AI Chip Race: Bubble Fears Mount Over Obsolete Tech

The artificial intelligence boom has spurred unprecedented investment, with the tech industry committing around $400 billion this year to specialized chips and data centers. However, a critical question looms: how long will this cutting-edge technology remain relevant? Analysts are increasingly voicing concerns that overly optimistic estimates of chip lifespans could be leading the industry towards a potentially brutal and costly AI bubble.

Previously, cloud computing giants assumed a six-year lifespan for their chips. Yet, with chipmakers like Nvidia rapidly releasing more powerful processors, this assumption is now in doubt. Some experts, including those from Princeton University, suggest a realistic lifespan of just two to three years due to wear and tear and swift technological advancements. This accelerated obsolescence means chips can lose up to 90 percent of their value within three to four years, a stark contrast to earlier projections.

The implications of this rapid depreciation are significant, especially for AI specialists such as Oracle and CoreWeave, which are heavily indebted and expanding their chip acquisition. If these companies are forced to shorten their depreciation timelines, it could directly impact their profitability and increase the cost of raising capital, particularly as some loans use the chips themselves as collateral, creating a precarious financial situation.