The Private Theorem Economy
Y. Wayne Wu@waynewu
−3.4%
since Oct 10
The idea
I believe AI is turning math itself into a tool companies can wield like a trade secret, and the winners will be businesses with private data and a fast way to test if an answer works, a modern version of the secret formula. This basket spans pricing engines where the formula is the business, chip and engineering design software, computational drug discovery, trading firms, and Alphabet as a steadier anchor given DeepMind's math breakthroughs. No single name dominates, a small cash buffer sits on the side, and this is built to be held for five to ten years, not traded on headlines.
What it owns
- These companies already live off a guarded formula, whether it scores credit, prices insurance, ranks ads or indexes risk, and better math can make those formulas sharper and more profitable. · Fair Isaac CorporationThe FICO score is a guarded formula that lenders and mortgage rules treat as the standard and pay to use., Meta PlatformsIts ad ranking engine runs on user data nobody else has, so every improvement in that math turns straight into ad revenue., ProgressiveIt prices car insurance with models built on its own driving and claims data, and the payoff shows up directly as underwriting profit when its pricing stays ahead of rivals., AppLovinIts ad auction algorithm is the business: better math means advertisers get more sales per dollar and send more money its way., Verisk AnalyticsIt sells insurers risk models built on decades of pooled claims data that a newcomer cannot recreate, even with the same AI tools., MSCI Inc.Its risk models and indexes are formulas that investors pay to use, and funds tracking them are locked into that math.35%
- Chip designers and manufacturers rely on simulation software to test their most secret designs before building them, so improvements in the math behind layout and testing flow directly into the tools these companies license. · SynopsysChip companies run their most secret designs through its simulation software, so as AI makes the math behind chip layout better, the payoff lands inside tools it sells., Cadence Design SystemsIt sells simulation software for chip design, entire systems and molecules, so it gets paid whenever companies want to test a private design before building it., PTC Inc.Manufacturers keep their product designs in its engineering software, which is exactly where a company's own formulas for a part or machine would live and get tested., Keysight TechnologiesIt sells simulation and test gear that checks chips, radios and networks before they ship, so companies keep paying for faster proof that a new design performs.25%
- Market makers and trading firms turn a better pricing formula into money the same day, which makes them the most direct payoff of proprietary math when their models stay ahead of rivals using the same techniques. · Interactive BrokersIt runs on heavily automated systems where its edge is software and math doing work that rivals still pay people to do, and its cost advantage shows in growing accounts., Virtu FinancialIts whole business is a private pricing formula that the market tests every day, and its trading income shows whether its models still have an edge against rivals.13%
- Drug discovery is becoming a search problem that simulation and AI can solve faster than lab work alone, especially when companies own private biology data that rivals cannot recreate. · SchrödingerIt uses physics-based math to predict how drug molecules behave before anyone makes them, letting drug companies test their private targets without the lab work., CertaraIts biosimulation software models how drugs move through the body, and regulators already accept those models, making it a working example of math that saves real trial time., Recursion PharmaceuticalsIt builds its own library of biology data and feeds it to AI, pairing private data with math to discover drug candidates.12%
- Alphabet owns DeepMind, a leading lab producing fundamental AI math results, and couples it with a large profitable business that generates the cash to fund long-term research. · AlphabetIt owns DeepMind and a huge profitable business to fund it, providing ballast for a take built on math results that take years to turn into products.8%
- Cash7%