Singapore’s GIC Adds $30 Bn to Hedge Funds While Widening AI Bets

Singapore’s sovereign wealth fund GIC, which manages part of the country’s reserves, is widening its investments across the AI economy while committing another $30 Bn to hedge funds that can reposition as markets turn. The two moves reflect a broader portfolio strategy: pursuing technology-led growth while adding sources of return that can perform across a wider range of market conditions.

Over the next three years, GIC plans to channel the $30 Bn into global macro, quantitative and multi-strategy funds, building on hedge-fund investments that have already tripled over the past decade, group chief investment officer Bryan Yeo told Reuters.

Together, the moves show how GIC is responding to an AI boom that is creating real economic value even as market gains, capital spending and investor expectations pile up around a relatively small group of companies.

GIC has not presented the hedge fund expansion as protection against an AI correction. Its purpose is broader. As markets become harder to read, the fund wants more strategies that can move across currencies, interest rates, commodities and securities rather than depend on shares or bonds rising.

That flexibility allows GIC to keep investing across the AI economy without tying the wider portfolio too closely to a small group of technology companies. The fund sees durable value emerging from semiconductors, data centres, power networks, software and business adoption, even as it warns that hype and stretched valuations are building in parts of the market.

The $30 Bn commitment is, therefore, not a bet against AI. It broadens the sources of return around one of the most powerful, expensive and crowded investment cycles in global markets.

Why the 3.4% Return Can Mislead

GIC reported an annualised real return of 3.4% for the 20 years ended on March 31, 2026, down from 3.8% a year earlier and its lowest since 2020, when the figure stood at 2.7%. Its annualised nominal return in US dollar terms was 5.6%.

That does not mean the fund earned 3.4% during its latest financial year. GIC does not publish a one-year portfolio return. Its main performance measure looks back over two decades and shows the average annual return after global inflation.

The calculation changes each year as the newest period enters the 20-year window and the oldest falls away. The lower reading, therefore, reflects not only the performance being added but also the year that has dropped out.

Even at 3.4%, the reserves managed by GIC came close to doubling their purchasing power over two decades. In nominal terms, they roughly tripled.

How GIC’s 20-Year Real Return Has Changed 

Source: GIC annual reports. GIC’s 2025/26 reporting year, for example, covers April 1, 2025 to March 31, 2026. The 3.4% figure shown for that year is GIC’s average annual return above global inflation over the 20 years to March 31, 2026, not its return for 2025/26 alone.

Chief Executive Lim Chow Kiat said the result also reflected a deliberate decision to take less risk and spread investments more widely. That approach can temper returns when a narrow group of assets is rallying but is intended to limit losses when markets reverse and preserve capital for opportunities that emerge during periods of stress.

GIC does not disclose the total value of the assets it manages. So, the new $30 Bn hedge fund commitment cannot be measured reliably as a proportion of the overall portfolio.

More Equities, More US Investment and Broader Diversification

GIC’s reported portfolio has moved further towards growth assets.

Equities accounted for 56% of the portfolio at the end of March, up from 51% a year earlier. Fixed income fell to 22% from 26%, while real assets slipped to 22% from 23%. The Americas remained GIC’s largest regional exposure at 53%, up from 49%, as the fund increased its investments in the US.

How GIC’s Portfolio Has Evolved

Sources: GIC and Reuters. Portfolio weights are measured at the end of each reporting year. Changes may reflect both investment activity and market movements. GIC monitors geographic exposure but does not set formal regional allocations.

Those shifts do not necessarily represent fresh purchases alone. Rising markets can increase the weight of an asset class without new money being added, and GIC has not separated investment flows from valuation gains. What the figures show is that a larger share of the reported portfolio now sits in equities and the Americas.

The hedge fund expansion adds a different source of return around that growth exposure. Global macro managers can invest across currencies, rates, commodities and securities. Quantitative funds use systematic models, while multi-strategy firms can move capital between approaches as opportunities change.

But none of these strategies guarantees protection when markets fall. Their appeal lies in flexibility. They can alter positions, reduce risk or move into different markets rather than remain locked into a conventional mix of shares, bonds, credit and private assets.

The move also comes as GIC transitions to a refreshed investment framework, a process it began on April 1, 2026. Instead of organising holdings mainly by traditional asset classes, the fund now focusses on what each investment is intended to deliver — be it growth, income or protection against inflation — across public and private markets.

GIC said the framework was designed to prepare the portfolio for a range of outcomes rather than rely on one forecast. That approach reflects a market shaped by geopolitical conflict, heavy government borrowing, energy constraints and rapid technological change.

There is also a track record behind the greater emphasis on active management. From April 2013 to March 2026, GIC’s active strategies generated an average 52 basis points of gross annual alpha over its Policy Portfolio (0.52 percentage point a year above the return from its long-term benchmark portfolio, before costs). The figure covers active investing across the fund and should not be read as a forecast for the new hedge-fund allocation.

Investing in AI Without Picking One Winner

GIC is applying the same logic to AI. Rather than treat it as a single trade, the fund divides the opportunity into three layers: companies building the infrastructure, businesses selling AI products and enterprises using the technology to improve their existing operations.

Where GIC Is Looking for AI Value

Source: GIC. Why It Matters and Key Risk analysis by Entrepreneur APAC

It then asks three questions of each company. Does it have a defensible advantage? Can management turn AI into revenue or cost savings? Can early gains produce the data, scale and operating momentum needed to stay ahead?

The clearest opportunity today lies in the infrastructure behind AI. Training and operating advanced AI models require semiconductors, memory, high-speed networks, data centres, electricity and cooling. Annual capital spending by hyperscalers, large cloud-computing companies that build and run vast data centre networks, is approaching 2% of US GDP, according to research cited by GIC.

That spending is creating opportunities in chip manufacturing, power equipment, electric grids and other infrastructure. But it is also creating risk. Capital is entering the sector at extraordinary speed, while lack of power, grid connections and specialised talent could delay projects or weaken returns. Rapid changes in technology could also leave some assets outdated before investors recover their costs.

The winners are harder to identify among model developers and software companies. Although models are improving quickly, product features can be copied, and it remains unclear which businesses will retain pricing power once the first wave of enthusiasm subsides.

Business adoption is advancing, but unevenly. A GIC survey of nearly 200 North American enterprises found that more than half ranked AI among their three leading priorities. Among that group, one in five placed it first. Respondents reported savings in finance, customer service, onboarding and non-software research and development.

The survey shows that some companies are already finding practical uses for AI, but it should not be treated as a measure of adoption across all North American businesses. GIC has not published detailed information on how the sample was selected, the response rate or whether the results were weighted.

For GIC, that combination of real value and uncertain winners argues for diversifying its exposure. The fund can invest in listed chipmakers and hyperscalers, private AI developers, physical infrastructure and established companies that use the technology to become more productive.

A Different Route from Temasek

The approach points in the same direction as that of Singapore’s other major state investor, but GIC and Temasek are using different instruments.

Temasek has set explicit portfolio targets. It plans to lift AI-related exposure from 6% to as much as 15% by 2031, while increasing private credit from 2% to 5% and core-plus infrastructure from 1% to 5%. Core-plus infrastructure covers assets such as power grids, energy storage and data centres that can produce long-term cash flows while retaining scope for expansion or upgrading. Private credit is intended to add recurring income and downside protection, while infrastructure broadens Temasek’s investments in energy, grids and other long-duration assets supporting the AI build-out.

GIC has not disclosed the size of its AI holdings or announced a numerical target. Its approach is more fluid: a larger hedge-fund programme, a bigger role for active management and a framework intended to move capital as risks and opportunities shift.

The comparison has limits. Temasek owns a portfolio whose net value and one-year shareholder return it reports publicly. GIC manages part of Singapore’s reserves, does not disclose its assets under management and is judged mainly on whether it can outperform global inflation over 20 years.

Yet both institutions are addressing the same investment problem. AI is too important to ignore, but too expensive, concentrated and fast-moving to approach as a simple one-way wager.

GIC’s $30 Bn hedge fund commitment is not a prediction that the AI boom will fail. It allows the fund to invest in the technology’s long-term promise without letting one cycle, one group of US companies, or one market consensus dictate the performance of Singapore’s reserves.
(With inputs from Reuters)