AI concentration, bullish positioning and the yen intervention puzzle

Summary

Global investors are facing a growing concentration risk as artificial intelligence spending increasingly links equities, investment-grade bonds, high-yield debt and private-market financing to the same underlying AI investment cycle. Bank of America’s latest Global Fund Manager Survey found cash allocations falling to 3.5%, the sixth-lowest level in the survey’s 28-year history, while a 56% net overweight to global equities pointed to highly bullish positioning. At the same time, the recent U.S.-Japan intervention to support the yen has come under scrutiny after the currency gave back much of its initial gains.

LONDON, August 19, 2026 — Global financial markets are grappling with three interconnected issues: the growing concentration of portfolios around artificial intelligence, increasingly bullish investor positioning and questions over the effectiveness of the recent U.S.-Japan intervention to support the Japanese yen. Reuters commentary highlights how AI exposure is spreading well beyond technology stocks into corporate credit, data-center financing and private markets, making traditional diversification less effective.

AI concentration spreads across asset classes

AI concentration has become a broader portfolio issue rather than one confined to benchmark technology stocks. Investors holding the S&P 500 now have roughly 50% exposure to AI-linked companies, including hyperscalers, chipmakers and AI infrastructure businesses, according to the Reuters analysis.

The concentration is also becoming increasingly visible in fixed-income markets. AI-related bond issuers accounted for about one-third of net U.S. investment-grade issuance so far this year, while Apollo’s credit team estimates that the five major AI hyperscalers could increase their combined share of the U.S. investment-grade credit index from less than 5% currently to almost 10% by 2030.

Apollo also estimates that the five major hyperscalers, together with SpaceX, could have a larger share of the investment-grade index than the five largest U.S. bank issuers by the end of the decade.

The high-yield market is experiencing a similar shift. Data-center financing has rapidly expanded, with companies including Core Scientific and TeraWulf raising capital to convert former cryptocurrency-mining facilities into AI data centers.

According to Apollo, direct data-center exposure has risen from virtually zero to nearly 5% of the U.S. high-yield index, making it comparable to a major standalone sector. GPU-backed financing and AI-related private-market funding add another layer of exposure.

The result is that portfolios that appear diversified across technology stocks, corporate bonds, infrastructure and real estate may still be exposed to the same underlying AI spending cycle.

BofA survey shows investors remain heavily bullish

Despite concerns surrounding AI valuations and hyperscaler debt, investors have not been moving aggressively into cash.

Bank of America’s August Global Fund Manager Survey, covering 180 fund managers overseeing about $525 billion, showed average cash holdings falling to 3.5%, from 3.6% in July. That was the sixth-lowest reading in the survey’s 28-year history and placed cash levels in the range that Bank of America considers its contrarian “sell” zone.

Investor positioning was similarly aggressive:

• Global equities: Managers were a net 56% overweight, the highest level since 2021. 
• U.S. economic outlook: Almost 60% expected a “no landing” scenario. 
• AI spending: More than 70% did not expect reductions in AI-related capital expenditure forecasts this year. 
• Crowded trade: Long semiconductor positions remained the most crowded trade. 
• Key risks: An AI bubble and hyperscaler capital-expenditure debt were among the major perceived risks. 

The combination of very low cash holdings and high equity exposure suggests that investors remain strongly positioned for continued economic and corporate earnings growth, leaving relatively little defensive positioning should the AI trade reverse.

Yen intervention raises questions over durability

The Japanese yen has also become a focus for global investors after the United States and Japan intervened to support the currency.

Japan intervened on July 30, followed by a coordinated U.S.-Japan operation on July 31. The intervention initially strengthened the yen, but much of that gain has since faded. Reuters reported on August 13 that the yen’s post-intervention strength was already under pressure, with markets increasingly focused on whether the Bank of Japan would raise interest rates.

Deutsche Bank FX strategist George Saravelos offered a different explanation for the intervention’s limited durability. He argued that the structure of the operation may itself have weakened its credibility.

The Federal Reserve’s Foreign and International Monetary Authorities (FIMA) Repo Facility allows foreign official institutions such as Japan to obtain dollar liquidity by using U.S. Treasuries as collateral rather than selling those securities outright.

That approach can reduce the need for Japan to liquidate Treasury holdings in the open market. But Saravelos argued that borrowing dollars through FIMA at relatively unattractive rates could make repeated intervention expensive, potentially raising the threshold for further yen-support operations.

Reuters also noted that Federal Reserve System Open Market Account data indicated limited U.S. financial commitment during the intervention. That has led currency traders to question how much firepower Washington would actually deploy in future operations.

The yen’s subsequent weakness provides an early market test of whether the intervention can produce a lasting change in currency expectations without a significant shift in Japanese monetary policy.

Why this matters

• AI risk is spreading beyond stocks: Hyperscaler bonds, data-center debt, power infrastructure and semiconductor investments are increasingly tied to the same AI capital-spending cycle.
• Bullish positioning leaves little cash buffer: BofA’s 3.5% cash reading and 56% net equity overweight show that investors are heavily positioned for continued gains.
• High-yield markets are gaining AI exposure: Data-center financing has grown to nearly 5% of the U.S. high-yield index, adding another channel through which an AI investment slowdown could spread.
• Yen intervention faces a credibility test: The currency’s renewed weakness suggests that intervention alone may not be sufficient without supportive Japanese monetary policy and credible follow-through.

FAQs

Q1: How much of the S&P 500 is exposed to AI?

According to the Reuters analysis, roughly 50% of the S&P 500 is exposed to AI-linked companies, including hyperscalers, chipmakers and AI infrastructure providers.

Q2: What did Bank of America’s August fund manager survey show?

The survey found average cash holdings at 3.5%, while managers were a net 56% overweight global equities. Almost 60% expected a “no landing” scenario for the U.S. economy.

Q3: How significant is AI exposure in the high-yield market?

Apollo estimates that direct data-center exposure has grown from virtually zero to nearly 5% of the U.S. high-yield index, with companies such as Core Scientific and TeraWulf raising debt to convert facilities for AI computing.

Q4: Why is the FIMA Repo Facility important to the yen intervention?

The FIMA facility allows Japan to obtain dollar liquidity using U.S. Treasuries as collateral rather than selling the bonds outright. Deutsche Bank’s George Saravelos argues that the relatively high cost of this funding could make repeated intervention less attractive.

Q5: Why has the yen struggled to maintain its post-intervention gains?

Market participants have pointed to expectations around Bank of Japan interest-rate policy as one reason. Reuters also reported concerns that the structure and limited follow-through of the intervention may have weakened its long-term credibility.