Chris Wood of Jeffries predicts massive capital destruction in AI due to a shift towards cheaper Chinese open-source models. Ed Dowd agrees with the overbuild and competitive pressure but adds that Western open-source models are improving. Dowd highlights a historical pattern of technological bubbles: excitement leading to overinvestment, financial collapse, distressed assets, and then recapitalization fueling future growth. He believes the current AI boom is in the overinvestment phase, with early signs of cracks appearing. Past examples like the US Railroad Bubble, the 1920s Utility Bubble, and the Dotcom Bubble illustrate this cyclical pattern of innovation and subsequent market correction. These bubbles involved speculative financing and aggressive expansion beyond immediate demand. Following the collapse, distressed infrastructure was acquired cheaply, forming the basis for future economic expansion. Dowd sees parallels in the massive capital flowing into AI data centers, chips, and equipment, with AI models becoming commoditized. He points to stalled private credit funding, weakening enterprise demand, power constraints, and open-source pricing pressure as emerging cracks. Circular deals and companies unable to cover compute commitments under realistic revenue further indicate stress. Credit markets historically end these infrastructure buildouts by ceasing to fund expansion. Dowd suggests that while equity markets might lead or lag, credit markets are likely to signal the end of the AI buildout. Ultimately, Dowd concludes that understanding these cycles of mania, overinvestment, and recapitalization is crucial for navigating market opportunities and risks.