Work

Dotcom-kupla ja tekoälyn investointisykli: arvon määrittely, markkinakäyttäytyminen ja markkinarakenne

Year: 2026
Type: dissertation
Abstract: This thesis looks into whether the current artificial intelligence infrastructure investment cycle shows characteristics of an asset price bubble per literature on bubbles. Additionally, this thesis compares this investment cycle with that of the Dotcom bubble in 1998-2001 using similar measurements. The comparison is split into three pillars adapted from literature on bubbles as well as on the Dotcom bubble. The three pillars are fundamental valuation, market cycle behaviour, and structural and institutional mechanics. The empirical analysis uses holdings data from CSRP, Compustat, WRDS Financial Ratios Suite and Thomson Reuters 13F filings, together with FINRA margin data as well as FRED macroeconomic series. Quarterly earnings reports from companies are also used in this thesis. A reverse discounted cash flow model is used to determine that the leading public AI companies require a median implied free cash flow growth of approximately 7 percent over thirty years compared to 12 percent for the Dotcom equivalent at their peak. Valuation multiples indicate that the two markets appear expensive in different ways. Trading velocity appears moderate whilst margin debt relative to GDP is close to record highs at 4 percent. Institutional ownership shows similar trading behaviour whilst short sale constraints that weighed on the Dotcom bubble are absent today. The thesis concludes that the current market is not a repeat of the Dotcom bubble and places today’s investment cycle in the earlier stages of a bubble per the Kindleberger-Minsky cycle. It also identifies debt-assisted intense capital expenditure spending for the following years as an emerging risk to valuations.
Author: Soult, Max
Language: English
Subfield: Finance
Open Access status: green
OpenAlex ID: W7213220356
Cites: 17
Cited by:
Related to: 10