The Nasdaq Market site is seen outside the Nasdaq Market site in New York City, U.S., March 26, 2024. REUTERS/Brendan McDermid
PETALING JAYA: Concerns of an artificial intelligence (AI) bubble are shaping up to be the market’s biggest wildcard this year, as intensifying debate over earnings delivery and monetisation continues to keep investors on edge.
Late last year, fears of an AI bubble mounted as explosive spending on chips and data centres by tech giants far outpaced revenue growth, with hyperscalers like Microsoft and Nvidia committing trillions despite thin monetisation returns.
iFAST Capital research analyst Kevin Khaw Khai Sheng said this debate will intensify in 2026 as markets closely watch whether the huge outlays on AI infrastructure and tools lead to real monetisation.
Investors are on guard for signals that demand for AI is tailing off or that the massive spending is not paying off as expected, with Khaw noting that this year will be a key test year for AI monetisation.
Any delays or disappointments in AI monetisation during the upcoming earnings season will likely trigger heightened market volatility.
He noted also that AI-related risks outweigh tariff-related concerns in the year ahead, given that tariffs have largely been anticipated and priced in by markets.
“Many investors are expecting 2026 to be the year when monetisation begins to materialise.
“If that does not happen and companies need more time, potentially until 2027 or 2028, it could become a major drag on the broader market,” he told StarBiz.
Khaw’s expectation of 2026 being a key test year for AI monetisation is grounded in visibility from semiconductor companies’ order books, particularly Taiwan Semiconductor Manufacturing Co Ltd (TSMC).
In December 2025, it was reported that TSMC had completely sold out its entire two-nanometer chip production capacity through the end of 2026, driven by AI workloads, high-performance computing requirements, and advanced smartphone technologies.
Its advanced packaging capacity has also been fully booked, with production expected to reach 150,000 wafers monthly by 2026.
Khaw said while Malaysia may not stand to benefit much from the global AI upside, it could see negative spillover effects if the AI theme turns sour stemming from a broader market rout.
Reports on stretched valuations and AI bubble talk prompted sell-offs in big tech names like Nvidia Corp and Broadcom Inc, with several AI-related companies posting drops amid renewed investor caution.
Tech-heavy Nasdaq saw its sharpest drop of 1.81% – or more than 400 points – at the peak of AI bubble jitters in mid December last year, as part of a multi-day slide triggered by a weaker AI outlook from Oracle and margin warnings from Broadcom.
The Bursa Malaysia Technology Index also fell by a similar magnitude over the same period. Despite elevated market expectations, iFAST Capital’s Khaw said AI monetisation is “quite unlikely” to materialise this year.
He noted that so far, only a few companies – such as Salesforce and Palantir Technologies Inc – have managed to translate AI directly into revenue.
“We need more such software companies which can translate AI to direct revenue. This is the turnaround that we are expecting, which I do not think will happen by this year.
“There are already indications that ChatGPT, for instance, plans to integrate advertising revenue by including advertisements within its search model.
“This suggests that more monetisation initiatives will emerge over time, in the form of premium pricing or advertising integration for example,” he said.
“However, while such steps may materialise this year, a full turnaround – where revenue is sufficient to cover costs – is unlikely, given how large AI-related costs are.”
Any pullbacks in the market following sharp or volatile sell-offs stemming from investor disappointment in the AI theme, however, should be viewed as buying opportunities.
“Buying the dip in large US technology stocks would be the investment case here,” Khaw said.
Meanwhile, another analyst told StarBiz the current AI boom does not fit the definition of a bubble, arguing that unlike past speculative episodes, AI has real products, revenues and rapidly growing demand.
“If you look back at the year 2000 on the dot-com bubble, it happened because there was no real business case.“
Bubbles are situations where there is no rational demand or business case. That is not what we are seeing here.
These are real products with real demand.
The only issue is that demand has not yet caught up with the scale of investment being poured in.
