🔥The AI fever far exceeds any CAPEX cycle in history

📌Currently, the dot-com bubble can no longer be compared to the AI craze. According to Financial Times forecasts, CAPEX for AI software is projected to reach 9% of U.S. GDP — an unprecedented figure in Wall Street history.
📌In fact, the dot-com bubble did not account for a large percentage of capital attraction; the historical booms in the United States even far surpassed the dot-com bubble, but are less often mentioned:
🔸Railroads in the 1880s: Peak around 5–6% of U.S. GDP, the largest infrastructure cycle before AI.
🔸Electrification in the 1920s: Attracting capital into power plants, grids, electrical equipment, and industrial manufacturing.
🔸Automobiles in the 1920s: Peak around 2–3% of GDP, involving factories, steel, rubber, oil, roads, and consumer credit.
🔸Post-World War II / 2000s housing boom: Residential investment once exceeded 6% of GDP, showing that physical assets can also attract capital on a massive scale.
🔸Telecom/dot-com buildout in the late 1990s: Fiber optics, telecommunications networks, and internet infrastructure once attracted around 1% of U.S. GDP.
🔸Shale oil boom in the 2010s: Capital poured into oil drilling, mining services, and energy logistics.
📌Today, AI data centers alone account for about 2.5% of global GDP, and AI software is forecasted to take up 9% of U.S. GDP. The dot-com bubble is even less worth mentioning.