AI applications are widely used among U.S. consumers, but the penetration rate of individual paid subscriptions remains extremely low. Paid user spending exhibits a steep power-law distribution, with the top 1% of paying users spending far more on a monthly average than the combined total of the bottom 50%. Currently, the landscape of leading companies is gradually becoming clear. ChatGPT leads in terms of traffic and subscription revenue, Claude focuses on professional consumers with a high proportion of premium-priced plans, and Google emphasizes creative models, showcasing significant differentiation among the three. The personal AI agent sector is rapidly emerging, with the potential to explore new business models through transaction-based commissions. Currently, consumer-grade AI primarily relies on subscription and pay-as-you-go models, contrasting with the advertising-dominated logic of traditional consumer internet. This divergence highlights a split between traffic and revenue logic. The compute market remains hot, with GPU rental prices continuing to rise despite expectations. As the cost of intelligence declines, demand for compute power continues to surge, bringing hardware and infrastructure back into the market spotlight. However, mature applications of AI in both enterprise and consumer sectors are still in their early stages. Previous bearish views on software were inaccurate, as the software sector is now entering a profit-driven revaluation and growth-proofing phase. AI is poised to expand demand boundaries across multiple fields. The gap between the rapid development of the compute side and the shallow penetration on the demand side represents both the core uncertainty in the current market and the greatest potential for future growth.