Hua Er Jie Jian Wen·2025-06-20 02:59Group 1 - The core metric for evaluating the commercialization of AI is Token, which is considered the "atomic unit" of AI, similar to user numbers in the internet era [1][3] - Google has demonstrated overwhelming advantages in AI inference traffic, processing approximately 634 trillion Tokens by Q1 2025, significantly outpacing Microsoft's 100 trillion Tokens [2][4] - The growth of Token usage is seen as a bullish signal for AI adoption, with companies like Coatue Management applying Token growth theory to investment decisions, benefiting firms like Broadcom and TSMC [3][4] Group 2 - Google's monthly inference volume surged to 480 trillion Tokens by April 2025, a 50-fold increase from 9.7 trillion Tokens a year prior, indicating a substantial scale advantage [2][4] - Cost control is a key factor for Google, with estimated inference costs of approximately $750 million in Q1 2025, representing only 1% of search revenue, even as Token usage continues to rise [7] - The growth in free AI Token usage, driven by features like AI Overviews in Google Search, is outpacing the revenue growth from paid models, with Google Search's user base being 5-6 times larger than ChatGPT's [7]