Asian Investors Grow More Selective on AI, Shifting Focus From Hype to Long-Term Winners

Artificial intelligence remains the dominant investment theme in global markets, but some of Asia’s largest investors are becoming increasingly selective as soaring valuations, record infrastructure spending and uncertainty over future returns prompt a reassessment of where the biggest opportunities lie.

Rather than chasing every company linked to AI, institutional investors are increasingly positioning portfolios around businesses they believe can either withstand AI-driven disruption or benefit from the technology’s expansion without depending on uncertain breakthroughs in AI applications.

The shift in sentiment was evident at the Reuters NEXT Asia conference in Singapore, where senior executives from some of the region’s largest investment firms said the next phase of the AI investment cycle will require greater discipline as markets begin asking tougher questions about valuations and profitability.

For much of the past two years, global equity markets have been propelled by enthusiasm surrounding artificial intelligence. Technology companies developing AI models, semiconductor manufacturers, cloud computing providers, and data center operators have all benefited from an unprecedented wave of investment.

That rally has lifted stock markets to record highs, but investors are increasingly debating whether corporate earnings can continue expanding fast enough to justify current valuations and whether the trillions of dollars being committed to AI infrastructure will ultimately generate attractive returns.

Rohit Sipahimalani, Chief Investment Officer of Singapore state investment company Temasek, said investors cannot afford to focus solely on companies building AI technology.

“You want to ride that trend,” Sipahimalani said during an interview at the Reuters NEXT Asia event.

“But the equally big issue is disruption because of AI to many other businesses.”

He explained that Temasek has increased its investments in businesses backed by tangible assets, arguing those companies are less vulnerable to disruption from rapid advances in artificial intelligence.

“We’ve increased our exposure to businesses that are more around hard assets, which are likely to be less disrupted by AI,” he said.

Temasek already holds stakes in AI companies, including OpenAI and Anthropic, and announced this week that it intends to increase its exposure to artificial intelligence significantly. The sovereign investment company plans to raise AI-related investments to as much as 15% of its portfolio over the next five years, up from approximately 6% today.

Even with that ambitious expansion, Sipahimalani said the investment approach will remain diversified.

“You’ve got to look at the entire value chain,” he said.

“There are some areas where there’s froth, the other areas where there’s real cash flows.”

“We try to play across the entire spectrum.”

This underscores a growing distinction within financial markets between companies benefiting from genuine commercial demand and those whose valuations have been driven primarily by investor enthusiasm. That distinction is becoming increasingly important as AI-related stocks experience sharper swings in share prices.

Investors have repeatedly questioned whether the rapid appreciation of AI companies and semiconductor manufacturers risks creating another speculative bubble similar to previous technology booms. Instead of attempting to predict which AI applications will ultimately dominate the market, some investors are choosing a simpler strategy.

Stephanie Hui, Head of Private and Growth Equity for Asia-Pacific at Goldman Sachs Asset Management, said her firm is concentrating on the infrastructure supporting AI rather than the applications themselves.

“I am not smart enough to tell you today which applications are going to be winning, it’s way too early,” Hui said during a panel discussion.

Goldman Sachs Asset Management has invested in businesses that supply the underlying infrastructure required for AI deployment, including companies specializing in liquid cooling systems and data centers, rather than betting on individual AI software companies.

As AI models become more powerful, they require more energy-intensive computing infrastructure. Advanced liquid cooling technologies are becoming essential for preventing overheating in densely packed AI servers, while new data centers continue to be built to accommodate rising computational demand.

“We are not going for the front end at this moment,” Hui said.

“We are going for the simple stuff that facilitates an end proxy for AI adoption.”

The strategy points to what many investors describe as a “picks and shovels” approach, borrowing from the California gold rush, where suppliers of essential equipment often generated more consistent returns than miners searching for gold.

Investors hope to benefit regardless of which companies ultimately emerge as long-term winners by investing in infrastructure providers rather than AI application developers.

Even among supporters of artificial intelligence, concerns about valuation are becoming more prominent.

Fred Hu, Chairman of Chinese investment firm Primavera Capital Group, said he remains convinced that AI will reshape industries but warned against excessive optimism in financial markets.

“I’m a big believer in the AI revolution but as valuations keep going up, as more and more capital goes into AI… it begs the question, how much is enough,” Hu said.

There has been growing unease that investor enthusiasm may be running ahead of commercial reality. Technology companies have announced hundreds of billions of dollars in spending on AI infrastructure, including data centers, advanced semiconductors and networking equipment.

While those investments have driven strong earnings for companies supplying AI hardware, investors now want evidence that businesses deploying AI can generate sustainable revenue growth sufficient to justify those enormous capital expenditures.

Satoshi Ueyama of Bain Capital Japan said the investment opportunities remain significant but stressed that infrastructure spending alone cannot sustain the industry’s momentum. For AI investments to generate attractive returns, businesses must ultimately create products and services that customers are willing to pay for.

“There were ample investment opportunities,” Ueyama said, but he cautioned that AI infrastructure requires end-users if the economics are to make sense.

His firm’s strategy is therefore focused on identifying companies capable of using AI to improve products and services in sectors such as consumer applications and business services.

“AI is real but at the same time there’s no denying some parts of the markets are over-excited,” Ueyama said.

“Not all AI investment is going to be successful at this stage.”

Being a major institutional investor suggests that the AI investment narrative is entering a more mature phase. During the early stages of the generative AI boom, investors largely rewarded companies simply for announcing AI strategies or increasing spending on AI infrastructure.

Today, attention is shifting toward more fundamental questions about business models, profitability, and long-term returns.

Rather than abandoning artificial intelligence, investors appear to be refining their strategies, seeking exposure across the AI value chain while avoiding areas where valuations have become detached from underlying cash flows.