The upside case is a real shipment-and-mix story, not a science project. More
edge AI workloads should raise demand for local vision compute, and Ambarella already has credible products, customers and a reusable software layer. If automotive ramps, security AI upgrades and robotics deployments land together, revenue can scale faster than operating expense growth. I stop short of a hypergrowth call because most capture still comes from silicon sockets, not a software tollbooth, and customer timing remains the main gate.