017 Perseus Tanzania Limited (100%) Application 31/05/2017 - - 3.53
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
5
In the central part of the NGP, the Nyanzaga and Kilimani groups are defined. The Nyanzaga Group is divided
into a sequence of three broad mappable units from drill core and outcrop mapping, and are in stratigraphical
order as follows:
1. Nyanzaga Upper Volcaniclastic Formation.
2. Nyanzaga Central Formation.
3. Nyanzaga Lower Volcaniclastic Formation.
These three units are folded into the north-northwest plunging Nyanzaga Anticline. These sequences are in
turn overlain by the Kilimani Group which appears to be tectonically dislocated from the underlying Nyanzaga
Group by faulting and potential earlier thrusts.
A series of different faulting sets has been recognised across the Nyanzaga project area. These include
observed outcrop positions with suggested sinistral movement; others with inferred dextral movement; a
series of steep faults sub-parallel to bedding; and a number of low angle thrust faults. Additionally, east-west
trending normal faults with minor displacement have been identified.
The interaction of these various fault orientations results in the definition of numerous fault block domains.
The geological model has focussed on those faults where evidence is observed across multiple sections, and
therefore likely to have influence on the geological model at a resource scale. Small scale and discontinuous
features are expected to be present but unlikely to materially influence the position of mineralisation within
the project.
The two mineralisation systems show similar alteration and mineralisation styles, though the Kilimani deposit
is interpreted to have been emplaced at shallower levels than at Tusker. Typical alteration features include
pervasive carbonate alteration, varying from distal ferroan-calcite to proximal ankerite-sericite dominated
and quartz-pyrite alteration. The current interpretation indicates that there is a distinct lithological control
and most of the gold appears to be hosted by iron rich sediments in the form of an extensive stock-work of
carbonate, quartz veins, and quartz-carbonate breccias. Preferential grade enhancement occurs in selected
altered units such as the thick cherts, silica-dolomite altered medium grained sandstones, brecciated silica-
carbonate altered mudstones, or in the late quartz veins as free gold.
As a result of the complex interactions between lithology and faults, definition of discrete mineralisation
envelopes is difficult, and attempts to do so are likely to overstate mineralisation continuity, particularly for
zones of higher grades.
To mitigate this risk, and to better reflect the anticipated large scale open pit mining method, the
mineralisation interpretation instead aimed to define the limits of the mineralisation within the deposit, with
the distribution of high-grade mineralisation within this shape reflected in the estimation parameters. A
mineralised envelope was defined using an economic compositing routine within Leapfrog using a grade
threshold of 0.3 g/t Au and a minimum downhole length of five metres. This length was selected as it was
broadly aligned with the expected bench height for the project, while also generating consistent shapes with
limited individual internal waste zones which were unlikely to be realistically excluded from the overall
domain.
An intrusion model was created in Leapfrog using the economic compositing outputs as a base. To reflect the
interpreted mineralisation associations with lithology trends and key faults, surfaces representing these
features were used to define a structural trend and guided the intrusion modelling. The resulting
mineralisation shape is presented in Figure 3 and forms the basis of the domain coding for the Tusker deposit
area.
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
6
Figure 3: Cross-section 9,672,450mN (±10 m) looking north showing Tusker mineralisation envelope against economic
compositing criteria
Drilling Techniques and Hole Spacing
Drilling completed at the NGP and used to support the Mineral Resource estimate includes 759 reverse
circulation holes for 96,716 m and 462 diamond holes (as tails on RC pre-collars or drilled from surface) for
188,643 m. RC drilling prior to 2010 used 6” diameter face-sampling bit. After 2010, RC drilling used a 5¼”
diameter face-sampling bit. Diamond drilling utilised PQ (85 mm diameter) or HQ triple-tube (61.1 mm dia.)
drilling in weathered materials and NQ2 (50.6 mm dia.) or NQ (47.6 mm dia.) core in fresh rock. Pre-collared
holes were normally drilled to NQ or NQ2 diameter from the commencement of coring.
The Tusker mineralisation was drilled to a nominal 20 mN × 40 mE, with selected infill to 20 mN × 20 mE.
Holes were typically drilled at a dip of - 60° towards 270°, with a minor number of holes drilled towards a
bearing of 090° to better define the western limb of the fold structure. Drilling at Kilimani was predominantly
via RC with minor diamond core at a nominal 20 m × 20 m pattern. Holes were aligned to either 035° or 215°
with inclinations nominally -60°. A plan showing the historical drill collars and those drilled by Perseus to
support this update is presented in Figure 4.
Perseus Mining Limited
Projection: Arc1960 UTM Zone 36S
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
7
Figure 4: Plan view showing existing drilling across the NGP
Sampling
For RC samples prior to 2005, samples were normally combined into 3 m composite samples for assaying.
Where composite samples returned gold assays greater than a nominal threshold, second splits were
generated for the constituent one metre samples and those were submitted for assay. The one metre assays
are prioritised over the original composite assays in the acQuire database. RC drilling since 2005 has been
sampled at one metre intervals for the entire length of the hole. Depending on the drill rig, the sample was
split at the cyclone, or the bulk sample was passed through a standalone riffle splitter at the drill site to
generate the subsample for analysis with a nominal weight of 3-4 kg.
Drill core was collected from a core barrel and placed in appropriately marked core trays. Down hole core
run depths were measured and marked with core blocks. Core was measured for core loss and core
photography and geological logging completed. Core samples are typically quarter core for PQ and half core
for HQ and NQ sizes, with the orientation line preserved in the tray. The entire hole is sampled with samples
collected nominally at 1 m or at geological contacts.
NEWS RELEASE | PERSEUS MINING INCREASES NYANZAGA GOLD PROJECT
ORE RESERVES TO 4.0 Moz
8
Certified reference materials, analytical blanks, and field duplicates were used as part of the QAQC
procedures and were each inserted at a rate of 1:25 samples. Umpire analysis has been completed on a range
of mineralised intervals.
Sample Analysis
For RC and diamond core samples, the entire sample was dried for 24 hours at 95°C ± 5°C. The sample was
then crushed in a jaw crusher to 85%, -2 mm, and riffle split to produce an 800 g to 1 kg split for pulverisation
and analysis. The sample was pulverised in a LM2 bowl (1 kg capacity) to 90% passing 75 μm.
Prior to 2024, analysis utilised fire assay techniques with a 50 g charge for the gold analysis. Lead buttons
produced after fusions are cupelled and then digested via aqua regia. The digest is analysed for gold using
atomic adsorption spectroscopy ( AAS). The value is recorded electronically to the laboratory database,
exported in CSV format and emailed for upload. Fire assay is considered a total assay method for gold.
From 2024 onwards, gold analyses have been attained via the photon assay determination method. Samples
are crushed in a jaw crusher to 85%, -2 mm, and riffle split to produce a nominal 500 g split into the sample
pots, ensuring fill levels as close to maximum as possible. This method is considered a measure of the total
gold content.
Estimation Methodology
Mineral Resources reported herein have been estimated using a geostatistical block modelling approach,
informed from gold assay data obtained from predominantly core and RC drill chip samples. The geological
wireframes consider structural and lithological interpretations of the gold mineralisation.
The NGP Mineral Resource estimate used Seequent Leapfrog Geo software for the geological and
mineralisation interpretation and associated drill hole coding, while Maptek Vulcan software was used for
block model creation, editing , grade estimation and reporting. The grade interpolation of the Tusker
component of the resource model utilised local uniform conditioning (LUC) while the Kilimani component
employed ordinary kriging. Dynamic anisotropy was used to guide the search orientations along the
mineralised trends. Bulk density was estimated for the project based on measurements of diamond core
using water immersion methods.
Sample data was composited to one-metre downhole lengths using a best fit-method. No residuals were
generated. Statistical analysis was carried out on mineralisation domains, with hard boundary techniques
employed within each estimation domain.
Analysis of the composite data indicated the presence of outlier values indicating grade capping was
required. Capped values were generally selected above the 99 th percentile. Distance restrictions on high
grades were also employed as part of the Tusker LUC estimate. Estimates were prepared for gold and density.
Interpolation parameters were derived using standard exploratory data analysis techniques of statistical and
continuity analysis. Appropriate interpolation strategies were developed on a domain basis using kriging
neighbourhood analysis (KNA) with a minimum number of 6 composites and a maximum of 16 for Tusker,
with 6 and 14 used for Kilimani. Octant restrictions were employed for each deposit. Tusker was estimated
in a single pass with a maximum search radius of 65 m, while Kilimani employed a two-pass strategy, with
the initial 125 m search distance increased by 50% for the second pass. An example cross section showing
the estimated blocks for the Tusker mineralisation is presented in Figure 5.