Sovereign Wealth Funds chase AI infrastructure boom

News

1. Sovereign Wealth Funds chase AI infrastructure boom

2. AI Capex Boom Could Turn Into Investment Bust, BIS Warns

3. Indian IT Firms See Agentic AI Opening $400 Billion Opportunity

4. OpenAI Hires Uber Veteran Prabhjeet Singh for India Push

5. Adani, Jabil Plan AI Data Center Hardware Platform in India

6. Arun Misra May Take over as Vedanta CEO in August

Sovereign wealth funds are pulling back from crowded listed equity markets and shifting more capital into infrastructure, private credit and private equity, as the artificial intelligence boom changes how some of the world’s largest state investors seek returns.

The move reflects a broader reassessment of portfolio construction among sovereign investors, according to Invesco’s 2026 Global Sovereign Asset Management Study, which surveyed 144 senior investment professionals from 90 sovereign wealth funds and 54 central banks managing about $29 trillion in assets.

The sharpest shift is away from public equities.

Invesco said a net 17% of sovereign wealth funds plan to cut listed equity exposure over the next 12 months. By contrast, infrastructure, private equity and private credit all showed strong positive allocation intentions, with net increases of 28% to 35%.

The change is not a retreat from risk but a more selective bet on where long-term returns may come from after years in which broad equity exposure was lifted by a narrow group of large technology companies.

Rotating capital

“Capital is rotating toward infrastructure and private credit, the bar for passive market exposure is rising,” Invesco said in the report.

Much of that rotation is now tied to AI. Sovereign investors see the technology as a structural investment theme, but many appear wary of capturing it mainly through listed technology stocks, where gains have become concentrated among a handful of companies.

A Middle Eastern development sovereign quoted in the study put it bluntly: “The AI wave is currently best captured in private credit and infrastructure opportunities.”

That view is pushing investors toward the physical foundations of AI rather than only the software layer. Data centers, power supply, digital infrastructure and energy systems are emerging as preferred routes into the AI build-out.

An Asia-Pacific investment sovereign told Invesco: “Infrastructure comes first because these technologies need power and data centers, no matter who wins the software race.”

The same investor said productivity and automation were “slower and less visible,” but added that this was where “lasting value probably builds over time.”

Investment themes

The report found that AI infrastructure and productivity or automation enablement were the most compelling long-term AI investment themes for sovereign wealth funds, with each cited by 69% of respondents.

Semiconductors and hardware followed at 54%, while software platforms and applications were cited by 39%.

Current exposure also shows the tilt toward hard assets. Half of sovereign wealth funds said they were gaining exposure to AI through infrastructure linked to the technology. Broad public equity exposure was cited by 33%, while private markets were cited by 41%.

Some of the largest sovereign investors already have portfolios that reflect that tilt. Temasek said 49% of its portfolio was in unlisted assets and funds as of 31 March 2025, while 51% was in liquid and listed assets.

Mubadala, Abu Dhabi’s sovereign investor, said private assets accounted for 42% of its portfolio in 2025, with real estate and infrastructure adding another 17%.

The figures show that the shift into private markets is not only an intention for the next allocation cycle. For large state investors, it is already a major part of portfolio construction.

The shift comes as investors question whether passive exposure to stock indices still provides the diversification they once assumed. Invesco said index-heavy strategies now carry significant exposure to a small number of large technology companies, prompting some sovereign funds to review concentration risk at the portfolio level.

Diversification model

A North American liability sovereign said: “We have drawn down from equities because of a high concern of deep concentration.”

Invesco said the concern was not with the equity risk premium itself, but with “the distribution of returns within it” and the extent to which broad index exposure had become a bet on a narrow set of companies.

That concern has grown alongside doubts about the older equity-bond diversification model. The inflation shock of recent years weakened the assumption that bonds would reliably cushion equity losses. Several institutions told Invesco that traditional diversification had become less dependable.

“Bonds still help, but they do not solve everything on their own,” an Asia-Pacific liability sovereign said.

Invesco said sovereign wealth funds are placing greater emphasis on private and illiquid assets, while central banks are more likely to reassess long-term return assumptions and increase scenario analysis and stress testing.

The private credit shift is still not without risk. The asset class has drawn heavy inflows in recent years, raising questions about whether competition will erode returns.

“There is a lot of enthusiasm for private credit, although I am personally cautious that a lot of money is chasing the space,” a European liability sovereign told Invesco.

Power availability is another emerging constraint. One Middle Eastern development entity with large AI infrastructure investments said inference capacity was unlikely to face a bubble in the same way other parts of the AI market might.

The power equation

“We are of the view that there is never going to be a bubble in inference capacity, because power is going to be your constraint,” the institution said.

That makes energy supply a central question for the next phase of AI investment. Invesco said power, rather than model performance alone, could determine where data centers are built and how quickly they come online.

The study also found that sovereign investors still see the US as the region best placed to lead in AI development and adoption, helped by its capital markets, technology companies and research base.

China was the second-most cited region, with some respondents saying its industrial framework and talent pipeline could help it narrow the gap.

One Asia-Pacific liability sovereign said: “For AI it currently has to be US against China. But the rest of the world will eventually have that realization that each country will need to have its own capability.”

For countries trying to build domestic AI capacity, that may make sovereign capital more important. The next phase of the AI race is no longer only about models and chips. It is also about financing power, data centers and infrastructure at a scale that few private balance sheets can carry alone.