A Comparative Analysis of AI and Dot-com Eras : Involving Market Indices and Financial Fundamentals of Firms

Sammanfattning

Artificial Intelligence (AI) has attracted strong investor attention and substantial capital allocation. This raises a practical question familiar from earlier technology cycles: whether current market pricing for AI mainly reflects justified growth expectations or also contains speculative elements comparable to the dot-com era. The thesis addresses this issue by comparing market-level index behavior and firm-levelvaluation and profitability patterns across the dot-com period and the current AI cycle. Background: Technology-driven optimism can produce periods in which equity prices move ahead of visible fundamentals. The dot-com bubble is a canonical benchmark for such a cycle. The AI boom mayshare some features with that episode, but it may also differ due to firms’ maturity, profitability, andbroader adoption of the technology across sectors. Objectives: The purpose of this study is to evaluate, using historical and descriptive evidence, to what extent the current AI hype resembles the dot-com bubble in terms of (i) market-level cycle behavior in major U.S. equity indices and (ii) firm-level valuation multiples and operating profitability for AI-positioned firms relative to non-AI controls and dot-com benchmark groups. Methodology: A comparative quantitative design combines two perspectives. At the market level, the study compares the National Association of Securities Dealers Automated Quotations (NASDAQ)Composite, NASDAQ 100, and Standard Poor’s (SP) 500 across the dot-com and AI periods using nor-malized index paths; correlation and dynamic time warping distance on LOcally Weighted ScatterplotSmoothing (LOWESS)-smoothed, standardized logarithmic trends; rolling dynamic time warping sim-ilarity between the AI-era path and segment-wise dot-com cycle phases; and joint analysis of rollingvolatility and log returns with kernel density visualization. At the firm level, it compares Price to Sales Ratio (P/S), Price to Earning (P/E), enterprise value to Enterprise Value to EBITDA (EV/EBITDA), and Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) margin across definedgroups: dot-com bubble versus dot-com control firms, and AI-positioned versus non-AI control firms forthe AI era, with group rules based on archival dot-com lists and external AI-positioning scores (LarridinAI Tracker, with complementary context from adoption indices). Most firm-level multiples and relatedfields were assembled from historical records in Macrotrends, supplemented by the external trackers usedto classify AI positioning; the analysis is descriptive and does not estimate intrinsic value or causal effects. Results: Market-level evidence suggests partial similarity between the AI period and specific phases of the dot-com cycle (notably acceleration in some windows), but overall trend similarity is moderate (correlation on smoothed trends roughly 0.59–0.63) rather than an exact replay. The AI-era index path appears comparatively gradual, with episodic corrections, while the dot-com episode shows a more complete boom–bust profile in the analyzed technology-heavy indices; volatility–return patterns share a central region of low-to-moderate volatility near zero returns but differ in tail dispersion. Firm-level multiples indicate that AI-positioned firms are not valued at the extremes observed for dot-com bubble firms; gaps between AI-positioned and non-AI controls are limited, especially when valuation is related to operating earnings. EBITDA-based profitability indicators point to stronger operating performance in the AI-era sample than in many dot-com bubble firms. Recovery analysis for the dot-com crash furthersuggests that less extreme valuations were associated with faster post-correction recovery, though the AIcycle is still ongoing. Conclusions: Taken together, the evidence supports a nuanced answer to the main research question: the AI period shows some bubble-like signals and limited market-level resemblance to parts of the dot-com cycle, but it should not be characterised as a straightforward repetition of the dot-com bubble based on the indicators used. The balance of index and firm-level results is more consistent with a technology boom in which expectations are partly priced, yet are often linked to more mature firms and stronger reported operating fundamentals than in the dot-com peak. Interpretation remains subject to survivorship bias inthe historical sample, incomplete outcome data for an ongoing AI cycle, and the use of external proxiesfor AI positioning. Recommendations for future research: Future work could extend the comparison to other historical episodes, refine AI-exposure measurement, improve coverage of delisted dot-com firms where data allow, incorporate private-market AI valuations, and revisit market- and firm-level patterns as post-peakoutcomes for the AI cycle become observable.