AI is likely to follow the internet's path: first a bubble burst, then true mature growth. Unlike the dot-com era, however, today's AI landscape looks very different. Back then, countless .com companies were already public with little to no real revenue, and most retail investors lacked strong cash flow to weather volatility. Any shock could — and did — trigger a rapid market collapse. In contrast, very few AI companies are currently public or nearing IPO. The sector is dominated by a handful of large foundation model giants, backed by investors with deep pockets and robust cash flows. This structure makes the AI capital markets far more resilient. Even if there's a bubble, a swift collapse is unlikely. That said, we may already be experiencing a "soft landing" version of it — just not at the top. The real pressure is shifting downstream to the software companies heavily using AI coding tools. These firms are quietly undergoing a soft collapse: widespread engineer layoffs, declining revenues, and stalled growth. The flashy "AI wrapper" apps that generated buzz last year have largely gone silent in 2026. Meanwhile, AI coding — by far the largest current use case — has yet to produce breakout new applications or explosive downstream products. In short, the AI sector has already seen its correction. The risks and pain from upstream AI infrastructure and foundation model companies have simply been transferred downstream.