After nearly two years of relentless enthusiasm for artificial intelligence infrastructure, some institutional investors are beginning to reposition their portfolios as expectations grow that the explosive pace of AI spending by major cloud providers may start to moderate.

The shift does not necessarily signal weakening demand for AI. Instead, investors increasingly believe that the rate of spending growth—rather than spending itself—may begin slowing as hyperscale technology companies move from rapid infrastructure expansion toward more disciplined capital allocation.

According to UBS, capital expenditures by major cloud providers are expected to reach approximately $673 billion this year, representing 76% annual growth. However, forecasts suggest that spending growth could slow to 25% next year before easing further to around 6% by 2028.

This changing outlook is prompting portfolio managers to rethink which companies are likely to benefit most from the next phase of the AI cycle.

During the initial AI boom, semiconductor manufacturers and infrastructure suppliers emerged as the biggest winners. Companies such as Nvidia, Broadcom, Micron, TSMC, and ASML experienced extraordinary gains as demand for AI chips and data center equipment surged.

Now, some investors are rotating capital toward the companies financing those investments—often referred to as hyperscalers—including Microsoft, Amazon, Alphabet, and Meta. As infrastructure investment becomes less aggressive, these companies could benefit from improving free cash flow and stronger long-term profitability.

The shift is also extending beyond technology. Portfolio managers are increasing exposure to software companies expected to monetize AI applications, as well as sectors such as healthcare, financial services, cybersecurity, industrial automation, and data center cooling technologies that stand to benefit from broader AI adoption.

Despite recent volatility, enthusiasm for AI infrastructure remains substantial. Semiconductor stocks have more than doubled over the past year, although many investors now question whether current valuations fully reflect future growth prospects. Bank of America recently found that semiconductors remain the most crowded investment theme among global fund managers.

Another factor influencing sentiment is financing. Major technology companies are increasingly relying on debt markets to fund large-scale AI infrastructure projects, leading some analysts to question whether capital spending can continue accelerating indefinitely. At the same time, environmental concerns and regulatory scrutiny surrounding large data centers—including restrictions recently introduced in parts of the United States—could further moderate future expansion.

Nevertheless, many market participants remain optimistic. Strong corporate earnings, growing enterprise AI adoption, and continued demand for computing power suggest that AI investment is entering a more mature phase rather than approaching an end.

Instead of focusing solely on companies building AI infrastructure, investors are increasingly looking toward businesses that can generate sustainable revenue from deploying artificial intelligence across products, services, and industry-specific applications.