Real-World Data Moats
Dice@dice
+2.5%
since Oct 6
The idea
I believe that as AI models become commoditized, the real edge moves to whoever owns proprietary real-world data, and consumer ops platforms are best positioned to collect it through every delivery, ride and booking. This basket concentrates on one layer of one value chain: twelve platforms spanning delivery, global commerce, mobility and travel, weighted toward the ones like DoorDash and Uber already feeding this data back into their own models. I keep a cash buffer and avoid betting on any single name, planning to hold for years unless the data-moat thesis itself breaks down.
What it owns
- Every order maps real streets, kitchens and shopping behavior that cannot be scraped from the internet, giving platforms ground-truth operational data to improve efficiency. · DoorDashEvery delivery produces ground-truth data on task timing, customer orders and city movement patterns that feed back into its own logistics optimization., ToastIt captures every order, menu change and kitchen timing inside restaurants, creating operations data that no web crawler can reach and that drives value for its customers., Instacart (Maplebear)It sees what is actually on grocery shelves across thousands of stores and sells that visibility back through ads and retailer tools that depend on proprietary catalog data.28%
- Emerging-market platforms own end-to-end commerce and logistics data in regions where most of the physical economy is not documented online, making their datasets difficult to replicate. · MercadoLibreIt sees commerce, payments and logistics across Latin America, where little physical economic activity is documented online, making its dataset close to irreplaceable in the region., Sea LimitedIts Shopee platform collects shopping, payments and delivery data across Southeast Asia at a scale no competitor in the region matches., CoupangIt runs its own end-to-end logistics in Korea and captures ground-truth delivery, inventory and demand data that pure marketplaces cannot access.22%
- Ride platforms log how people and vehicles actually move through cities, minute by minute, creating proprietary datasets that feed directly into routing and pricing models. · Uber TechnologiesIt maps how people and goods move through cities minute by minute and already feeds that graph into routing and pricing models that improve with every trip., LyftIt collects the same ground-truth city movement data as other ride platforms and turns that dataset into better matching and pricing models., Grab HoldingsIt collects rides, deliveries and payments data across Southeast Asia, a region where the physical world is especially poorly mapped online, making its data hard to replicate.21%
- Booking and stay data captures real demand, pricing and trip behavior across the world, creating datasets that AI systems can train on to improve recommendations and search. · AirbnbIt knows where people actually want to stay, what they will pay and how hosts respond, creating booking data for physical spaces that cannot be replicated elsewhere., Booking HoldingsIt holds the largest record of how people actually travel globally and trains AI trip-planning systems on that historical booking behavior., Expedia GroupIt captures trip behavior across flights, hotels and packages, creating a second major proprietary travel dataset that feeds into AI-driven product improvements.19%
- Cash10%