Investors are beginning to rethink their artificial intelligence strategies as concerns grow that the rapid expansion in hyperscaler spending could lose momentum over the next few years, prompting some fund managers to reduce exposure to semiconductor stocks in favour of software companies and major cloud providers.
AI infrastructure rally faces fresh scrutiny
For much of the past two years, the dominant investment strategy has been to buy companies supplying chips and infrastructure used to power AI, based on expectations that Microsoft (NASDAQ:MSFT), Amazon (NASDAQ:AMZN), Alphabet (NASDAQ:GOOG) and Meta (NASDAQ:META) would continue increasing investment in data centre expansion.
That outlook is beginning to shift.
According to UBS, hyperscaler capital expenditure is expected to rise 76% this year to approximately $673 billion. However, spending growth is forecast to slow to 25% in 2027 and just 6% in 2028.
The prospect of moderating investment has encouraged some active managers to reduce positions in semiconductor companies while increasing exposure to hyperscalers, whose share prices have lagged AI chipmakers. Others are rotating into software businesses and sectors expected to benefit from wider AI adoption, including financials and healthcare.
Portfolio managers adjust positioning
Alexis Bossard, Global Equity Portfolio Manager at Edmond de Rothschild Asset Management, said his team has significantly reduced semiconductor exposure because valuations have become increasingly demanding.
“Once they stop increasing their capex, it will definitely be a relief for hyperscalers and a negative signal for the semi industry,” he said.
The Philadelphia Semiconductor Index, whose largest holdings include Nvidia, Broadcom, Micron, ASML and TSMC, has more than doubled over the past year despite retreating nearly 18% from its June high. Over the same period, the equal-weighted S&P 500 has gained 11%, while Europe’s STOXX 600 has risen around 8%.
Bank of America’s July survey of global fund managers found that 82% identified semiconductors as the market’s most crowded trade, while none reported holding short positions in the sector.
Bossard has increased exposure to Amazon while favouring businesses involved in liquid cooling, cybersecurity and selected software companies.
“We have a massive underexposure to semis right now.”
Financing questions emerge
LFG+ZEST Chief Investment Officer Alberto Conca has also reduced holdings in memory chip manufacturers and semiconductor equipment companies, while increasing allocations to hyperscalers and healthcare stocks. He has additionally purchased put options on selected semiconductor names.
After initially funding AI expansion using their own balance sheets, hyperscalers are increasingly turning to debt markets, raising questions about whether financing conditions could eventually limit future investment.
Apollo Chief Economist Torsten Slok noted that demand for new corporate bond issuance has weakened. Cover ratios have fallen to below two times in July from nearly five times in February, indicating softer investor appetite.
The Bank for International Settlements also warned in June that weaker-than-expected returns on AI investments could reduce funding availability and eventually reverse the current capital expenditure boom.
“Cash flow is starting to be almost completely drained by capex,” Conca said.
Empirical Research has also highlighted a widening gap between slowing capital expenditure growth and the optimistic revenue forecasts currently assigned to semiconductor manufacturers and AI infrastructure suppliers.
“Either the capex trajectory of the hyperscalers will be upgraded again, or the revenue growth pencilled in for their suppliers will have to come from elsewhere,” the firm said.
Long-term confidence remains intact
Not all investors are turning cautious.
Madeleine Ronner, Senior Portfolio Manager at DWS, expects upcoming earnings commentary from hyperscalers to continue supporting additional AI investment.
“The surprise would be if it’s not like that,” she said.
Although DWS has taken some profits following the strong rally in semiconductor shares, the firm continues to maintain an overweight position in the sector while selectively adding exposure to industrial and electrical equipment companies.
Data centre expansion faces new obstacles
Beyond financing, regulatory and community opposition is emerging as another potential headwind.
Empirical Research estimates that around 70% of planned U.S. data centre projects now face some degree of local resistance.
New York recently became the first U.S. state to introduce a one-year moratorium on large new data centres, citing concerns over electricity demand, water usage and local infrastructure.
Despite those challenges, investor appetite for AI infrastructure remains robust. Morningstar data show semiconductor-focused funds attracted a record $10 billion of net inflows through May.
Fidelity Investments Director of Global Macro Jurrien Timmer believes recent volatility should be viewed as part of a normal technology cycle.
“The AI story is well known, it’s ongoing, the earnings are still supporting the trend,” Timmer said.
He also argued that investors should broaden their exposure beyond AI hardware alone.
“I want to participate in the boom, but I also want to protect myself in case that boom is overdone,” Timmer said.