NVIDIA’s open-sourcing of a 34-billion-parameter robotaxi AI model in early August is viewed by Autonomy AI co-founders Grayson Brulte and Rob Grant as a strategic platform capture event, not charity. On their Road to Autonomy podcast, they argued that NVIDIA is giving away the model to make it economically irrational for AV developers to build their own, while locking the downstream pipeline into NVIDIA’s Drive Thor silicon and DGX cloud infrastructure. The model never touches a production vehicle; it trains distilled edge models that run exclusively on NVIDIA hardware. The episode highlighted three signals: NVIDIA’s open-source play reshapes who can afford to compete and who controls the compute; WeRide’s partnership with Denmark’s Green Mobility accelerates the “autonomous Belt and Road,” with Chinese AV companies using domestic cash flow to finance European expansion; and Aurora Innovation’s scramble to retrofit 200 driverless trucks by year-end, after OEM partner PACCAR refused to allow driverless operations, reveals brittle scaling architecture. Grant described Alpamayo as a multitask framework synthesizing 360-degree spatial awareness from camera inputs, producing chain-of-causation reasoning, and outputting tactical intent and physical trajectories. NVIDIA’s roughly $5 trillion market capitalization enables this play. “They’re basically giving away the model in order to sell the silicon and the cloud,” Grant said. The target is second-tier AV developers, not Waymo or Tesla, as the