GenAi has ushered in a new gold rush. Successive generational leaps in the underlying NVIDIA GPU infrastructure, together with the rapid evolution of both commercial and open-source LLMs, have disrupted how enterprises traditionally plan for growth. Speed and scale, however, are key to long-term enterprise success. Enterprises today need to adopt a foundational enterprise architectural model for GenAI to not only move faster, but also to scale for growth. In this talk, we examine lessons learned from earlier disruptive eras in Internet infrastructure, a reference architecture for building and maintaining a pipeline of continuous GenAI innovation, and the importance of integrating GenAI with core Cloud-native engineering tools and practices.