Tesla Supply Chain Consulting Project

Highlights

• Commissioned by Tesla to model a multi-country manufacturing location decision under real-world uncertainty.
• Replaces static cost comparisons with empirically calibrated posteriors drawn from macroeconomic time series.
• Produces actionable uncertainty bands across tariff shocks, FX swings, logistics disruptions, and event risk.

The brief

Tesla needed a rigorous, defensible answer to a factory location question that involved simultaneous uncertainty across raw material costs, labor markets, logistics networks, tariff exposure, and foreign exchange. A static spreadsheet model was not adequate. They needed a framework that could reason honestly about uncertainty rather than paper over it with a single expected value.

What we built

We replaced hand-crafted Normal priors with Student-t and Beta posteriors fit to FRED economic time series, capturing fat tails and asymmetric risks that standard models ignore. The output surfaces genuine uncertainty intervals across all three factory scenarios, giving decision-makers a clear view of downside exposure before committing to a location.