The archive · Space, Robots, Defence · Strategic decision · 2025–2026
Terranox bets AI on 70 years of exploration data can find uranium humans miss
YC W26's Terranox runs its own North American uranium exploration with AI trained on 70+ years of outcomes; backed by General Catalyst and 776.
Terranox AI
What the business is
Terranox runs its own uranium exploration projects in North America, using AI trained on decades of geoscience data and exploration outcomes to map high-prospectivity areas and choose the next drill location.
How it started
Jade Checlair (CEO), a UChicago geophysics PhD who developed statistical methods adopted by NASA flagship missions and spent 3.5 years at BCG on nuclear and mining strategy, and Leeav Lipton (CTO), an astrophysics-trained AI builder with eight years at Borealis AI and time at NASA JPL, met in first-year physics. They founded Terranox in 2025 and joined Y Combinator's Winter 2026 batch (Demo Day, March 2026).
What happened
Terranox's pitch rests on a structural gap: global uranium production needs to roughly quadruple by 2050, the world's largest mines approach end-of-life starting in the mid-2030s, and new mines take 10-15 years from discovery to production — while exploration still runs on intuition, with hit rates under 1%. The company says its system extracts decades of fragmented geoscience data into a unified context base, maps prospectivity humans would miss, and compounds a learning flywheel. It is backed by General Catalyst, 776 Ventures and Y Combinator, and announced its own North American exploration projects via Launch YC in August 2026, hiring exploration geologists based in Saskatchewan.
How it ended up
No deposit or drill result has been announced: as of late August 2026 Terranox was launching its first exploration projects, so the bet that AI can beat the industry's sub-1% hit rate remains untested in public.
Background
Terranox AI is a San Francisco company that says it is the first vertically integrated AI-powered uranium discovery business: it runs its own exploration projects in North America using AI trained on 70+ years of exploration outcomes. It was founded in 2025 by Jade Checlair (CEO, UChicago geophysics PhD, ex-NASA Ames, 3.5 years at BCG on nuclear and mining strategy) and Leeav Lipton (CTO, ex-NASA JPL scientist, eight years building AI/ML at Borealis AI), who met in first-year physics.
The bet is on a structural gap: global uranium production needs to roughly quadruple by 2050 to fuel the nuclear renaissance, the world's largest mines approach end-of-life starting in the mid-2030s, and new mines take 10-15 years from discovery to production — yet exploration still works like the 1960s, with hit rates under 1% and decades of data trapped in legacy PDFs and hand-drawn maps. Terranox's system combines multi-modal geoscience intelligence, prospectivity mapping tuned to uranium, and sequential decision intelligence that picks the next action to maximize information per dollar.
The company joined Y Combinator's Winter 2026 batch (Demo Day, March 2026) and is backed by General Catalyst, 776 Ventures and Y Combinator. PingWest's W26 recap singled it out as an example of YC's pivot toward extremely vertical niches — AI for uranium exploration alongside protein characterization and CAD tools — as generic AI workflow-automation pitches lost favor.
Terranox announced via Launch YC in August 2026 that it was launching its own exploration projects in North America and hiring exploration geologists based in Saskatchewan. As of that date no drill result or deposit had been announced, so the core claim — that a learning flywheel can beat an industry where hit rates are below 1% — remains untested in public.
What has to be true
- Structural scarcity: uranium supply is already in deficit, demand for nuclear fuel is rising, and new mines take a decade-plus, so discovery speed is the bottleneck.
- Data moat: 70+ years of exploration outcomes sit unused in legacy documents; the wedge is assembling and learning from that corpus, not the model itself.
- Learning flywheel: every drill hole, hit or miss, improves predictions across all projects, compounding with each dollar spent.
- Vertical integration: running its own exploration captures the value of discovery rather than selling software into a slow-moving mining industry.
- Founder fit: NASA-adjacent geophysics and AI experience aimed at a capital-intensive industry where hit rates are under 1%.
What can be applied
When an industry's asset is decades of unused, fragmented documents, the moat is the corpus and the learning loop, not the model — every failure improves the next decision.
Aftermath
As of August 19, 2026 Terranox was a two-person company based in San Francisco with exploration-geologist roles based in Saskatchewan, backed by General Catalyst, 776 Ventures and Y Combinator, and publicly launching its first North American exploration projects. It had not announced a drill program, deposit, or discovery, and had no disclosed revenue; the public record consists of the company's own claims about its system and the industry gap, with PingWest's W26 recap confirming its positioning within the batch.
Sources
- Launch YC: Terranox AI: AI-powered uranium discovery
- Terranox AI: The first AI-powered uranium discovery company
- YC Demo Day（W26）现场复盘：Make Something Agents Want
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card