SINGAPORE, Nov. 17, 2026 (GLOBE NEWSWIRE) -- Verde AgriTech Ltd. (TSX: NPK | OTCQX: VNPKF) ("Verde" or the "Company") reports SGS bench-scale leach results from Minas Americas Global Alliance ("Minas Americas"), Brazil, held through Magnes Rare Earths ("Magnes"), Verde's subsidiary. Expressed as oxide equivalents, the SGS results indicate up to 982 grams of magnet rare earth oxides (MREO) per tonne of prepared minus-2 mm laboratory feed, combining primary solution and both washes. MREO comprises neodymium, praseodymium, dysprosium and terbium oxides.
SGS tested seven selected geological samples from four auger holes, plus a quality-control (QC) repeat. The findings add processing evidence to the previously reported rare-earth intersections and guide two planned workstreams: integrated testing to produce a characterized concentrate, and wider bench testing to understand differences across the project.
Highlights
- 982 and 892 g/t estimated MREO in the two highest-result selected samples, combining Nd, Pr, Dy and Tb oxide equivalents.
- 30-minute main leaches at ambient temperature, using ammonium sulphate at nominal pH 4.2 with mechanical agitation, followed by two separately assayed washes.
- 409 and 355 g/t estimated samarium, europium and gadolinium (SEG) oxides plus heavy rare earth oxides, including yttrium, in those samples. This overlapping measure is not additive to MREO.
- Uranium, cadmium and chromium below detection in all primary leach and wash solutions; iron below detection in 23 of 24 streams.
- Integrated concentrate production and broader bench testing are planned to support preliminary processor evaluation, future mining priorities and area-specific flowsheet optimization.
"Our ambition is to make Magnes a significant Brazilian source of magnet rare earths," said Cristiano Veloso, Founder and CEO of Magnes. "These SGS results give us a concrete reason to advance, with selected material showing an encouraging leach response at ambient temperature. Now we are focused on the next step that matters commercially: producing a concentrate that processors can evaluate. The ANSTO programme and broader SGS testing will help us determine how much we can recover through the full process and where to focus development across Minas Americas."
What the SGS results show
SGS tracked rare earths through the prepared feed, primary leach solution, both washes and residue, recording reagent use, collected volumes and filtration behaviour. The section "SGS test conditions and reporting basis" explains how the Company calculated the oxide-equivalent estimates.
The 982 and 892 g/t MREO estimates comprise 941 and 856 g/t of neodymium-praseodymium (NdPr) oxides, plus 41 and 36 g/t of dysprosium-terbium (DyTb) oxides, respectively. Nd and Pr are principal ingredients in high-performance permanent magnets; Dy and Tb can enhance their performance.
Tables 2 and 3 show variability across the selected samples and QC repeat. These preliminary results do not establish reproducibility at larger scale, an optimized process or commercial viability.
Development strategy and next testwork
Magnes plans integrated testing with the Australian Nuclear Science and Technology Organisation (ANSTO), whose minerals team develops rare-earth flowsheets.[13] Work will focus on a priority area of Minas Americas that management considers sufficiently encouraging to progress. It would follow leaching, solid-liquid separation and impurity removal through production of mixed rare-earth carbonate (MREC) for preliminary processor evaluation, not direct magnet use. Tank leaching is the proposed base case, with preliminary heap-leach and in-situ recovery assessment. Management has approved the proposed scope. The ANSTO quotation has been signed and acknowledged, and samples have been dispatched for the programme.
In parallel, Magnes plans extensive SGS bench tests on dozens of additional samples from different areas of Minas Americas, selected using assays and geological interpretation. This wider campaign will assess geological and metallurgical domains, variability in leach response and impurities, and area-specific processing requirements.
Together, these workstreams would guide the prioritization and sequencing of areas for potential future mining and flowsheet optimization, measure full-process recovery and provide inputs for later engineering and economic assessment, including actual-volume balances, filtration, washing, reagent and water requirements.
Impurities in the leach and wash solutions
Table 1 summarizes the liquid impurity results. Samples 1107 and 1108, which gave the highest estimated MREO, also returned thorium below detection in all three streams. Despite 23.01-34.17% Fe2O3 in the prepared feed, the limited dissolved iron supports further investigation of selectivity, not a removal percentage or processing-cost conclusion. Impurity removal can consume reagents and cause rare-earth losses.[12]
Table 1. Selected impurities in primary leach and wash solutions
| Element | SGS result (mg/L) | Coverage and interpretation |
| Iron (Fe) | <1.28 | Below detection in 23 of 24 streams |
| Iron (Fe) | 1.6 | Sample 0168, second wash |
| Cadmium (Cd) | <0.08 | Below detection in all 24 streams |
| Chromium (Cr) | <0.64 | Below detection in all 24 streams |
| Uranium (U) | <0.04 | Below detection in all 24 streams |
| Thorium (Th) | <0.12 | Below detection in 21 of 24 streams |
| Thorium (Th) | 0.28 / 0.15 / 0.18 | 0168 primary / 0832 primary / 0168 second wash, respectively |
Sources: SGS HM2600006-8, pages 2-3. The 24 streams cover a primary solution and two washes for seven samples plus the 1109 QC repeat. "<" means below detection, not zero. The uranium threshold is 0.04 mg/L (40 micrograms per litre). These are analytical thresholds, not safety or product-specification limits.
Other measured constituents reached 161.9 mg/L aluminium, 219.99 mg/L calcium, 389.12 mg/L magnesium, 0.8 mg/L arsenic and 1.07 mg/L lead. Integrated testing must assess their purification behaviour, rare-earth losses and final concentrate specifications.
These separately assayed streams are not combined liquor or purified concentrate. Uranium and thorium were present in the solid feed; chemical concentrations are not radioactivity measurements, and below detection does not establish zero radioactivity.
Complete selected results and sample origins
Table 2. Estimated oxide equivalents in solution plus both washes in g/t of prepared laboratory feed
| SGS sample | Auger hole | Interval (m) | MREO (g/t) | HREO (g/t) | TREO (g/t) |
| 0168 | MAV_AD_0013 | 8-9 | 55 | 24 | 155 |
| 0832 | MAV_AD_0068 | 10-11 | 251 | 82 | 558 |
| 1093 | MAV_AD_0113 | 0-1 | 391 | 83 | 895 |
| 1095 | MAV_AD_0113 | 2-3 | 403 | 127 | 852 |
| 1107 | MAV_AD_0115 | 6-7 | 982 | 194 | 2,322 |
| 1108 | MAV_AD_0115 | 7-8 | 892 | 168 | 2,073 |
| 1109 | MAV_AD_0115 | 8-9 | 549 | 112 | 1,324 |
| 1109 QC | MAV_AD_0115 | 8-9 repeat | 542 | 111 | 1,319 |
Table basis: Company-calculated oxide equivalents from SGS report Table 6 and Tables 8-15, using the calculation and feed basis in "SGS test conditions and reporting basis". Original IDs are the SGS labels prefixed AD; QC is a repeat of AD1109. Heavy rare earth oxides (HREO) overlap MREO through Dy and Tb; total rare earth oxides (TREO) include both.
Table 3. Estimated composition in solution plus both washes in g/t of prepared laboratory feed
| SGS sample | NdPr oxides | DyTb oxides | SEG oxides | SEG plus HREO | REO excluding La and Ce |
| 0168 | 51 | 4 | 15 | 39 | 90 |
| 0832 | 240 | 11 | 53 | 134 | 374 |
| 1093 | 378 | 13 | 70 | 153 | 531 |
| 1095 | 383 | 21 | 92 | 219 | 602 |
| 1107 | 941 | 41 | 214 | 409 | 1,350 |
| 1108 | 856 | 36 | 187 | 355 | 1,211 |
| 1109 | 525 | 24 | 118 | 231 | 756 |
| 1109 QC | 519 | 23 | 120 | 231 | 750 |
Table basis: As in Table 2; all results are g/t. SEG plus HREO overlaps MREO through DyTb and is not additive. REO excluding La and Ce counts each oxide once, including MREO.
pH tests and the aluminium burden
Sample 1109's nominal-pH-5 test showed approximately 85-86% lower primary-solution aluminium concentration than the earlier pH-4.2 test and repeat. Table 4 shows the associated oxide-equivalent estimates; Figure 1 shows aluminium per gram of dissolved magnet rare earths.
Table 4. Sample 1109 estimated oxide equivalents in primary solution plus the first wash
| Nominal pH | MREO (g/t) | HREO (g/t) | TREO (g/t) |
| 3 | 576 | 111 | 1,399 |
| 4.2 - earlier test | 549 | 112 | 1,324 |
| 4.2 - earlier repeat | 542 | 111 | 1,318 |
| 5 - trajectory | 493 | 96 | 1,209 |
Table basis: Company-calculated oxide equivalents in g/t of prepared feed, using primary solution plus the first wash in every row. Earlier pH-4.2 rows exclude the second wash; Table 2 includes both. See "SGS test conditions and reporting basis" for the calculation basis.
Figure 1. Measured aluminium relative to magnet rare earths in primary solution

Figure basis: Elemental primary-solution assays, excluding washes. Separate test observations, not a continuous pH response curve; lower values mean less measured Al per gram of dissolved Nd + Pr + Dy + Tb. Sources: SGS baseline and sample-1109 workbooks, 30 September 2026.
Iron was 3.0 mg/L at pH 3 and below 1.28 mg/L otherwise. Uranium and thorium were below 0.04 and 0.12 mg/L, respectively, in all four primary liquors. Calcium and magnesium did not show comparable reductions; lower aluminium burden does not establish concentrate purity.
The nominal-pH-5 test began near pH 4.3 and was raised with ammonium hydroxide after five minutes; it was not replicated. Earlier pH-4.2 tests were separate, at about 18 degrees C versus 25 degrees C, with two washes versus one. The differing conditions and limited replication prevent attributing the response solely to pH or defining an optimum.
Table 4 sources: SGS report dated 14 October 2026, Tables 14-15 and 22-23; Company element-by-element oxide calculations checked against the 30 September 2026 baseline and sample-1109 workbooks. Baseline distributions use Table 6. Primary-liquor aluminium comparisons use Tables 19-21 and the same workbooks.
SGS test conditions and reporting basis
This selected material differs from Verde's earlier diagnostic intervals and is not a before-and-after recovery comparison. Follow-up pH and reagent tests used sample 1109.
Main tests used approximately 40 grams of homogenized minus-2 mm material and 160 mL of 0.5 M ammonium sulphate, with mechanical agitation and pH monitored and adjusted around 4.2. Leaching lasted 30 minutes at ambient temperature, followed by vacuum filtration and two dilute-sulphate washes. SGS used inductively coupled plasma mass spectrometry (ICP-MS) and optical emission spectrometry (ICP-OES). No high-temperature cracking stage was used.
SGS supplied the assays and normalized balances; the Company applied each element's normalized distribution to its experimental head assay, converted to oxide equivalents and summed the groups before rounding. The balances assume the same nominal volume for each liquid stream as the initial leach solution, rather than actual collected volumes, which SGS describes as a maximum-recovery calculation. Below-detection results are zero-coded in the source balances. These liquid-stage estimates are per tonne of the prepared feed described above, not measured collected-liquid yields, physically produced oxides or independently reconciled dry-whole-ore product yields. Actual recovered volumes require further filtration testing.
Drilling progress and resource work
The geological dataset provides the basis for resource modelling and wider metallurgical sample selection. Previously reported intersections of 30 metres in MAV_RC_0063 and 53 metres in MAV_RC_0001 both begin at surface. These are downhole lengths, not established true widths. They are head-grade results and do not demonstrate that the selected SGS leaching response extends over those intersections.
Verde's 6 October 2026 release used a 25 September cutoff and covered 7,468.8 assayed metres in 879 holes, against 12,640.6 recorded drilling metres in 1,191 holes. These remain the previously published reporting populations, not updated totals for the whole physical drilling programme.
Field and sample-preparation records extend into October, and additional assay results have been received since that cutoff. Further results will be reported following reconciliation and technical review. Geological interpretation and preparation of an initial mineral-resource estimate and supporting report remain in progress. No mineral resource, reserve, mine plan, production rate or economic study is announced for Minas Americas.
Industry context and product benchmarks
Permanent magnets are used in industrial motors, robotics, AI data centres, aerospace and defence. The International Energy Agency puts permanent magnets at around 95% of rare-earth consumption by value.[1] These end uses explain the focus on MREO.
Samarium, europium and gadolinium (SEG), together with other heavy rare earths, may broaden the potential product suite. Separated materials are used in samarium-cobalt magnets, europium phosphors and gadolinium-based MRI contrast agents, while yttrium is used in ceramics, phosphors and laser crystals.[5-8]
Industry studies provide context for the intended product, but report different processing endpoints. Magnes's selected 982 and 892 g/t MREO estimates are at the liquid stage; Table 5 shows full-life forecasts of MREO contained in product per tonne of feed. Magnes has not yet produced a characterized MREC from this programme; product yield, specifications, payability and production costs remain to be established.
Table 5. Feasibility study forecasts of MREO contained in product per tonne of feed
| Project | Full life study basis | MREO in forecast product (g/t feed) |
| Meteoric Caldeira | 23-year DFS forecast MREC; 151.071 million tonnes of processed feed | 607.4 |
| Viridis Colossus | 25-year DFS forecast product; 123.5 million tonnes of scheduled feed | 600.9 |
| Aclara Carina | 18-year FS forecast MREC divided by processed dry feed | 140.1 |
Comparison basis: Company calculations from the cited studies.[2-4] Differences in feed selection, preparation, process endpoint and accounting prevent direct recovery or economic comparisons. The forecasts do not establish project superiority and are not assumed Magnes outcomes.
Table 6. Illustrative MREO calculations for resource-stage projects
| Project | Published basis used in Company calculation | Illustrative MREO (approx. g/t) |
| IMC - Itarantim [9] | Average resource oxide grades multiplied by element-specific recovery assumptions in the initial assessment. | 135 |
| Rare Earths Americas - Alpha [10] | Average block-model oxide grades multiplied by mean modelled leach-extraction percentages. | 101 |
| Rare Earths Americas - Constellation [11] | Average block-model oxide grades multiplied by mean modelled leach-extraction percentages. | 241 |
Calculation basis: The Company multiplied each magnet oxide grade (Nd2O3, Pr6O11, Dy2O3 and Tb4O7) by its corresponding recovery assumption or modelled extraction percentage, then summed the results, rounded to the nearest g/t of mineralized material.[9-11] For Alpha and Constellation, these are products of separately published averages; they do not establish average extracted grades from matched blocks. The figures are illustrations, not issuer-reported recovered mineral resources, measured yields or forecasts of saleable product. No additional allowance is made for downstream losses. They are not directly comparable with Magnes's selected bench tests or the full-life product forecasts in Table 5.
Definitions technical disclosure and Qualified Person
MREO = Nd2O3 + Pr6O11 + Dy2O3 + Tb4O7. NdPr = Nd2O3 + Pr6O11; DyTb = Dy2O3 + Tb4O7. SEG = Sm2O3 + Eu2O3 + Gd2O3. Company HREO = Y2O3 plus oxides of Tb, Dy, Ho, Er, Tm, Yb and Lu, excluding Eu and Gd; SGS's elemental HREE grouping includes Eu and Gd. Y contributes materially to HREO, which must not be treated entirely as magnet metals. TREO includes Y and excludes Sc, using CeO2, Pr6O11, Tb4O7 and the other rare-earth sesquioxides. REO excluding La and Ce = TREO minus La2O3 and CeO2. These groups express oxide-equivalent mass, not economic value based on prices, payability or recoveries.
Sampling analytical procedures and quality control
The samples in Table 2 are one-metre auger intervals from Minas Americas, Minas Gerais, Brazil. The four holes are recorded as vertical, with dip -90 degrees and azimuth 0 degrees. Intervals are downhole; true mineralized widths are undetermined. Selection has not been shown to be statistically representative of the property or wider drilling database.
SGS Geosol Laboratórios S.A., Vespasiano, Minas Gerais, performed the bench tests and analyses under the conditions in "SGS test conditions and reporting basis". SGS's 14 October 2026 report states that it applied QA/QC procedures and operational traceability. Prepared-feed analyses included a preparation blank, a repeat of sample 1109 and reference materials OREAS460, OREAS461 and SG_241; leach results include a repeat of 1109.
Qualified Person and data verification
The scientific and technical information in this release relating to exploration results has been reviewed and approved by geologist Carlos Leite.
Carlos Leite is an independent consultant to Magnes Rare Earths, MAusIMM CP(Geo) 1006695, and has sufficient experience of relevance to the styles of mineralisation and types of deposits under consideration, and to the activities undertaken to qualify as Qualified Person under National Instrument 43-101 — Standards of Disclosure for Mineral Projects.
About Verde AgriTech and Magnes Rare Earths
Verde AgriTech Ltd. trades on the Toronto Stock Exchange as NPK and on OTCQX as VNPKF. Magnes Rare Earths, Verde's subsidiary, is advancing the exploration-stage Minas Americas Global Alliance rare-earth project in Minas Gerais, Brazil. The listed securities referenced in this release are Verde AgriTech shares.
For further information: investor@verde.ag. Verde investor information: https://investor.verde.ag. Magnes project and company information: https://www.magnesrareearths.com.
Sources
SGS source: SGS Geosol, "Programa piloto de testes de dessorcao dos elementos terras raras por meio de lixiviacao com sulfato de amonio - Projeto Magnes Rare Earths", technical report dated 14 October 2026, sections 5.2-5.3, Table 6 and Tables 8-15. Sample origins: Company auger sampling register. Prior drilling disclosure: Verde release of 6 October 2026, Tables 1 and 2 and drilling reporting-progress table, with a 25 September 2026 cutoff.
Impurity sources: SGS certificates HM2600006 (primary leach solutions, issued 29 July 2026), HM2600007 (first wash) and HM2600008 (second wash), both issued 27 July 2026, pages 2-3, analysed by ICP-MS and ICP-OES; solid-feed context from HM2600003, issued 30 July 2026, pages 2-3. Table 1 uses the reported sample-row results from all three liquid certificates.
Reference [1]: IEA Rare Earth Elements Executive Summary 2026. Peer references: [2] Meteoric Caldeira DFS, 31 July 2026, pages 1 and 4; [3] Viridis Colossus DFS, 20 August 2026, pages 9 and 32; [4] Aclara Carina FS results, 13 April 2026, Table 4 and Hatch FS, effective 20 March 2026, Table 16-1.
Application references: [5] DOE Ames Laboratory, samarium-cobalt magnets; [6] USGS critical minerals and uses; [7] FDA, MRI and gadolinium-based contrast agents; [8] USGS Mineral Commodity Summaries 2026 Yttrium.
Additional project references: [9] IMC Itarantim Initial Assessment and Technical Report Summary, effective 31 March 2025, section 11.8.1, Table 11-6 (page 66) and the calculated-assumed-revenue Table 11-8 (page 67); [10] Rare Earths Americas Alpha Technical Report Summary, updated 25 March 2026, Tables 10-2 and 11-15; [11] Rare Earths Americas Constellation Technical Report Summary, updated 25 March 2026, Tables 10-2 and 11-16.
Processing context [12]: Li et al., Iron, aluminum, and thorium impurity removal from a rare earth element pregnant leach solution using magnesium carbonate, Scientific Reports 15, 15951 (2025), published 7 May 2025. General purification principles only; different feed and process conditions, not a Magnes performance benchmark.
Development and drilling update sources: [13] ANSTO Minerals process-development capabilities, https://www.ansto.gov.au/services/resources-sector/minerals/consultancy/capabilities; Company FIELD_DATA.xlsx, current live version checked 20 October 2026, HEADER_AD and BATCH sheets; SGS certificate GQ2611383, batch MAV_AD_26138, printed 16 October 2026.
Cautionary statement regarding forward looking information
This release contains forward-looking information under applicable Canadian securities laws, including potential future supply of magnet rare-earth products, planned SGS testing, the ANSTO programme, assays, resource work, potential future mining priorities, flowsheet development and later engineering and economic assessment. These statements concern possible future events or performance and are based on assumptions as of the release date.
Material assumptions include sufficient suitable samples and funding; available laboratories and qualified personnel; adequate geological data and successful completion and technical review of sampling, assays and variability tests; acceptable contracts and transport arrangements; and required rights, access and permits.
Actual results, timing and outcomes could differ materially because of geological and metallurgical variability or unrepresentative samples; delayed or revised assays and interpretations; errors or limitations identified through QA/QC and verification; differences between nominal and collected liquid volumes; filtration, washing and scale-up difficulties; rare-earth losses during purification or precipitation; impurity or radionuclide behaviour; reagent, water and residue-management requirements; or failure to meet a processor's product specifications.
Other risks include insufficient funds, equipment, services or personnel; contract, transport and laboratory delays; mineral-title, land-access, environmental, permitting or regulatory constraints; changes in costs, exchange rates, rare-earth prices, demand, trade policy or downstream capacity; and failure to define a resource or establish economic viability. These release-specific risks are supplemented by those in Verde's continuous-disclosure filings under the Company's profile on SEDAR+ at https://www.sedarplus.ca.
Forward-looking information is not a guarantee, and readers should not place undue reliance on it.
A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/6e144d02-d14b-4c8c-bcf4-b10d02861887