Tweet Overview
View this X/Twitter post from @StockSavvyShay published on April 15, 2026 at 05:26 PM. This post contains 1 video.
Jensen Huang basically argued that $GOOGL TPU is a specialized chip while $NVDA owns the broader AI compute stack that developers actually build on. The real moat is not just the chip but CUDA, programmability and distribution across every major cloud.
