Part 4/12:
Kevin delves into the technical and economic intricacies of AI hardware investment. He explains how the depreciation schedule of chips—especially those like GPUs used for AI—impacts financial planning and investment. A typical chip with a 2-year depreciation schedule becomes "worthless" after two years, forcing companies to continually invest in new processors to maintain their technological edge.
He illustrates this with simplified depreciation models: spreading the cost of a $100 chip over four or five years reduces annual expenses and preserves financial health. However, with supply chain constraints, companies like Nvidia dominate manufacturing capacity, creating a "moat" that results in longer chip lifespans being the exception rather than the norm.