Sherbrooke, Quebec–based startup Quminex said on October 14 that it has closed a CAD 1.5 million (approximately $1.06 million) pre-seed round led by local fund Quantacet, with participation from Swiss QAI Ventures and angel investors.
Quminex’s Discovery Platform
Quminex will use the funding to develop the Quminex Discovery Platform, a service designed to select drilling locations. The company reported its first commercial projects: it identified targets on ground held by private Tech-X Resources of British Columbia and on ground held by an unnamed mid-sized mining company. Contract amounts were not disclosed. Quminex also stated on LinkedIn that it will deploy the platform on its own account to acquire prospective exploration licences in Canada.
Features over intuition
According to QAI Ventures CEO Alexandra Beckstein, geologists at early exploration stages still rely heavily on expert judgement because source data are scarce and fragmented. S&P Global reported that in 2025 budgets for finding new targets fell 8% to $2.57 billion, representing 21% of global subsurface study spending in the nonferrous metals segment—the lowest share on record.
Quminex’s platform ingests a client’s accumulated geochemical, geophysical and spatial data, cleans and harmonises them, and brings them to a common format. It then constructs derived indicators (features) intended to distinguish prospective rock from barren material, selects the most informative features, and trains machine learning algorithms on that set. The company describes the output as an explainable mineral system model that shows which factors influenced the forecast and can be validated by drilling. Customers receive a list of targets across their tenements. Quminex does not disclose accuracy metrics.
Qubits take a back seat
When Quminex announced its first pre-seed round in April 2025, with participation from QV Studio, Quantacet and Quantonation, the startup described its technology as combining AI and qubit-based computing. In the new release the second component is not explicitly named, and on LinkedIn Quminex presents it as high-performance computing (HPC). The team includes two specialists exploring whether quantum algorithms could improve predictive models, although they have not demonstrated a practical advantage over classical computers in any applied domain.
Industry context and comparable efforts
U.S. company KoBold Metals raised $537 million in January 2025 at a $2.96 billion valuation. KoBold’s tools include TerraShed, which consolidates public and private geodata into visual maps, and Machine Prospector, which helps determine likely deposit locations. In March 2026 KoBold said it had begun developing the Mingomba copper deposit in Zambia: shaft sinking is planned for 2027 and output of 300,000 tonnes of copper per year is targeted for the early 2030s; the project will require $2.3–2.5 billion.
Adelaide-based Fleet Space combines low-Earth-orbit satellites, “smart” seismic sensors and AI in its ExoSphere system to obtain near real-time 3D images of the subsurface. In December 2024 the company closed a Series D round of AUD 150 million and said its valuation exceeded AUD 800 million; Fleet Space reports more than 40 exploration firms have used ExoSphere, including Rio Tinto and Barrick Gold.
Startup Earth AI drills algorithm-identified targets itself and then sells rights to prospective areas to miners; in January 2025 it raised $20 million from funds led by Tamarack Global and Cantos Ventures.
Quantum technologies are also entering exploration through sensors. In March 2024 Sherbrooke’s SBQuantum and the UK’s Silicon Microgravity received CAD 500,000 and £414,000 from their governments to build a drone equipped with a diamond quantum magnetometer and a gravimeter. In April 2026 Quantacet became a co-lead investor in SBQuantum’s $4 million seed round. In October, Quantinuum announced a world record for quantum computing accuracy.
