Micron says AI boom is here to stay as it ramps up Singapore investment

If you have any doubts about the longevity of the artificial intelligence boom, look at the stellar performance of and outlook for semiconductor firms such as Micron Technology, a US memory chipmaker that also has a large manufacturing footprint in Singapore.

The market value of New York-listed Micron hit US$1 trillion (S$1.28 trillion) for the first time on 26 May, up from US$110 billion a year ago. It is now among the US trillion-dollar club that already includes chipmakers NVIDIA, Broadcom, Taiwan Semiconductor Manufacturing Company, and Samsung Electronics.

Justifying the surge in stock value, Micron delivered a record second-quarter 2026 result in March, with revenue reaching US$23.9 billion – up 196 per cent year on year. The company expects third-quarter revenue to beat that record and reach US$33.5 billion.

While the company’s management has played a key role in the achievements, none of this would have been possible without the unprecedented investment boom in building the infrastructure to support AI technology.

“The extent by which demand is exceeding supply is at historically high levels,” said Mr Sumit Sadana, executive vice-president and chief business officer at Micron.

“We continue to see supply growth in the medium term, meaning in the 2026 and 2027 calendar years, at levels that will not be able to close the gap between supply and demand,” he said.

AI has bolstered economies such as Singapore, leading to increased exports and the creation of high-value jobs, as new investments pour into its electronics sector.
 

Earlier in 2026, Micron announced plans to invest US$24 billion over 10 years in a new advanced wafer fabrication plant in Singapore, on top of a separate US$7 billion facility announced in 2025.

Together, the new plants will create around 3,000 new jobs.

Micron believes the AI build-out will endure for many years to come, keeping demand for semiconductors – particularly memory chips that process and store data – outpacing supply.

Memory chips are distinct from other computer chips such as logic processors that calculate and process instructions.

Hence, the company will continue to invest in boosting the supply of AI-powering semiconductors, including high-bandwidth memory chips and NAND flash memory chips used in solid-state drives that support data centres.

Research and advisory firm Gartner said on 19 May that worldwide spending on AI will total US$2.59 trillion in 2026, a 47 per cent annual jump, and reach US$3.49 trillion in 2027.

Gartner believes that through the next several years, the need for computational capacity will make AI infrastructure – including AI-optimised data centres, servers, network fabric, semiconductors, and devices – the largest segment of the market, accounting for more than 45 per cent of spending.

Micron’s global capital expenditure – the money spent to acquire, upgrade, or maintain long-term physical assets – will total US$25 billion in fiscal 2026, up from US$13 billion in 2025.

The firm’s fiscal year ends in September.

Its closest peers, such as South Korea’s Samsung and SK Hynix, are also investing heavily in ramping up production. Still, the demand-and-supply gap is likely to persist for at least a few more years.

Mr Sadana said that even with all the new investments, a meaningful supply of memory chips will start to come online only late in the 2027 calendar year, and ramp up in 2028.

“Even in that environment, the expectation is for demand to continue to increase. And so, we don’t have line of sight today as to when the gap between supply and demand can be closed,” he said.

The lack of supply has resulted in an unprecedented surge in memory chip prices. According to different estimates, contract prices of advanced memory chips have soared between 400 per cent and 600 per cent since 2025.

Mr Sadana said Micron is working hard not just to maximise supply, but also to ensure that its pricing approaches are thoughtful.

“We are always mindful of the impact pricing may have on our customers’ businesses and the impact the lack of adequate supply may have,” he said.
 

For the memory chip business, an environment where demand persistently exceeds supply is a sea change from the boom-and-bust cycles it suffered in the years before the AI boom.

Since the emergence of AI chatbots – starting with ChatGPT in 2023 – that simulate human conversation using natural language processing and machine learning, tech giants such as Amazon, Alphabet, Microsoft, and Meta have collectively committed hundreds of billions of dollars to expand data centres and computing capacity.

As the race to build ever-larger facilities to power AI applications intensifies, servers are being designed with far more bandwidth memory and require increasing amounts of conventional memory chips to handle training data and support cloud workloads.

“We expect to be primary beneficiaries of the growth of AI because AI is fundamentally a very memory-intensive technology,” said Mr Sadana.

“We have always had the vision that memory is a strategic asset, but what AI has done is make that very, very evident to everybody, not just to our customers, but also to the whole ecosystem. That’s part of the reason why we are in an unprecedented gap between supply and demand.”

While the demand for AI is mostly centred on data centres so far, experts see the technology proliferating across all electronic devices and applications, resulting in far more intelligent mobile phones, cars, computers, and industrial machines.

AI can also help create new drugs, boost industrial productivity, and improve workflows.

Hence, to deploy AI at scale across the economy, it is going to take a massively bigger computing infrastructure than we have in place today, said Mr Sadana, adding: “We see this as certainly a very long-term trend, and it will be enormously memory-centric and require a lot bigger aggregate supply of memory in the future than we can produce today.”

The AI-powered explosive growth in the semiconductor industry has also raised the need to ensure a steady talent supply.

To that end, Micron works closely with local universities and ecosystem partners, in Singapore and elsewhere, to build a future-ready semiconductor workforce.

The collaboration includes designing structured learning journeys that combine classroom learning with real industry context, including mentorships, site learning visits, internships, and applied projects in advanced semiconductor manufacturing and research and development.

“We are working actively to develop the talent to train our teams on the usage of cutting-edge AI tools internally, and also work with the local community to help them understand the benefits of AI and pull more people into the world of AI, educate people, show them the possibilities,” Mr Sadana said.
 

In January, Micron and AI Singapore announced a strategic partnership to advance AI literacy and talent development across Singapore. The collaboration will focus on making AI education accessible and inclusive for young people and underserved communities in Singapore with funding from the Micron Foundation.

Mr Sadana said Micron is also partnering with Singapore’s institutes of higher learning to create internship programmes to provide students with real industry experience in advanced semiconductor manufacturing.
 

Source: The Straits Times © SPH Media Limited. Permission required for reproduction.