“The next Great Split separates economies that finance intelligence into productivity from those that turn it into a higher cost of capital.”
The AI Supercycle is reshaping not only the demand for capital, but the mechanisms through which central banks control its price and markets distribute its risks.
Three developments between November 4 and 5 reveal a new phase in the Great Split. The central question is changing: no longer simply how AI affects inflation and productivity, but whether monetary authorities can preserve control over financial conditions as capital formation concentrates, intermediation migrates outside banks, and increasingly autonomous systems join financial transactions.
The Savings Glut Has Become A Bond Glut
Schnabel, November 5
Schnabel's November 5 speech, Navigating Stormy Waters, identifies a structural change in the balance between savings and investment. AI infrastructure, digital transformation, energy, defence and government financing are raising demand for capital, and the resulting debt issuance is producing what she calls a bond glut — upward pressure on the natural rate r, even before inflation shocks. She announces no new numerical estimate.
The assumption that technology inevitably means cheaper money is incomplete. AI capital formation raises financing requirements and potentially neutral rates now; industrial AI adoption lowers unit costs and may disinflate later. The central risk for Europe is that the global cost of capital rises faster than firms can deploy technologies that improve returns — so the Great Split may emerge through the discount rate before it shows in productivity statistics.
The US Treasury Challenges The Crowding-Out Thesis
TBAC, November 4
The Treasury Borrowing Advisory Committee's November 4 report finds no evidence that AI-related borrowing has weakened demand at long-term Treasury auctions. To the extent technology investment is moving yields, stronger economic growth is more convincing than direct displacement. It is not a Fed decision.
The report flags widening sector spreads, financing via GPU-backed loans and data-centre structures, and international flows strengthening the dollar. The squeeze is not uniform: the largest firms self-finance through cash flow, equity and investment-grade debt, while less established infrastructure firms depend on secured lending and refinancing. Capital formation is advancing faster than the diffusion of its labour-market benefits.
Monetary Policy Hostage To Non-Bank Finance
Schnabel's second warning
Non-banks now hold assets comparable in scale to banks, while euro-area public debt has risen roughly 20 percentage points of GDP since the early 2000s — much intermediated by leveraged funds reliant on short-term financing. The sequence: tightening → higher funding costs → non-bank deleveraging → sovereign-bond volatility → liquidity pressure → tighter corporate financing. A disorderly repricing of AI-related credit could coincide with sovereign-funding stress.
This is a risk scenario, not a crisis underway. Schnabel advocates stronger oversight of non-banks rather than weaker bank safeguards, and raises the possibility that more robust regulation could support access to central-bank liquidity.
Autonomous Finance: Preparing For AI-Agent Payments
ECB and OCC
Executive Board member Piero Cipollone outlined the Eurosystem's digital-payments strategy. The digital euro innovation programme is examining AI-agent payments and micropayments — exploratory, not approved. The pilot is planned for H2 2027 with 36 payment service providers selected, and the Pontes infrastructure recently settled a €1 billion ESM tokenised benchmark-bond in central-bank money.
On November 4 the OCC issued its Super Intelligence: Supervisory Framework, covering agentic AI. Its principle: attention should reflect the consequences of an AI system's actions, not sophistication alone. It is examiner guidance, not binding rules — but it adds competitive differentiation for European banks that can deploy AI agents with human accountability and transaction controls.
The Great Split — Three Scenarios
the test is cash flow
Productivity-Led Capital Expansion
AI investment stays strong and productivity spreads. Profitability strengthens, financing remains accessible, and higher equilibrium real rates are supported by sustainable growth.
Higher-For-Longer Capital Costs
Energy constraints, public borrowing and AI capex sustain financing demand while productivity stays concentrated. Spreads widen and weaker balance sheets postpone investment — deepening the split.
Financial Repricing And Monetary Constraint
Disappointing AI returns trigger a correction in technology equities and credit. Leveraged non-banks face funding pressure, sovereign markets stay vulnerable, and central banks confront inflation alongside instability.
position for the transition
- Reassess the neutral-rate assumption; test long-duration valuations across real-rate scenarios.
- Distinguish sovereign crowding-out from corporate capital competition.
- Treat non-bank leverage as a critical transmission channel.
- Position for the move from AI infrastructure to Industrial AI productivity.
- Accelerate Europe's financial integration and a common corporate framework.
What Comes Next — The Monetary Test
the immediate events
The Bank of England's November 5 decision, due at noon London time, is unpublished at this assessment. The Bank Rate is 3.75% and markets widely expect 4.00% — a forecast, not an announcement. This morning Norges Bank held at 4.50% while preserving the option of further tightening. No new AI-specific framework or neutral-rate estimate has been announced by the ECB, Fed, BoE, MAS or SNB.
The next Great Split will not simply separate those who possess intelligence from those who do not. It will separate economies capable of financing intelligence into productivity from those whose financial structures turn technological opportunity into a higher cost of capital.