From Poison to Power

The AI buildout secured the demand and skipped the inputs. Power was the first wall. Water is the next. We solved the inputs first — the way a mechanic always does.

Water isn't fungible like power. You can wheel electrons across a region. You can't wheel in a watershed.

A nine-stage refinery on a free, co-located feedstock. FEOC-free. Domestic. Conserved mass balance — atoms in, atoms out.

A bench-scale validation rig at Rensselaer Polytechnic Institute, running real Permian produced water through the full nine-stage train. It converts a red-teamed, investor-grade model from demonstrated on paper to demonstrated on feed — closing every open technical question before we build the pilot.

Computed bottom-up from the unit operations at conserved throughput — not targeted. The margin isn't magic: it's the water-midstream, soda-ash, and chlor-alkali margins the market pays three separate companies, captured once, on a free feedstock.

One check funds the rig. The rig de-risks the pilot. The pilot's cash flow funds the scale. Small entry, sequenced de-risking, clear ladder up.

That's how much sodium-ion battery capacity China controls today. We're building the domestic answer — into a market that already exists.

Pilot output (~14,300 t/yr) is a rounding error against 73M tons — so we don't need share, we need to exist as the domestic option. (Illustrative ceiling: two mega-plants ≈ 14% of global soda ash from a single domestic source.)