Is AI Really a Bubble?

Is AI Really a Bubble?

Is the AI Market Really a Bubble?

"The End of the Capital-Light Era and the Shift in Investment"

 

  • The essence of the recent 'AI bubble' argument is not simply that stocks are overvalued. It lies in the structural change in which the 'growth without investment (Capital Light)' model that big tech has enjoyed for the past 10 years is shifting to a 'high-cost structure (High Capex)'.

 

 

1. The Core Basis of the Bubble Argument: The Efficiency Paradox

"Astronomical Spending, Falling ROE"

 

  • The collapse of the old formula for success: Apple, Google and others have generated enormous cash with minimal investment by outsourcing production and building on existing networks. (The secret behind Apple's 171% ROE)

 

  • Today's game of chicken: To compete for AI dominance, companies must pour astronomical sums into data centers and power facilities. The projected 2026 capital expenditure of the four big tech companies (Amazon, Meta, Google, MS) alone reaches 660 billion dollars.

 

  • Shrinking shareholder returns: Since the money earned must all go into capital expenditure (Capex), the capacity for share buybacks and dividends, which drove share price gains, inevitably shrinks. This means a decline in capital efficiency (ROE).

 

 

2. Cracks in the Market: Signs of a Change in Market Leaders

"The Breakup of the M7 and the Rise of Traditional Industries"

 

  • Differentiation within the M7: The era when the 'Magnificent 7' all rose together is over. 
    - Winners: Alphabet (+65.9%), Tesla and Apple, which have proven real results 
    - Losers: Microsoft (-25.5%), Meta and Amazon, which bear heavy investment costs

 

  • The great migration of capital: The market is already leaving high-cost tech stocks and taking refuge in traditional industries (Dow Jones +13.7%) and small and mid caps (Russell 2000 +21.3%). The string of record highs for Caterpillar, Walmart, financial stocks and others proves this.

 

 

3. Strategy Going Forward: Separating the Wheat from the Chaff

"Rather Than Selling Unconditionally, Find the 'Far End of the Beneficiaries'"

 

  • Strategy A. The 'Picks and Jeans' Strategy (Infrastructure): When big tech spends money, you should focus on the companies that earn that money. The power equipment, cooling systems and construction machinery (Caterpillar, etc.) sectors, which are essential for building data centers, are the real beneficiaries.

 

  • Strategy B. Hedging with Traditional Blue Chips (Value): You should increase the portfolio's stability by raising the weight of traditional industrials with good cash flow (retail, finance, healthcare), which are highly defensive when the AI bubble bursts.

 

  • Strategy C. Selecting 'Real AI' Companies: Exclude companies that merely wear the AI theme, and narrow your focus to companies that achieve both real revenue growth and cost reduction through AI, like Alphabet or Tesla.