AI Buildout Drives Yields and Market Rotation

Last week’s market action reinforced a critical market driver: the AI buildout has become one of the most important forces influencing not only technology stocks, but also long-term interest rates and the relative performance of growth and value. CPI came in above expectations, and I thought the initial market reaction was substantially overdone. The 10-year Treasury yield jumped roughly 10 basis points and the implied probability of an October Fed hike moved close to 50%. But as concerns about the sustainability of AI spending intensified, technology shares weakened sharply and bond yields fell back below where they had been before the inflation report.

That tells us something important. The biggest driver of long-term rates right now is not oil and may not even be what the Federal Reserve does at its next meeting. It is the enormous demand for capital associated with the AI buildout. If that spending continues at its current pace, stronger growth and greater demand for funds puts upward pressure on long-term yields. If investors begin to question the scale or durability of the buildout, those yields can come down very quickly. We saw exactly that dynamic last week, with a sharp rotation out of growth and into value as the Nasdaq weakened while the Dow and value stocks moved higher.

This is also why diversification remains so important. I am not predicting anything resembling the technology collapse of 2000. Today Nasdaq trades at roughly 22 to 24 times forward earnings compared with approximately 100 times at the peak of the internet bubble. Those are completely different valuation environments. But if AI spending or token demand disappoints, I could certainly see a 20% to 30% correction in Nasdaq. Because that would also bring long-term interest rates down, value stocks should hold up much better, and the overall S&P 500 could decline only 5% to 10%—a normal pullback rather than a bear market.

The key question for me is whether all this AI investment begins to show up in productivity. Hyperscalers are spending aggressively, and companies are spending heavily to use these models, but so far, I do not see a broad productivity acceleration in the economic data. Firms may be maintaining margins, but the average non-AI centric companies are not yet showing the type of margin improvement that would justify the full scale of the investment. That productivity payoff needs to emerge.

Meanwhile, the broader economy remains very healthy. Third-quarter GDP appears headed for a three-handle, according to both the Federal Reserve and private-sector estimates cited in our discussion. Weekly jobless claims remain near the low end of their range, and the labor market continues to look strong. Importantly, money growth has also become more consistent with sustainable nominal economic growth. Deposits have risen about 8% since the start of the war, but since June the pace has slowed to roughly 5.5%. I view that moderation as a positive development for the inflation outlook.

The latest CPI and PPI reports were not good, and estimates for the upcoming core consumption deflator have moved higher. But I still do not expect the Fed to raise rates in October. There is enough evidence of slower money growth and enough time for the inflation data to improve for the Fed to remain patient. December is less certain, and I would currently put the probability of a hike extremely high, with two more months of data still to come.

Earnings now have become the central market focus this week. Bank results have been somewhat mixed, and the market will need confirmation that profits outside the largest technology companies are strengthening. I still do not believe the overall equity market is excessively valued, but investors should expect continued volatility between growth and value as expectations for AI spending shift. The essential point is that an AI disappointment would not affect every asset class in the same way. Technology could correct sharply, but lower bond yields would provide a powerful counterweight for value stocks and Treasuries. In this environment, diversification is not simply defensive—it may be one of the most effective ways to navigate the next phase of this market.