ElasticRun

Their pitch was 'we own the rural retailer relationships and the hyperlocal demand data' — which worked beautifully until AI-powered demand forecasting started letting brands skip the middleman entirely. Claude doesn't need a relationship with the kirana store; it just needs to know when they're running low on soap.

Route optimization, inventory management, and demand prediction — the entire operational backbone — are textbook AI automation targets. Their field teams still matter for relationship management, but the algorithms that made those relationships profitable are becoming commodity.

The physical network is real — actual relationships with actual stores in actual villages that took years to build and would take competitors years to replicate. But moats only matter if you control the value creation, and when AI can optimize inventory and routing from satellite data, the network becomes expensive infrastructure for a function that's being commoditized.

They've been quiet on AI integration despite running exactly the kind of complex logistics operation that benefits most from machine learning optimization — either they're building it in stealth mode or they're hoping the relationships outlast the algorithms.

Getting products to rural India survives. Needing a specialized platform to figure out how to do it doesn't — AI turns hyperlocal distribution from a relationship game into a math problem, and math problems don't need middlemen.