Two consequential developments on October 13 mark a new phase in the AI Supercycle. The IMF has formally identified AI-related market corrections, sovereign bond stress and cyberattacks as interconnected threats to global financial stability. Simultaneously, banking supervisors across the United States and Europe are intensifying scrutiny of autonomous AI systems.
The Great Split is entering a more dangerous financial phase: AI capital formation is supporting growth while increasing leverage, market concentration and operational interdependence.
The critical question is shifting from whether AI investment can generate productivity to whether the financial system can absorb the consequences if expected returns arrive too slowly.
The IMF Identifies a Three-Way Systemic Risk
Interconnected instability
The IMF’s new assessment connects three previously distinct sources of financial instability: disappointing AI investment returns, rising sovereign yields and disruptive cyber incidents.
The interaction is more important than any individual risk.
AI EARNINGS DISAPPOINTMENT → EQUITY REPRICING → CREDIT SPREAD WIDENING → FORCED DELEVERAGING → HIGHER FINANCING COSTS → LOWER INVESTMENT.
A correction could spread through corporate bonds, leveraged investment strategies, private credit and institutions holding concentrated technology exposures.
The IMF also identifies increasing competition between sovereign and private borrowers for global capital, while central banks continue reducing their balance sheets.
This creates an important monetary asymmetry: financial conditions could tighten sharply without an additional increase in policy rates.
The IMF’s recommendation is therefore preventive—contain excessive leverage, preserve monetary-policy credibility and strengthen financial resilience before shocks materialise.
Central Banks Move Toward Autonomous-System Supervision
Agentic AI oversight
A separate Reuters investigation published October 13 reveals that US and European banking authorities are increasing their scrutiny of agentic AI.
The Federal Reserve, OCC and FDIC are questioning banks about autonomous decision-making, human oversight, auditability and mechanisms for shutting down AI systems.
In Europe, the Bank of England and ECB are focusing more closely on operational recovery, technology-provider dependencies and the ability to restore critical banking functions after AI-enabled disruptions.
These developments have been reported through supervisory sources. They do not constitute newly announced binding regulations.
The second-order implication is substantial.
AI AUTONOMY → FASTER FINANCIAL DECISIONS → GREATER SYSTEM INTERDEPENDENCE → SHORTER RESPONSE TIMES DURING STRESS.
Supervision must increasingly address not only whether an AI model behaves reliably, but whether financial institutions can retain control when autonomous systems malfunction or are compromised.
This reinforces the significance of Singapore’s recent agentic-finance safeguards and broader AI risk-management guidelines.
The Great Split Becomes a Capital-Structure Divide
Two financial economies
The IMF findings sharpen our central thesis.
The AI Supercycle creates two increasingly different financial economies. The first consists of companies with strong balance sheets, recurring cash flows and sufficient capital to finance investment internally.
The second consists of companies dependent on external financing and valuations that assume rapid future monetisation.
As borrowing requirements increase, the distinction becomes decisive.
The next market rotation may therefore be driven less by technological capabilities than by differences in financing resilience and realised returns on invested capital.
Three second-order effects deserve attention.
First, greater AI debt issuance can increase competition for capital, although the precise contribution to sovereign term premiums remains difficult to isolate.
Second, rising financing costs may accelerate consolidation among smaller AI infrastructure providers, potentially increasing concentration.
Third, a slowdown in infrastructure investment could reduce demand-driven inflation while simultaneously weakening growth and tightening credit.
That would complicate central-bank policy: a disinflationary correction need not be economically benign.
Industrial-AI deflation remains a possible productivity outcome. It should not be confused with falling prices caused by collapsing investment demand.
What to watch now
- Credit becomes a leading indicator. Monitor AI-related bond issuance, refinancing requirements, credit spreads and private-credit exposures—not only technology equity valuations.
- Capital efficiency becomes decisive. The critical valuation metric is increasingly the spread between realised returns on AI capital and the weighted-average cost of capital.
- Central-bank financial stability becomes part of the AI trade. A disorderly AI correction could tighten monetary conditions independently of policy-rate decisions. No new AI-driven policy-rate decision or formal neutral-rate revision is established by today’s developments.
- Europe needs a coordinated capital-formation strategy. Deeper capital markets, reliable energy, compute infrastructure and industrial adoption must advance together. Otherwise, Europe risks importing higher financing costs without capturing sufficient productive returns.
- Operational resilience becomes a competitive advantage. Institutions capable of controlling, auditing and recovering autonomous AI systems should be better positioned to manage supervisory expectations and operational disruption.
The next Great Split will separate economies and companies that convert AI capital into sustainable returns from those whose dependence on rising leverage makes the Supercycle itself a source of financial instability.