This talk will explore how the commercialization of AI innovation and associated AI startup activity has happened in waves, each with its distinctive pattern. Since deep learning's ImageNet moment a decade ago, broad-based research-driven commercialization has given way to use case-specific AI applications. In the next major wave, AI technologies will become embedded across all software applications. The talk will cover the progression and successes of AI commercialization wave by wave with illustrative examples. It will conclude with a discussion of current and future startup opportunities with time for Q&A.