
Every industry gets rebuilt, function by function.Until none of the original planks remain.
Intelligence does not land on an industry all at once. It replaces one function, then another, then another, until the company on the other side is a new thing that still remembers the shape of the old. This is the book of what we believe about those rebuilds, before they happen to you.
Four industries. Four wedges.
Each thesis is a point of view we're willing to stake an engagement on. Not research. Not a survey. What we believe, why we believe it, and how we would attack it if we were inside the company. Principals write them. Pods prove them. Bastion carries them forward.
-
Energy
Deregulated power markets are beatable with state-space and sequence models.
Power prices in deregulated US markets move faster than utility planning cycles can absorb. Load forecasting, price forecasting, and congestion prediction need models that handle regime shifts, not regressions tuned to last year's curve. Moative deploys state-space and sequence models on the same feeds traders use, turning raw volatility into positions on the desk and operating decisions on the floor.
The same engine, retrained on the facility's own load, moves from trading desks to datacenter curtailment to industrial demand response.
Deployment Live on energy trading desks. Extending into datacenter load curtailment and industrial demand response.
-
Healthcare
Healthcare's back office runs on labor that AI agents already do better.
U.S. practices spend nearly two dollars on admin for every dollar on care. Denials sit at 11.8 percent, physicians lose 13 hours a week to prior auth, and 62 percent of revenue cycle work is directly automatable. Traditional MSOs absorb this burden by hiring, so their cost base scales linearly with every practice they roll up. The wedge is an AI-first MSO where agents are the primary workforce and humans handle the exceptions.
Claims statusing drops from 19 minutes to 45 seconds. EBITDA moves from 15 percent to 30 percent. Headcount stops being the answer.
Deployment Live in production: claim statusing across 12+ payer portals and Lisa voice AI handling payer calls, intake, and scheduling. Prior auth, denials, and coding in build.Read the full thesis →
-
Equipment Leasing
Equipment leasing runs on residual curves no one actually believes.
A trillion-dollar industry prices its assets with spreadsheets, OEM self-reports biased to move equipment, and industry residual tables that lag the market by years. Every lessor quietly knows the numbers on their book do not match the numbers in the world. The consequence is overpriced deals walking away, underpriced deals eating margin, and portfolios that surprise the credit committee six months too late.
Underwrite in 40 minutes instead of 9 days. See portfolio deterioration six months before it hits the P&L.
Deployment Building now: portfolio-specific residual curve engines trained on each lessor's own disposition history. Dynamic curves replace static industry tables.
-
Paper & Pulp
State-space models turn paper mill energy systems into controllable surfaces.
Integrated paper mills burn 10 to 40 percent of production cost on energy, and the recovery boiler plus turbine island is where most of that money leaks. Fixed-cycle soot blowing, HP steam dumped into air preheaters, deaerator pegged on instinct, turbine inlet drift. No single operator holds the heat-mass balance in their head. Moative fits state-space models to plant historian data and learns the hidden cleanliness, fouling, and efficiency states the DCS was never built to expose.
One to three percent off steam and power cost. No capex. Paid for by assets the mill already owns.
Deployment Deploying intelligent soot blowing, SCAPH rebalancing, and deaerator pegging control on live recovery boiler islands. Every recommendation carries a direct dollar linkage to the mill's own unit cost basis.
The argument your industry is not yet making.
Wedges, mechanisms, and engagement shapes across four industries. Walk into your next board meeting with the thesis already written. Updated every quarter as the work ships.
Four is the starting line.
A new thesis goes in the book every time our pods run an engagement, ship a model, or spot a wedge we can defend. The overflow industries below are already on the whiteboard. Programmatic thesis pages for each are on the way.
- Insurance
- Distribution
- Logistics
- Pharma
- Retail
- Construction
- Agriculture
- Hospitality
- Education
- Real Estate
- Government
- SaaS
Think your industry belongs in the book?
The book grows by engagement, not by research. If you can see the wedge in your industry and want an operator at the table, that is how a thesis gets written.
- A midmarket industry with a known AI wedge.
- Real data the thesis can be grounded in.
- An operator on your side who will own the outcome.
- Willingness to structure around outcomes, not hours.