Weekly Commentary (10/13/25) – Is AI a Bubble? Yes. Now Here’s the Real Risk.

By the numbers:
Last week, markets pulled back. The S&P 500 fell roughly 2.7%, the Nasdaq dropped about 3.6%, and the Dow lost around 1.9%. The 10-year Treasury yield finished near 4.05%. Oil slid to a five-month low near $59 (WTI). Gold briefly crossed $4,000 and remained elevated. Investors rotated defensively as policy uncertainty, trade tensions, and the shutdown weighed on confidence.
________________________________________
1) Current State
People keep asking if AI is a bubble.
It probably is.
But bubbles rarely pop just because stocks look expensive.
They pop when confidence breaks.
That is why this government shutdown matters.
It is not happening because the economy is weak—it is political.
But it still takes income away from workers, slows services, and, more importantly, shuts off data.
Markets can handle bad news.
They hate no news.
________________________________________
2) Perspective
Let’s start with AI.
One of the biggest drivers of this market is not just customer demand—it is circular financing.
AI companies invest billions into each other, often in exchange for equity or compute credits. Those investments show up as revenue. Higher revenue boosts valuations. Higher valuations attract more capital. The capital gets reinvested again.
It looks like growth.
Often it is just the same dollars moving in a loop at the top of the market.
When capital, revenue, and valuation all feed each other, the system becomes extremely sensitive to any hit to confidence.
Now layer on the macro environment.
The shutdown makes the economy harder to read.
The September jobs report was already collected before the shutdown. October is at risk. If data collection is delayed or replaced by estimates, no one can clearly see what is happening. When decision makers cannot see, they hesitate.
The labor market looks stable—until you look deeper.
Hiring has been soft all year.
Wage growth is still elevated, which sounds good, but high wage growth with weak hiring often means companies are cutting lower-level roles and only paying up for top talent. Think fewer entry-level jobs, fewer backfills, and more dollars concentrated at the top, especially in AI.
That combination leads to:
• Slower participation
• Weaker on-ramps for young workers
• Rising youth unemployment, which has historically been a leading indicator of broader labor weakness
Immigration adds another pressure point.
A large share of high-participation, high-efficiency labor in the U.S. comes from immigrants. When enforcement headlines increase, even in small numbers, the chilling effect is large. Participation falls, productivity suffers, and labor shortages appear in critical industries. This is not a political issue—it is a math and labor supply issue.
Participation is growth. When participation drops, the engine loses power.
________________________________________
3) Where do I find Confidence?
This is not 1999.
Valuations are high, but nowhere near dot-com extremes.
History is clear: even if you invest near a market top, long-term returns are overwhelmingly positive if you stay invested.
Calling recessions rarely helps. There is usually a lag between the peak and the official recession announcement. Trying to time it feels smart—it almost never improves outcomes.
The bigger risk is losing confidence in the system itself.
Confidence relies on:
• Reliable data
• Steady labor participation
• Efficient capital allocation
• Predictable policy
When data goes dark, when participation slips, or when capital is recycled instead of earned, confidence erodes.
________________________________________
4) What to Consider
• Stay exposed to innovation, but not through only ten companies. Diversification matters more than ever.
• Watch the labor market the right way: participation, job quality, and youth unemployment trends.
• Treat the shutdown as a confidence and data risk. If it ends quickly, it is noise. If it drags, expect more volatility and less clarity.
Either way, do not let short-term events rewrite long-term portfolios.
If bubbles are part of markets,
discipline must be part of the portfolio.
Own what wins over ten years.
Refuse to give it back in ten weeks. That is the real risk you can control.