Temasek, Singapore's sovereign wealth fund managing more than $400 billion, plans to nearly triple its AI-related portfolio exposure from roughly 6% today to 15% by 2031, according to The Register. The fund's portfolio value reached a record high this year, with AI named alongside infrastructure and private credit as a priority growth area.
What's notable is the reasoning behind the target, not just the number: a Temasek executive argued that direct investment in AI companies -- the labs, the chipmakers, the infrastructure providers -- captures only a fraction of the technology's total economic value. The larger opportunity, in Temasek's framing, sits in the other 85% of the global economy that adopts and is reshaped by AI, rather than in the relatively narrow set of companies actually building the underlying models and hardware.
“That thesis puts Temasek in an interesting position relative to other large sovereign and institutional investors.”
That thesis puts Temasek in an interesting position relative to other large sovereign and institutional investors. Funds like Saudi Arabia's PIF and Norway's Government Pension Fund Global have also increased AI exposure, but much of that has concentrated in direct bets on infrastructure names -- chipmakers, data centers, foundation-model labs -- rather than explicitly betting on adoption across the broader economy the way Temasek is framing its target.
A sovereign fund of Temasek's size moving from 6% to 15% AI exposure over five years represents tens of billions of dollars in incremental capital deployment, regardless of exactly how it's split between direct AI infrastructure bets and adoption-driven plays across its existing portfolio companies. That scale of committed, patient capital is meaningfully different from venture capital's typical 3-7 year return horizon -- sovereign wealth funds can underwrite theses that take a decade to fully play out.
For GPs raising from sovereign LPs, Temasek's explicit adoption thesis is worth internalizing: the pitch that resonates with this class of capital may increasingly be about which portfolio companies are using AI to transform existing large markets, not just which ones are building AI infrastructure directly. For founders, a $400 billion fund publicly reasoning that most AI value capture happens outside the AI industry itself is a useful data point when framing your own company's AI narrative to institutional investors -- "AI-native disruptor of a large existing market" may land better with this capital base than "AI infrastructure company" alone.