Windfall Geotek and Gold Hunter deploy AI for Newfoundland gold exploration

This is a small-dollar deal, but a good illustration of how mature, specialized AI is becoming embedded in old-economy sectors like mining. Windfall Geotek isn’t training giant frontier models; it’s using well-honed machine-learning pipelines to fuse magnetic surveys, VTEM data and tens of thousands of historical assays into high-probability drill targets, with enough confidence to accept partial payment in client equity.

For the AGI race, the message is that domain-specific AI continues to quietly compound value even as attention focuses on foundation models. Mineral exploration is data-rich, spatial, and uncertain—exactly the kind of problem where agentic systems that iterate between hypothesis, simulation and action can shine. If these workflows prove meaningfully more efficient than traditional geologist-led targeting, capital will keep flowing into narrow, applied AI, even without AGI. That, in turn, builds complementary capabilities: data infrastructure, simulation pipelines, and human–AI collaboration patterns that will make it easier to deploy more general agents later.