The central bank expects global financial resilience to be threatened by persistently higher cost of capital
[SINGAPORE] Around 32 per cent of Singapore-listed companies will be at risk under a severe downturn in artificial intelligence-related investment and loss of revenue along the AI supply chain, a stress test by the Monetary Authority of Singapore (MAS) has shown.
These companies would account for about 16 per cent of overall corporate debt, with the increase in at-risk firms concentrated among highly leveraged, capital-intensive firms and those reliant on working capital financing.
Smaller firms were also disproportionately represented due to their lower profit margins and thinner cash buffers, MAS said in its annual Financial Stability Review released on Tuesday (Sep 22).
The stress test subjected companies listed on the Singapore Exchange to revenue shocks of as much as 30 per cent and interest-rate shocks of up to 400 basis points, calibrated according to the individual firm’s exposure to the AI supply chain.
MAS defined an at-risk company as one with an interest coverage ratio of below one, or negative cashflow with cash buffers that cover less than six months of the shortfall.
Still, most firms were able to weather the shock, supported by their earnings and cash reserves, and corporate balance sheets are “generally sound”, the central bank said.
But it noted that the possible reassessment of AI-related asset valuations is a key risk factor – which could prompt a retrenchment in related investments, weighing on the revenues and earnings of exposed firms.
This comes as MAS expects global financial resilience to be threatened by persistently higher cost of capital as a result of AI investment, fiscal risks and recurrent supply shocks.
Higher borrowing costs raise risks across Asia
Global sovereign yields have risen across major advanced economies, raising sovereign debt-servicing costs that can spill over to firms and households, while weighing on asset valuations.
Abrupt shifts in interest-rate expectations could also trigger sharp asset repricing, amplified by elevated leverage, maturity mismatches and concentrated exposures, including in AI-related sectors, MAS said.
The rapid expansion of AI infrastructure has become an increasingly important driver of global capital demand, and should continue to support economic activity and corporate earnings.
However, current equity valuations require sustained strong revenue growth and “rather sizeable” eventual profitability of large investments in data centres and advanced semiconductors.
This means a material shortfall in earnings or expected returns could trigger a broader reassessment of AI-related valuations, with losses propagating across public equity, corporate bond and private credit markets, MAS said.
It expects higher global real interest rates to affect Asian economies unevenly.
Economies benefiting from AI-related investment and exports are better placed to absorb higher borrowing costs, but are concurrently more exposed to a pullback in the AI cycle.
Meanwhile, economies with weaker AI linkages and larger fiscal or current account deficits could be more affected by tighter global financial conditions, with portfolio outflows and currency depreciation raising risks.
For Singapore, financial conditions have been broadly supportive, supported by easing borrowing costs, tighter credit spreads and a pickup in bank credit growth, MAS said.
Corporates had stable earnings, lower borrowing costs and improved debt-servicing capacity, while firms maintained healthy debt maturity profiles and adequate liquidity buffers.
Corporate debt stood at 117 per cent of gross domestic product in the first quarter of 2026, while short-term debt made up 32 per cent of total debt, below the 10-year average.
Bonds due by the end of 2027 represented 19 per cent of outstanding corporate bonds as at August.
While household debt rose over the past year as mortgage and personal borrowing picked up, households remained resilient as financial asset growth outpaced liabilities. Debt-servicing capacity also improved as interest rates, including mortgage rates, declined over the past year.
Meanwhile, the banking sector continues to benefit from strong capital, liquidity buffers and healthy provisioning coverage. In particular, the banking system’s corporate non-performing loan ratio fell to an 18-year low of 1.2 per cent in Q2.
In the non-bank sector, insurers have remained well-capitalised and investment funds have managed liquidity risks well.
Shifting risk profiles
The risk landscape is also shifting for financial institutions in Singapore, with MAS’ 2026 survey of chief risk officers finding that cyber and operational risks have overtaken geopolitical risks as the most cited concern.
This reflects heightened worries over AI-assisted cyberattacks and concentration risks from critical third-party vendors.
The advancing capabilities of frontier AI models could enable easier vulnerability discovery and exploitation, while improvements in generative AI could lead to increasingly sophisticated fraud and social engineering.
Financial institutions also highlighted concentration risks from common reliance on shared cloud and AI infrastructure providers.
Expected advances in quantum decryption capabilities have led to concerns that sensitive, encrypted data could be pre-emptively stolen today to be unlocked in the future.
In addition, financial market stresses, including AI-driven market fragilities, have emerged as a new frequently cited category.
“In view of the uncertain macroeconomic outlook, firms and households should manage their finances prudently and maintain adequate liquidity buffers against potential stress,” MAS said.
