Nvidia vs. Cisco: Parallels Between the AI Era and the Dot-ComNVIDIA CorporationBATS:NVDAnothingchangehere Before the dot-com bubble peaked in 2000, Netscape's IPO marked the dawn of the software wave. What followed was a five-year upward impulse driven by internet hardware, with Cisco as the flagship. The bubble in hardware eventually topped in March 2000. Today, Anthropic's IPO is expected around November. If history rhymes, does this signal that the AI bubble could similarly peak roughly five years from now — around 2031? That's my first thesis. My second thesis involves a relative valuation comparison. I charted Cisco's market cap as a percentage of global gold market cap at its 2000 peak, then overlaid Nvidia's same ratio today. While the underlying data may not be precise, the exercise serves as an early warning framework — a way to gauge whether current AI exuberance has reached historical extremes. Finally, a lesson from 2000: when the internet bubble burst, many healthcare and pharma names posted strong gains. Think LLY, UNH, VRTX, PFE, MDT, ABT. My view is that even if the AI cycle hasn't peaked, we are very likely in the second half — and in this second half, AI applications are proliferating across sectors. Large language models are reshaping industries at a structural level, with particularly clear impact on biopharma and manufacturing, traditional robotics, and emerging fields like brain-computer interfaces. These are application-layer segments where LLMs are likely to deliver the earliest and most tangible transformations. If the second half of the AI cycle is indeed application-driven, these domains should outperform. That's why I'm leaning toward sectors like AI-powered drug discovery and healthcare, AI-enabled robotics, and AI safety — as both growth opportunities and portfolio diversifiers alongside selective AI hardware exposure.