Invest1 publisherNot yet confirmed elsewhere2 min readPublished
Temasek targets 15% of its portfolio in AI as its investment chief ranks failed AI spending a top risk
Temasek's investment chief Rohit Sipahimalani ranks AI spending that fails to pay off as a top risk for next year, yet the fund aims to hold 15% in AI by 2031. Because that goal is a portfolio share, a fall in AI valuations would raise how much Temasek must buy to reach it.
The Investor · Invest desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- Sipahimalani's worry is US spending on AI data centers, chips and power; if the returns do not show up, he cautioned, valuation pressure could cause significant market disruption.
- Inflation, the second risk, is tied to the first: AI demand for chips and energy is feeding higher prices, and higher inflation raises the hurdle rates projects must clear to get funded.
- Temasek's net portfolio value reached a record S$518 billion at March 31, 2026, up 10.5% from a year earlier.
- Singapore's central bank, in its Financial Stability Review, flagged capital costs inflated by AI spending and potentially severe fiscal pressure if an AI downturn hit.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- cost At March's portfolio size, the nine-point move into AI takes about S$47 billion of allocation away from Temasek's other sectors by 2031.
- constraint If chip and energy inflation keeps hurdle rates high, fewer AI deals clear Temasek's bar, and the 2031 date slips unless the fund accepts lower expected returns.
- exposure With MAS and AMRO both pointing to Singapore and Asia taking an outsized hit in an AI downturn, Temasek's home region would fall alongside its AI book.
About 6% of Temasek's S$518 billion March portfolio is roughly S$31 billion of AI-related holdings [12], and 15% of the same portfolio would be about S$78 billion [13]. At the March size, the nine-point increase [18] comes to about S$47 billion [14]. The whole portfolio gained about S$49 billion in the year to March [15]. Spread across the roughly five years to 2031, the gap works out to around S$9 billion a year of net new AI exposure [16].
The more interesting term is that the goal is a percentage. A percentage target asks for less new cash when AI does well and more when it does badly. If the US capex Sipahimalani worries about pays off, Temasek's existing AI holdings mark up and part of the gap closes through revaluation. If the spending fails and valuations fall, as he cautioned they could [2], the 6% shrinks and the gap widens. To stay on course for 2031, the fund would then have to buy more AI, at lower prices [5]. The third path is sticky inflation with AI intact, and it works through the hurdle rates new deals must clear [4].
I think the plan is a bet on buying AI after a correction, with the stated emphasis on resilience against inflation, rates and uncertainty [6] covering the wait. The counter-thesis is plainer. A fund with a record portfolio [7] can name a risk during its annual review [3] and still want 2.5 times its current relative AI exposure [17] without holding any view on timing. The source does not say what the resilience measures are. If Temasek's next review shows the AI share rising mainly through new commitments while AI valuations stay high, the buy-on-weakness view is wrong.
Singapore's regulators describe a home market on the same side of the trade. A Monetary Authority of Singapore stress test found 32% of Singapore-listed firms could face significant revenue shocks in a severe AI downturn [9]. AMRO, the ASEAN+3 research office, reported that Asian economies are disproportionately exposed when the boom cools [10]. Crypto Briefing takes Temasek's stance as a sign that big institutional money is starting to separate AI enthusiasm from AI pricing [11]. Temasek is the only allocator in its account.
What to watch
- Whether US AI capex starts producing visible returns, the event that decides which path Temasek's 15% target follows.
- Any update to the MAS AI stress test; a rise from 32% of Singapore-listed firms facing revenue shocks would deepen the overlap between Temasek's home market and its AI bet.
- How central banks respond if AI demand keeps chip and energy prices elevated.