THE MASTER PLANAUGUST 2026
The Widget Enterprises
Master Plan
Directed evolution for consumer goods: one system launches the brands, the market applies the selection pressure, and every launch inherits what the last one learned.
This page is dated and left standing. We will not edit it to look right in hindsight. Judge us against it.
01 / THE PROBLEM
The product is cheap. The company that makes and sells it is not.
Walk any aisle and the price on the shelf is mostly not the product. It is the company around the product: hundreds of people across functions and agencies, dozens of disconnected software tools, weeks for a single decision to cross the org. The consumer pays for all of it, now at record prices.
is all the brand keeps from every dollar the consumer spends.
The rest is the company in between: the retailer, the overhead, the people and the tools. We carry the product, not the company around it.
Take that structure out and the same saving shows up three ways. A new brand reaches market on a fraction of the usual capital, at equal or better margin. Concept to live takes days, not quarters. And at scale the consumer pays roughly half the incumbent's price for the same product on the same shelf, while we still keep equal or better margin. That last number is the point of the whole plan.
02 / THE ORGANISM
One operator. The system runs the brands.
Widget Enterprises is not a holding company with a playbook. It is a single organism, Metis, that senses the market, makes the product (a widget), and sells it. The work that needs a department at an incumbent needs no one here. Add a widget and the cost barely moves: one system runs every widget at once, under one operator's oversight.
The engine alone is not the moat. Anyone can rent the same intelligence tomorrow. What they cannot rent is the genome: the ledger of every decision the organism has made and what the market did in response. Every decision is captured with its prediction frozen at the moment it is made. When the market answers, the outcome is recorded against that prediction. Before the next like decision, the genome is read back as a prior. The gap between what we expected and what happened is the learning, and it compounds.
WHY IT COMPOUNDS
This is created data, not collected data. It does not exist anywhere until we run the widget, and a competitor would have to run the same rounds to get it. Each launch is born carrying everything every previous launch learned, so a competitor starting today is not behind by a feature. They are behind by every cycle we have already run. That is directed evolution, and the genome is the only reason it compounds instead of repeating.
03 / THE PLAN
Three phases. Each one funds the next.
Prove it.
Launch brands from the pipeline and let the market grade the system. Six brands are greenlit and in launch now, with ten more scored and queued for 2027, each one chosen by the system reading live demand rather than instinct. The engine is funded by the operating business it already runs, so the plan does not wait on anyone's permission to keep moving.
Every launch in this phase is a labeled experiment. The result matters twice: once as revenue, and once as an entry in the genome that makes the next call better.
Build and buy.
Keep incubating widgets we conceive, launch and own outright, carrying the cost and keeping the P&L. And extend the organism into the making itself: acquire and rebuild manufacturers, installing Metabolism, the making half of the system, so the plant runs the way the brands already do. The books close daily, quality gates enforce themselves, and a human is paged only for the calls that need judgment.
Owning the shelf and the plant on the same system is what lets the whole margin structure move at once. A business we rebuild pays us in the economics we unlock, and every business we run deepens the same genome.
Return the margin to the consumer.
At scale, the structural saving goes where it belongs: to price. Same product, same shelf, roughly half the incumbent's price, and we still keep equal or better margin. The incumbents cannot follow without dismantling the very cost structure they are made of. That is not a promotion. It is the permanent consequence of removing the company from the price of the product.
04 / THE HONEST PART
Most attempts at this fail.
Brand studios fail when every launch starts from zero and the holding company's overhead eats the margin it promised to return. Roll-ups fail when the platform is a slide, not a system. We are not exempt from those forces, and this page does not pretend we are.
The odds are different here for three reasons. The system is not a plan for software: it is in market today, running live businesses, and funded by the operating business it runs. The judgment encoded in it comes from people who have run the factories and the P&Ls in consumer goods, not from guesses at the category. And the genome only records decisions at the moment they are made, never reconstructed after the fact, so the learning it compounds is real.
And here is what would kill it. Launching faster than the loop can close, so the genome fills with bets and no outcomes. Letting the operating business that funds the engine decay while attention chases the new. Buying a plant whose problems are physical rather than organizational, because no system fixes a broken asset. We name these because we watch them, and because a plan that cannot say what would kill it is not a plan.
The company that runs itself.
This is the plan of record. The platform walkthrough, the moat thesis and the full data room open to serious investors and partners on request.
WIDGET ENTERPRISES
DIRECTED EVOLUTION FOR CONSUMER GOODS