MAS warns AI retreat could unsettle markets
The Monetary Authority of Singapore said uncertainty over whether companies can sustain vast spending on data centres, advanced chips and computing infrastructure has become a significant risk to the economic outlook. The warning reflects growing scrutiny of whether revenue generated by AI services can justify the scale of capital committed by technology groups.
MAS Managing Director Chia Der Jiun said the global economy should continue to benefit from AI-related expenditure in the near term, but the longer-term impact remains difficult to assess. A sudden reduction in investment could affect manufacturing, trade and financial asset prices, particularly in economies closely linked to electronics supply chains.
The assessment carries particular significance for Singapore, where technology-related production has driven much of the economic expansion this year. The economy grew by about 6 per cent during the first half of 2026, while AI-linked electronics accounted for more than 70 per cent of export growth. Semiconductor production, data-centre equipment and related services have helped offset pressure from trade restrictions, higher energy costs and geopolitical instability.
MAS expects economic activity to remain firm during the second half of the year, supported by technology, construction and financial services. However, it cautioned that tighter global financing conditions or a reversal in AI expenditure could undermine the durability of that growth and ease inflation more rapidly than expected.
The warning comes as the world’s largest technology companies raise spending on processors, servers, power systems and data centres. Capital expenditure by major AI-linked companies is projected to rise by more than 75 per cent to around $700 billion in 2026. Large technology groups are also turning increasingly to debt markets to finance the buildout, strengthening the connection between AI investment, credit conditions and the wider economy.
Fitch Ratings has identified an AI market correction as an emerging global credit risk, citing elevated valuations, concentrated investment and uncertainty over future returns. Amazon and Alphabet are among companies whose combined bond issuance has reached about $182 billion, underlining the growing role of borrowed funds in supporting infrastructure spending.
Supporters of the investment surge argue that generative AI could raise productivity, reduce costs and create new markets comparable to earlier technological transformations. Demand for advanced chips and cloud capacity remains strong, while business investment data continue to show momentum. US orders for core capital goods rose 0.9 per cent in June, and shipments increased 1.9 per cent, their strongest gain in four and a half years, helped by spending on computers and electronic equipment.
Questions persist over how quickly that spending will generate sustainable profits. Data centres require large quantities of electricity, specialised chips and cooling infrastructure, while rapid improvements in computing equipment can shorten replacement cycles. Companies may face pressure to keep investing even before earlier projects achieve their expected returns.
Academic research indicates that the AI boom combines strong technological foundations with pockets of speculative vulnerability. Enterprise adoption, revenue growth and productivity improvements support parts of the valuation expansion, but investment has outpaced monetisation in some segments. Private-market valuations are also concentrated among a limited number of companies, increasing the risk that a reassessment of growth expectations could spread rapidly.
Global equity markets are already showing greater sensitivity to concerns about returns from AI spending. Semiconductor shares have faced selling pressure, while technology indices have retreated from earlier peaks. Investors are closely examining earnings from leading cloud and chip companies for evidence that demand remains strong enough to support infrastructure budgets.
Singapore is vulnerable to such shifts because of its position in global electronics, finance and trade networks. A decline in chip orders could affect factory output and exports, while a fall in technology valuations could weaken investment sentiment and household wealth. Financial institutions could also face losses through corporate lending, bond holdings and market exposure.
MAS has separately highlighted cybersecurity threats arising from AI and quantum-computing advances. AI tools can improve fraud detection and network monitoring, but they can also make phishing, impersonation and hacking attempts more convincing. The authority has established an industry task force to help financial institutions strengthen their defences as technology-driven risks become more sophisticated.
