The Great Repricing of AI

“Velocity identifies the leaders. Economic capture explains the profits. Valuation determines the return.”

The AI race is no longer about who leads technologically. It is about who captures the economics—and what investors are paying for that capture.

With the U.S. 10-year Treasury yield at 5.158% at the September 25 close, the market is imposing a harder discipline. Velocity identifies the leaders. Economic capture explains the profits. Valuation determines the return.

The fundamental question: who captures the economics of AI after compute, power, security, infrastructure and capital costs are paid—and how much of that future is already priced in? Capability without monetization is optionality. Backlog without utilization is promise. Capex without excess ROIC is capital accumulation. And exceptional economics can still disappoint when valuation already discounts too much success.

1. 01 Rule 1 — High GVI + High Expectations = Valuation Sensitivity

   Where valuation already discounts sustained growth, margin expansion and near-perfect execution, the burden of proof rises. A modest operating miss can create a disproportionate equity reaction while structural Velocity remains high. Excellent businesses can still be priced for perfection. signal: The critical reading for Nvidia and Palantir.

2. 02 Rule 2 — High GVI + Strong Cash Conversion = Quality

   The strongest combination is high structural Velocity plus pricing power, utilization, free-cash-flow conversion and ROIC above WACC. Velocity becomes investable quality when it converts into cash without requiring proportionally more capital. signal: The key reading for Microsoft, TSMC and Schneider Electric.

1. 01 AI Capex Creates a Scarcity Premium

   Every additional AI system requires packaging, memory, networking, power and cooling. The scarcer the enabling layer, the stronger its pricing power. signal: AI capex is repricing the infrastructure surrounding the processor.

2. 02 Capital Cost Becomes a Competitive Advantage

   With the 10-year at 5.158%, internal cash generation and balance-sheet strength matter more. The strongest platforms can keep scaling while weaker balance sheets face higher hurdle rates. signal: The AI race is becoming partly a financing race.

3. 03 Agentic AI Creates a Security Supercycle

   More agents create more machine identities, permissions and transactions. More autonomy → greater security intensity → more recurring cybersecurity revenue. signal: That structurally strengthens Palo Alto, Microsoft and Palantir.

Board of Directors Takeaways

governance for the next economy

• Make AI ROIC a board KPI.
• Track time-to-cash, not just capex.
• Separate backlog from economic backlog.
• Treat AI governance as margin and balance-sheet protection.
• Use the 5.158% U.S. 10-year Treasury yield as the common market-rate reference—not as company-specific WACC.
• Identify your bottleneck position: are you paying the bottleneck—or becoming the bottleneck others must pay?

The Next Economy will be shaped by the companies that move fastest and capture the economics—and returns will accrue where that leadership converts into cash and valuation still leaves room for upside.