Many people say that Cursor's cash flow is good right now, but pay attention to a few details in the article: 1. Cursor's personal subscription profit margin is negative, while the enterprise package's profit margin is positive. 2. As of last November, enterprise contracts accounted for only 13.6% of Cursor's annualized revenue, but now about 60% of Cursor's revenue comes from enterprise clients. 3. There are signals of customer churn; pressure is continuously building up. In other words, Cursor's main profit source is still enterprise users. A characteristic of enterprise users is that they switch products very slowly, and it's quite normal for them to lag behind popular products used by individual users by six months to a year. For a slightly larger company, choosing which product to use involves a lengthy process. Many companies adopted Cursor earlier, and even if employees are now using Claude Code more, enterprises won't immediately switch to Claude Code due to inertia; there will still be a long process. Only after this process is completed will they completely unsubscribe from Cursor and switch to Claude Code, and that will be the real danger for Cursor. Of course, Cursor itself is also actively seeking change, and it is quite possible that they will find a good ecological niche in the future. In the AI era, everything changes too quickly.