Bubble or nothing? Looking at the AI boom as part of a cycle

[SINGAPORE] Even if artificial intelligence-related market activity is unlikely to be a bubble, as the Monetary Authority of Singapore (MAS) seems to suggest – much less one on the verge of popping – Singapore’s economic strategy must still be built to accommodate its regular ups and downs.

In its latest macroeconomic review on Wednesday (Oct 14), Singapore’s central bank assessed that “downside risks from an AI-cycle downturn” do not appear immediate.

Yet, the expectation that AI-led economic growth is cyclical still behooves planners to prepare for when the wheel of fortune inevitably turns – even if they do not think it imminent.

Shaking off shocks

MAS signalled that a downturn could be caused by “an AI infrastructure credit shock”, in which AI adoption fails to deliver expected profits, tech and data centre giants cut back on spending, and businesses have trouble financing the debts that fund AI-related projects.

But companies can get ahead of this risk by addressing the most fundamental part of the equation: the strategic deployment of AI to develop and enhance viable business cases.

In other words, it redounds to no good if AI adoption manifests solely as “wasting money, spending tokens”, as Broadcom CEO Hock Tan colourfully put it in a panel on Monday.

Though Singapore is already a leader in AI adoption relative to many other economies, this reminder offers industry practitioners a chance to take stock of how well the technology has actually been employed and integrated into their work processes beyond the superficial.

In this regard, Singapore and other governments also have a clear role to play in ensuring that the business and societal take-up of AI is intentional and targeted, instead of willy-nilly.

Rather, investment in AI must be directed towards technology with demonstrable benefits for companies and workers alike. Business and political leaders must thus resist any indiscriminate spending on AI that could be foisted upon them by firms with vested interests.

Meanwhile, MAS has forecast Singapore’s economic growth to moderate “as the AI cycle eases from its exceptional pace and global financial conditions tighten”.

Given that credit risk is inherent to the concept of an AI shock, prudence on this front is warranted as well. Such a consideration surely influenced MAS’ latest decision to stand pat on monetary policy after two rounds of tightening.

Its stance – a notable contrast from the interest rate hikes other central banks have recently undertaken – suggests a considered position against the possibility of overcorrection.

Riding the wave

AI-driven activity has propelled above-forecast growth for Singapore in 2026.

Third-quarter gross domestic product was spurred by factors such as higher electronics manufacturing output and stronger electronics exports in the wholesale trade sector, said the Ministry of Energy, Trade and Industry.

MAS still expects trade and technology-related sectors to “remain the principal drivers of growth” in 2027, despite moderation in the global AI investment cycle.

Still, in the medium term, economic projections need to compensate for the high base effect set by the sterling AI-driven growth in recent quarters.

That’s as DBS analysts Chua Han Teng and Philip Wee predicted that modern services – which comprise finance and insurance, information and communications, and professional services – will “remain the resilient anchor of Singapore’s economy”.

Under this model, growth comes on the back of continued AI adoption and digitalisation – but in tandem with other drivers, such as Singapore’s status as a mature financial hub.

In the longer term, the economy must stay broad-based, with multiple pillars that do not rely on direct or indirect uplift from AI. And workforce transformation has to be part of this story.

Standard Chartered economists Jonathan Koh and Edward Lee remarked in a report on Wednesday that “the AI-led expansion has been relatively capital and productivity-intensive, allowing output to rise without a commensurate increase in labour costs” in Singapore.

But allowing AI to set a moving target of infinite productivity is a fool’s errand.

Instead, policymakers must continue to guarantee a structural shift towards higher-value jobs – which can sustain healthy wage growth and meet the rising aspirations of the citizenry.