Pension Funds Cut U.S. Stocks

Global Pension Funds Rethink U.S. Equity Exposure

Global pension investors are reassessing U.S. equity concentration as concerns grow over AI-linked valuations and the dominance of mega-cap technology stocks.

Large pension investors in Australia, Canada and the United Kingdom are reducing or maintaining below-benchmark exposure to U.S. equities as concerns grow about valuations and the market's increasing dependence on a relatively small group of artificial-intelligence-linked technology companies. The shift does not represent a wholesale withdrawal from American markets, nor does it establish that an AI-driven correction is imminent. Instead, it shows how some of the world's largest long-term investors are reassessing concentration risk after years in which U.S. mega-cap technology stocks generated a substantial share of global equity returns. Fresh Financial Times reporting on October 5, 2026, highlights a broader institutional debate over whether benchmark diversification still provides enough protection when a handful of companies increasingly influence both market performance and investor expectations.

Why Pension Funds Are Reconsidering U.S. Stocks

U.S. equities have been one of the strongest components of global portfolios during the technology and AI investment boom. Companies including Nvidia, Microsoft and Alphabet have become enormous components of major benchmarks, while the broader technology sector has benefited from expectations for continued AI-related earnings and capital spending.

That success has created a different problem for institutional investors: concentration.

A pension fund can own hundreds or thousands of securities and still have a substantial portion of its equity risk tied to the same underlying theme. When the largest companies in a benchmark are increasingly connected to AI infrastructure, cloud computing, semiconductors and related technologies, geographic diversification can become less effective as a measure of economic diversification.

Financial Times reporting published October 5 identified several major pension investors that are either reducing their U.S. allocation or maintaining exposure below global benchmark weights because of these concerns.

The Concern Is Concentration, Not A U.S. Exit

It is important to distinguish between reducing exposure and abandoning the U.S. market.

The institutions highlighted in the latest reporting continue to own American equities. Their decisions are primarily about relative portfolio weights.

For a global pension fund, being underweight U.S. stocks means holding less than the amount represented by its chosen benchmark. The fund can still maintain significant American exposure while directing additional capital toward Europe, Asia, emerging markets or other asset classes.

This distinction matters because headlines about pension funds “leaving” U.S. stocks can exaggerate the underlying development.

The evidence instead points toward a gradual portfolio-allocation debate: how much exposure should a long-term investor maintain when the U.S. market has become unusually concentrated in companies associated with the AI investment cycle?

Australia's Largest Pension Investor Is Watching Valuations

Australian Retirement Trust, one of Australia's largest pension investors, is among the institutions highlighted by the Financial Times.

The fund manages approximately $260 billion and has reduced its U.S. equity exposure relative to its benchmark. Its portfolio managers have expressed concern that valuations in U.S. equities and particularly the AI sector have become stretched.

The fund's position is not that American technology companies are necessarily poor businesses. Rather, the concern is whether current market prices already incorporate an unusually large amount of future growth.

That distinction is fundamental to institutional investing.

A company can have excellent technology, strong revenue growth and an important role in the AI economy while its stock still delivers disappointing returns if investors have already priced in even stronger future performance.

For pension funds with obligations extending decades into the future, valuation discipline can therefore become more important after a prolonged period of market outperformance.

Canada's La Caisse Is Also Diversifying

Canada's La Caisse is another major institution reassessing the balance of its portfolio.

The fund continues to have a large U.S. allocation but has been diversifying away from some of the largest technology companies.

The underlying reasoning is similar: investors can become exposed to substantial valuation risk when a small number of companies dominate a major market index.

This is especially relevant when the same companies are simultaneously driving expectations for earnings growth, AI investment and index performance.

For an institutional portfolio, reducing concentration does not necessarily mean predicting that technology stocks will fall. It can simply mean accepting that future returns may be less evenly distributed than they were during the earlier stages of the AI boom.

The UK's People's Pension Has Reduced Its U.S. Weight

The People's Pension in the United Kingdom provides another concrete example.

According to figures cited in current reporting, U.S. equities represented 49% of the global equity exposure of its main fund, down from 53% at the end of the previous year. That allocation is substantially below the U.S. weighting of the MSCI ACWI global equity benchmark, which is around 64% in the figures cited by the Financial Times.

The difference illustrates the scale of the portfolio decision.

A benchmark-oriented investor may automatically maintain a large U.S. allocation because American companies represent such a large portion of global market capitalization. An active asset allocator can instead decide that the benchmark's weight is too concentrated and deliberately hold less.

That decision can reduce exposure to U.S. mega-cap technology companies, but it also creates a different risk: underperformance if those companies continue to outperform global markets.

Why AI Has Changed The Diversification Debate

The AI boom has made market concentration more consequential because AI is no longer confined to one narrow technology industry.

Artificial intelligence connects several large sectors:

• Semiconductors: Advanced processors and memory are essential for AI training and inference.
• Cloud computing: AI developers rely on enormous computing infrastructure.
• Data centers: AI workloads require expanding physical capacity.
• Software: Companies are embedding AI into productivity, engineering and enterprise applications.
• Advertising: Major internet platforms use AI to improve targeting and recommendation systems.
• Energy: AI data centers require substantial electricity and supporting infrastructure.

Because many of these businesses are represented by the same large technology companies, a single investment theme can influence several sectors at once.

That makes traditional geographic diversification less straightforward.

Passive Investing Can Increase Concentration

Another issue is the growing importance of market-capitalization-weighted benchmarks.

In a market-cap-weighted index, companies with larger market values automatically receive larger weights. If a handful of technology companies rise faster than the rest of the market, their index weights increase.

That can create a feedback mechanism.

Strong performance increases a company's market capitalization. The larger capitalization increases its benchmark weight. Investors tracking the benchmark then own more of that company.

This is not necessarily a flaw in index investing. It is simply how market-cap weighting works.

But it means that an investor who believes an index provides broad diversification may still have substantial exposure to a small number of companies.

For pension funds, which often manage extremely large portfolios, that concentration can become a major risk-management consideration.

Institutional Research Shows The Concern Is Broadening

The current debate extends beyond a handful of pension funds.

Research from Marsh cited by the Financial Times found that approximately one-third of more than 430 global institutions surveyed planned to reduce U.S. equity exposure during the following year. The combined institutions represented more than $5 trillion in assets under management.

That survey does not mean $5 trillion is leaving U.S. stocks. It describes institutions participating in a survey and their stated intentions regarding allocation.

Still, the direction of the responses is significant because institutional portfolios represent a large portion of the capital supporting global equity markets.

If more institutions gradually reduce their U.S. allocation, the result could be a broader redistribution of capital rather than a sudden market exit.

Denmark's ATP Is Watching Earnings Expectations

Denmark's ATP, another major pension investor, is also examining the relationship between market valuations and expected corporate earnings.

The concern is straightforward: when stock prices incorporate very strong assumptions about future earnings growth, even companies that continue to grow can experience sharp declines if results fall short of expectations.

This is particularly relevant to AI-related stocks because the sector has attracted enormous expectations around future productivity, software adoption, semiconductor demand and infrastructure spending.

Investors therefore have to assess two different questions.

1. Will AI create substantial economic and corporate value?
2. Have investors already paid too much for that expected value?

The first question can have a positive answer while the second is still a concern.

Why This Matters For The S&P 500

The S&P 500 has become one of the world's most important equity benchmarks. Its performance influences global asset allocation, retirement portfolios, exchange-traded funds and institutional investment decisions.

When the index becomes more concentrated, its headline performance can become increasingly sensitive to the earnings and valuations of its largest constituents.

That means a major change in expectations for AI spending could have consequences beyond the technology sector.

If investors become more optimistic about AI monetization, large technology companies could continue supporting index returns.

If expectations weaken, the same concentration could amplify market volatility.

This is the risk institutional investors are trying to manage—not necessarily by predicting the next market crash, but by avoiding excessive dependence on a single group of companies or economic assumptions.

Europe And Emerging Markets Could Benefit

If large institutional investors reduce their relative U.S. allocations, some of that capital could move toward other regions.

Europe could benefit from renewed attention to industrial companies, financial institutions, healthcare businesses and other sectors that have been less dominant in the AI rally.

Emerging markets could also receive additional interest if valuations remain attractive and economic growth prospects improve.

However, diversification does not automatically mean higher returns.

Investors moving away from U.S. technology stocks could miss further gains if American AI companies continue to outperform.

The decision therefore represents a trade-off between reducing concentration risk and accepting the possibility of lower returns relative to a U.S.-heavy benchmark.

The Market Is Not Yet Showing A Mass Exit

There is an important difference between institutional caution and a market-wide reversal.

The latest developments show selected pension funds becoming more cautious about U.S. equity valuations and AI concentration. They do not demonstrate that global investors have broadly abandoned American stocks.

U.S. equities remain deeply embedded in global portfolios, and many of the world's largest and most profitable companies remain headquartered in the United States.

For that reason, even institutions that are underweight the U.S. can continue to hold substantial American exposure.

The more significant development is that the default assumption of ever-increasing U.S. allocation is being challenged.

What Could Change Institutional Sentiment

Several developments could determine whether this diversification trend accelerates or reverses.

• AI earnings: Stronger-than-expected earnings from major technology companies could justify current valuations.
• AI capital spending: Continued investment in chips and data centers would support the infrastructure side of the technology cycle.
• Productivity growth: Evidence that AI is materially increasing economic productivity could strengthen the long-term investment case.
• Interest rates: Higher long-term yields can pressure richly valued growth companies.
• Valuations: Further expansion in technology multiples could encourage additional diversification.
• Global growth: Stronger growth outside the United States could make international markets more attractive on a relative basis.

The Bigger Issue Is Portfolio Construction

The pension-fund debate is ultimately about portfolio construction rather than a simple judgment on American technology.

AI could become one of the most important productivity technologies in modern economic history. If that happens, owning successful AI companies could remain highly valuable.

But institutional investors also have obligations to manage downside risk. Pension portfolios must function across different economic environments, including periods when technology valuations fall, interest rates rise or economic growth disappoints.

That makes diversification valuable even when a particular market has been extraordinarily successful.

The current reallocation discussion therefore reflects a classic investment tension: investors want exposure to the companies shaping the future, but they do not want their entire portfolio to depend on one forecast about that future.

What Investors Should Watch Now

The next stage of the story will be visible in institutional allocation data, corporate earnings and market breadth.

If more pension funds, sovereign investors and large asset managers reduce U.S. exposure, the trend could become a meaningful structural shift in global capital allocation.

If U.S. technology companies continue delivering earnings growth that exceeds expectations, institutional investors may find that reducing exposure carries a significant opportunity cost.

The most important signal may therefore not be whether pension funds are reducing U.S. equities today, but whether their decisions persist through the next several earnings cycles.

For now, the message from several major institutional investors is measured rather than dramatic: U.S. equities remain important, but concentration around AI-linked mega-cap technology companies deserves greater scrutiny.

Frequently Asked Questions

Why are pension funds reducing U.S. equity exposure?

Several large pension investors are concerned about high valuations and the growing concentration of U.S. market performance in a relatively small group of AI-linked technology companies.

Are pension funds abandoning U.S. stocks?

No. The reported changes are primarily relative allocation decisions. The institutions continue to own U.S. equities but are reducing exposure relative to their benchmarks or strategic targets.

How concentrated is the U.S. stock market?

The Financial Times reported that more than one-third of the S&P 500 is represented by large-cap companies closely connected to the AI investment cycle, increasing the importance of concentration risk for institutional investors.

What did the Marsh survey find?

Marsh reported that approximately one-third of more than 430 global institutions surveyed planned to reduce their U.S. equity exposure over the following year. The institutions represented more than $5 trillion in assets under management.

Why does AI increase portfolio concentration risk?

AI connects several major technology businesses, including semiconductors, cloud computing, data centers, software and digital advertising. Many of the largest companies in these areas are already major components of U.S. stock indexes.

Could diversification hurt pension-fund returns?

Yes. If U.S. mega-cap technology companies continue to outperform other global markets, funds that are underweight the United States could lag their benchmarks. Diversification reduces some concentration risk but does not guarantee higher returns.

What could reverse the trend?

Stronger-than-expected AI earnings, continued productivity gains, attractive U.S. valuations or weaker performance from overseas markets could encourage institutional investors to increase their U.S. allocations again.