The Great Split Enters Price Discovery

“The next winners will turn falling computing costs into rising returns on capital.”

AI is sustaining global growth while financial markets begin repricing its infrastructure, financing requirements and competitive advantages. The next market rotation may already be underway.

Three developments overnight mark a potentially decisive transition in the supercycle: Asian semiconductor shares are falling sharply as advances by China's memory-chip industry challenge durable pricing power; Tencent is launching a multibillion-dollar bond to finance accelerating investment; and the IMF warns that AI-related leverage and sovereign borrowing costs could amplify the next financial shock.

The contradiction is no longer between technological acceleration and financial restraint. It is between the extraordinary growth AI creates and the increasingly contested distribution of the returns it generates. The market is beginning to separate three forms of value — technological leadership, infrastructure pricing power and sustainable financial returns — and that distinction is becoming the organizing principle of the next rotation.

1. 01 The semiconductor selloff challenges the scarcity premium

   Asian chip stocks posted their sharpest collective losses in months — Samsung near −8%, SK Hynix near −10%, TSMC −1.8%, KOSPI more than −6% — after reports of memory-density progress at China's ChangXin Memory Technologies. The reports do not establish parity with leading suppliers, yet the prospect alone triggered substantial repricing. The transmission is clear: technological catch-up points to anticipated capacity competition, lower expected scarcity rents and margin revisions. The paradox: lower memory and computing costs could eventually accelerate AI adoption downstream. signal: What damages semiconductor valuations today could accelerate AI adoption tomorrow.

2. 02 Tencent is testing the price of AI capital

   Tencent is marketing dollar and offshore-yuan bonds to raise at least $4.5 billion, with initial dollar guidance around 115–140 basis points over Treasuries. First-half capex reached RMB 96.3 billion, more than double a year earlier. AI infrastructure is increasingly financed through global credit markets rather than internal cash flow, introducing a new determinant of advantage: the financing cost of intelligence. Access to competitively priced long-duration capital now matters as much as access to advanced computing. signal: AI financing is becoming a multicurrency contest measured in spreads and maturities.

3. 03 The IMF confirms the Great Split — and warns about it

   The IMF raised 2026 global growth to 3.1% (from 3.0%), keeping 2027 at 3.4%, crediting sustained AI investment. But just 32 countries now account for 95% of global growth, down from 54 in April. The eurozone was lifted to 1.1%, the U.S. to 2.3%. Its Global Financial Stability Report warns that AI-related concentration and leverage could amplify a sharp correction — turning an industry-specific revision into a broader market shock. signal: Resilience is real, but generated by a narrowing group of countries and industries.

4. 04 Energy and sovereign debt are limiting the room for error

   Brent holds around $105–106 per barrel amid Middle Eastern and Russian-supply risk, while the U.S. 10-year yield has fallen toward 5.19% on safety demand. But lower yields in a risk-off episode do not signal structurally easier financing. Fiscal requirements remain substantial and investment needs for AI, defense and grids keep expanding. For Europe — with France's sovereign difficulties in focus — the danger is that financing constraints delay the very investments needed for future productivity. signal: Temporary Treasury rallies are not permanently easier financing conditions.

5. 05 Bank earnings will test whether the capital collision is spreading

   JPMorgan, Goldman Sachs, Citigroup and Wells Fargo report today. The decisive readings are funding conditions — deposit costs, credit selectivity, provisions — not headline profits. The IMF's stress tests show banks holding 22% of global assets with weak capital under the adverse scenario, up from 18%. As banks finance complex AI infrastructure, credit becomes tied to electricity, utilization and obsolescence — closer to financing rapidly evolving industrial assets than conventional lending. signal: Advantage belongs to lenders who can distinguish strategic from economically productive infrastructure.

The five developments describe one process: a redistribution of economic advantage. First-order, AI investment expands and supports growth. Second-order, competition challenges pricing power and investors reassess the durability of projected cash flows. Third-order — the most important — capital shifts toward companies that generate productivity gains with lower incremental investment.

That could drive a rotation from AI infrastructure ownership toward AI-driven productivity. Three distinctions will define performance: scarcity rents versus sustainable margins; capital expenditure versus financial productivity; technological leadership versus economic capture. The next rotation may not move from growth to value, but within the AI ecosystem itself — infrastructure suppliers facing margin normalization, enterprise adopters generating measurable productivity. AI exposure is no longer enough; AI economic capture becomes the decisive variable.

The Five Morning Takeaways

October 13

• The semiconductor selloff is a warning about margins, not necessarily demand — Chinese progress challenges the durability of scarcity rents.
• AI is becoming a major force in global credit allocation; Tencent's bond confirms the growing importance of financing structures and risk appetite.
• The IMF validated AI's growth benefits while exposing their concentration in a narrower group of countries and industries.
• Sovereign borrowing and energy costs remain structural constraints; temporary Treasury rallies should not be mistaken for permanently easier financing.
• The next rotation may move from AI infrastructure to AI productivity — advantage will depend on financial returns, not technological exposure alone.

The next great rotation will not be decided by who controls the most expensive intelligence. It will be decided by who creates the greatest financial value as intelligence becomes cheaper.