Repricing the Cost of Capital

“The first AI trade repriced intelligence. The second repriced infrastructure. The third will reprice the cost of capital.”

Markets received what they wanted today. U.S. producer prices were unchanged in July, strengthening the argument that the inflation impulse is moderating. Treasury yields declined, the dollar weakened, technology rallied and the S&P 500 moved to another intraday record.

But the important signal was not the index move. It was the change underneath it — the probability assigned by markets to the Federal Reserve holding rates unchanged in September rose to roughly 65%, from about 50% before today's producer-price release.

Equity markets continue to celebrate AI. Credit markets are beginning to ask who must finance it. That distinction matters enormously.

1. 01 Treasury yields

   Today's decline matters because AI is extraordinarily capital intensive. Lower long-duration yields simultaneously support technology multiples and reduce the financing burden of the physical infrastructure behind AI. signal: Lower yields ease both valuations and the cost of building AI.

2. 02 Credit versus equity

   Watch whether the divergence between equity optimism and credit caution persists. If credit spreads begin differentiating aggressively between AI-relevant and AI-exposed-but-overleveraged borrowers, the next Great Split has begun. signal: Credit differentiation marks the start of the next split.

3. 03 Semiconductors

   The semiconductor index had fallen roughly 29% between late June and late July, yet stronger infrastructure earnings have revived confidence that AI spending remains constrained more by supply than by demand. Nvidia's August 26 earnings are becoming the next major system-wide test. signal: Nvidia's August 26 print is the system-wide test.

4. 04 Oil

   Brent fell toward roughly $87 today despite continuing geopolitical uncertainty. Lower energy prices simultaneously reduce inflation pressure and improve the economics of energy-intensive infrastructure investment. signal: Cheaper energy aids both inflation and infrastructure economics.

5. 05 Physical AI infrastructure

   Watch engineering, construction, electrical equipment, cooling, power, memory, optical networking and logistics. One striking example today: Larsen & Toubro announced a major order to construct an Nvidia B300 AI Factory infrastructure project in Chennai. signal: The AI trade is propagating outward — that is what it looks like.

Key takeaways

the AI trade becomes a capital-formation cycle

• For three years markets treated AI primarily as a question of technology leadership; that framework is becoming obsolete.
• AI is becoming a mechanism for reallocating capital across the entire economy — the unit of analysis moves from stocks to cash flows to credit to infrastructure to industries, cities and countries.
• Once AI changes expected cash flows, it changes credit risk; once it changes credit risk, it changes the cost of capital; once it changes the cost of capital, it changes investment.
• The ultimate winner may not be the company producing the most intelligence, but the company, network, city or country able to turn intelligence into productive capital fastest.