Singapore Central Bank Warns AI Investment Boom Could Destabilize Global Markets

A Trillion-Dollar Question

The artificial intelligence investment wave sweeping through global markets has become a double-edged sword for financial stability, according to Singapore's central bank chief. While near-term capital commitments remain solid, backed by hyperscaler cash flows and committed orders, the medium-term picture grows murkier as revenue growth struggles to justify the scale of spending on data centers and semiconductor capacity.

Chia Der Jiun, managing director of the Monetary Authority of Singapore, outlined the central tension during the release of the authority's annual report on July 28. Global growth trajectories, investment patterns, and financial market valuations have grown deeply intertwined with assumptions about continued AI capital deployment, particularly across the United States and semiconductor-exporting Asian economies including South Korea, Taiwan, and Japan.

The concern centers on a widening gap. Markets will increasingly scrutinize whether commercial revenue growth can validate the financing risks embedded in current valuations, Chia noted. That validation depends on early productivity gains at individual firms spreading across entire economies and the emergence of transformative applications that move beyond experimental phases.

Two Paths, Both Risky

The monetary authority sees two distinct scenarios unfolding, each carrying significant implications for policymakers and market participants.

In the first scenario, AI investment sustains its current trajectory as revenue growth accelerates and productivity gains broaden across sectors. This outcome would reshape income distribution, aggregate demand, and inflationary pressures, forcing central banks to recalibrate their interest rate frameworks. The mechanics of monetary policy would need to account for productivity shifts that alter the traditional relationship between growth and price stability.

The second scenario poses sharper dangers. A pullback in AI-related capital spending could trigger rapid deterioration in global growth through multiple channels: collapsing business investment, plummeting semiconductor demand, and negative wealth effects as equity portfolios revalue downward. Financial conditions could tighten swiftly as markets reassess exposure to business models that prove unsustainable without continued capital infusion.

Equity markets, credit instruments, and loan portfolios carry concentrated exposure to AI-linked companies and supply chains. A sharp reversal would test the resilience of financial institutions holding these assets, potentially amplifying losses through forced selling and liquidity strains. The speed of any downturn matters as much as its magnitude, given how rapidly sentiment can shift when growth narratives unravel.

Quantum Threats and Frontier Risks

Beyond macroeconomic surveillance, Singapore's monetary authority is intensifying operational defenses against AI-enabled threats to the financial system itself. On July 28, MAS and the Association of Banks in Singapore established the AI-Driven Cyber and Technology Risk Taskforce, drawing members from DBS, OCBC, UOB, Singapore Exchange, Nets, and Banking Computer Services.

The task force addresses an evolving threat landscape. Frontier AI models enable phishing campaigns with unprecedented personalization and scale, deploying deep-fake impersonation and customized deception tactics. Automated vulnerability scanning can identify and exploit weaknesses across systems faster than human security teams can respond.

Quantum computing introduces a separate layer of risk. As quantum capabilities mature, current encryption standards protecting financial data and communications face obsolescence. MAS plans to issue supervisory expectations later this year, establishing clear milestones and timelines for financial institutions to migrate toward quantum-resistant cryptography.

The authority is testing whether AI models trained on pooled data from multiple banks and public-private sector sources can improve detection rates for fraudulent transactions. Results are expected in 2027 and will inform decisions about establishing an industry-level utility for data sharing and AI-powered threat detection.

Since July, MAS has required key financial institutions to deploy advanced AI models that map potential attack paths, identifying routes cybercriminals might exploit to disrupt critical services or access customer information. Lessons from these exercises will be disseminated across the broader financial sector.

The Balancing Act

Singapore's approach reflects a broader challenge facing financial regulators across Asia. Economies deeply integrated into semiconductor supply chains and data center construction stand to benefit substantially if AI investment proves durable. South Korea, Taiwan, and Japan have seen surging orders for memory chips, advanced processors, and specialized components.

Yet that same exposure creates vulnerability. A reversal in spending would hit export-dependent economies with particular force, rippling through manufacturing employment, corporate earnings, and tax revenues. Financial institutions in these markets carry loan books weighted toward technology and manufacturing sectors that would face immediate stress.

The stakes extend beyond individual economies. Global financial markets have priced in assumptions about AI-driven productivity growth that remain unproven at scale. Equity valuations in technology sectors embed expectations of revenue expansion that may take years to materialize, if they materialize at all. Credit markets have extended financing on similar assumptions, creating potential for rapid repricing if confidence erodes.

Chia emphasized the need for markets and corporations to chart a sustainable course through this uncertainty. The high growth rates achieved so far could reverse sharply, with damage to both real economic activity and financial market stability. Finding that sustainable path requires continuous reassessment as evidence accumulates about whether AI investments are generating returns that justify their scale.

For now, near-term momentum remains strong. Hyperscalers continue to commit capital, and semiconductor manufacturers are expanding capacity to meet demand. The question is what happens when those committed investments complete and markets demand proof that the spending has generated lasting value. The answer will shape not just technology sector fortunes, but the trajectory of global growth and the stability of financial systems built on assumptions that have yet to be tested.