Fueled by AI enthusiasm, Taiwan's market now trades at a 32x cyclically adjusted P/E, approaching India's historically high valuation. A single company, TSMC, represents 13% of the benchmark and trades at 65x CAPE, creating significant concentration and valuation risk for the entire market.
Related Insights
Current AI-driven equity valuations are not a repeat of the 1990s dot-com bubble because of fundamentally stronger companies. Today's major index components have net margins around 14%, compared to just 8% during the 90s bubble. This superior profitability and cash flow, along with a favorable policy backdrop, supports higher multiples.
Unlike the 1990s tech bubble, today's companies have higher net margins (14% vs. 8%) and better cash flow. This, combined with a rare mix of monetary easing, fiscal stimulus, and deregulation outside of a recession, makes current equity multiples look more reasonable.
Today's market is more fragile than during the dot-com bubble because value is even more concentrated in a few tech giants. Ten companies now represent 40% of the S&P 500. This hyper-concentration means the failure of a single company or trend (like AI) doesn't just impact a sector; it threatens the entire global economy, removing all robustness from the system.
Current market multiples appear rich compared to history, but this view may be shortsighted. The long-term earnings potential unleashed by AI, combined with a higher-quality market composition, could make today's valuations seem artificially high ahead of a major earnings inflection.
The startup landscape now operates under two different sets of rules. Non-AI companies face intense scrutiny on traditional business fundamentals like profitability. In contrast, AI companies exist in a parallel reality of 'irrational exuberance,' where compelling narratives justify sky-high valuations.
Beyond financial metrics, the most significant 'tail risk' to the AI boom is the high concentration of advanced semiconductor manufacturing overseas, particularly in Taiwan. A geopolitical conflict could sever the supply of essential hardware, posing a much more fundamental threat to the industry's growth than market volatility or corporate overspending.
The stock market's enthusiasm for AI has created valuations based on future potential, not current reality. The average company using AI-powered products isn't yet seeing significant revenue generation or value, signaling a potential market correction.
The stock market is not overvalued based on historical metrics; it's a forward-looking mechanism pricing in massive future productivity gains from AI and deregulation. Investors are betting on a fundamentally more efficient economy, justifying valuations that seem detached from today's reality.
A circular economy is forming in AI, where capital flows between major players. NVIDIA invests $100B in OpenAI, which uses the funds to buy compute from Oracle, who in turn buys GPUs from NVIDIA. This self-reinforcing loop concentrates capital and drives up valuations across the ecosystem.
Ben Thompson argues that while investing in unproven fabs from Intel or Samsung seems risky, the greater risk is the entire AI industry being constrained by TSMC's singular capacity. The future opportunity cost of foregone revenue from this bottleneck far outweighs the expense of building up viable competitors.
