Is AI on the verge of a bubble?

ARTIFICIAL intelligence (AI) is set to reshape industries in ways that many are only beginning to grasp.

From factory floors to finance desks, the technology is increasingly embedded in the backbone of production and services, promising efficiency gains that were unimaginable just a decade ago.

Yet as AI-related stocks continue their meteoric rise, investors are asking whether this rally is a harbinger of a bubble – or a genuine revaluation of the economy’s future.

RHB Research offers its take on the question.

“We do not recognise a bubble (yet), and we like AI-related industries such as manufacturing, technology, data storage, finance and healthcare,” RHB Research group chief economist and head of market research Barnabas Gan says in a recent report.

“Our conviction is that AI is here to stay and will likely be a revolutionary force reshaping manufacturing, services, and everyday life,” he writes.

The beneficiaries of this transformation, according to Gan, extend far beyond the usual software giants to include data centres, semiconductors, advanced manufacturing, automation and next-generation communications technologies.

At the same time, it is cautious about the broader economic repercussions.

“However, the repercussions – particularly for labour markets, wealth distribution, investor confidence, and liquidity – are complex and potentially destabilising,” Gan warns.

No evidence of bubble

To gauge whether the current AI rally resembles a classic bubble, RHB Research employs rigorous statistical approach, specifically the Generalised Supremum Augmented Dickey-Fuller test (GSADF). This test may be used to detect the presence of speculative bubbles or periods of explosive asset price behaviour within a specified time series.

Using the GSADF study to detect the presence of an AI bubble over the period 2023-2025, RHB Research says: “The statistical result suggests that we fail to reject the null hypothesis, indicating no evidence of a bubble.”

It notes that the study uses the Nasdaq CTA Artificial Intelligence and Robotics Index, which tracks companies involved in AI and robotics across technology, industrial, medical, and other sectors.

Nasdaq classifies these firms as “enablers, engagers or enhancers” capturing the range of roles AI plays in modern business. Daily data over the three-year period, RHB Research notes, shows the rally is already in its third year.

The brokerage is quick to emphasise the inherent difficulty of labelling any fast-moving market phenomenon.

“Determining whether an asset price rally constitutes a bubble is, by nature, an exercise fraught with uncertainty,” it argues.

Bubbles often take years to form and can persist far longer than expected, propped up by narratives that feel coherent at the time.

In hindsight, what looks like a speculative frenzy can appear in real time as a rational market response to structural change.

“In real time, what later becomes labelled a ‘bubble’ is more commonly experienced as a powerful market rally driven by structural change, technological innovation, or shifts in investor expectations,” RHB Research observes.

The current surge in AI stocks fits neatly into this dilemma: it could be excessive, or it could reflect a constructive repricing of industries critical to the economy of the future.

Optimistic but cautious

One reason RHB Research remains optimistic is the quality of the sectors drawing AI capital.

“But for what it is worth, what we are currently observing may be a concentration of capital towards sectors viewed as foundational to the next phase of economic development, including manufacturing, semiconductors, cloud infrastructure, data storage, finance and healthcare,” it explains.

In these fields, AI is not an optional add-on – it is increasingly embedded in core processes.

This, RHB Research argues, is a major distinction from past bubbles in property or cryptocurrency.

“AI is not a standalone asset class nor a speculative instrument detached from real economic activity.

“Rather, it is deeply integrated into global manufacturing and service supply chains and is widely regarded as a critical input into future productivity gains,” it states.

Still, the brokerage cautions investors not to underestimate the risks, as AI-driven expansion is being financed on a scale unprecedented in modern technology history.

“Large technology firms have pledged to accelerate AI-related capital expenditures at an unprecedented scale, much of which will ultimately be financed through the US debt markets,” it notes.

This means that retail and corporate investors, via bond purchases, pensions, and borrowing costs, are underwriting the investment wave. There are concerns about risk concentration and misallocated capital if returns fail to materialise.

RHB Research also highlights the longevity problem: infrastructure investments meant to last decades could become obsolete in a fraction of that time.

“Investors may find themselves funding 30- to 40-year infrastructure debt instruments that could become obsolete within five to 10 years, or even sooner, given the pace of technological advancement,” it suggests.

For now, though, the AI market is more about inevitability than speculation, as RHB Research sees it.

The technology is woven into the world’s production and service chains, and investors’ enthusiasm reflects both the promise and uncertainty of this transformation.