The most recent quarterly earnings season revived investor excitement around artificial intelligence, but also underscored a clear tension: spending to build out the infrastructure that supports AI is lifting expectations for long-term gains even as it threatens near-term margins, according to a client note from analysts at Aubrey Capital Management.
Investors are looking for clear returns from sizable capital deployments, particularly among the large cloud providers often labeled "hyperscalers" - companies that are integrating AI across their product suites and expanding data-center footprints to support the compute those models require.
The market reaction has been uneven. One prominent example the analysts highlighted is Oracle, whose stock fell roughly 44% after a spike in September. That earlier jump followed Oracle's announcement of a $300 billion contract with OpenAI, the developer of ChatGPT. Traders later expressed doubts about OpenAI's ability to uphold the commitments and about Oracle's strategy of using debt to fund a rapid data-center build-out, prompting the share-price reversal.
In their note Aubrey's team argued that the earlier broad market enthusiasm for AI-related capital expenditures has become more conditional. "This trend has been broken, and not all capex is created equal," they wrote. The analysts said the market is differentiating between companies that deploy existing cash reserves to fund AI investments and those that take on new debt to finance them - the former are generally receiving positive investor recognition, the latter are facing closer scrutiny.
Despite the growing selectivity, Aubrey's analysts said much of the increase in hyperscaler spending is supported by strong operational results. They point to double-digit percentage growth in cloud revenues at Microsoft and Alphabet and to a 26% expansion in Meta's core advertising business. The combination of improving top-line performance and robust operating cash flow at the largest names has been central to underwriting increased AI capital spending.
Quantifying that concentration, Aubrey estimated that hyperscaler operating cash flows account for about 60% of total AI spending. They also projected that Meta, Microsoft, Alphabet and Amazon together are expected to commit more than $480 billion in capital expenditures in 2026 - a sum Aubrey described as roughly 60% of industry-wide AI spending. The analysts added that many participants expect that figure to climb further "as we close out the year."
Looking ahead, Aubrey said the influx of hyperscaler investment should help support fundamentals into the start of the next year. Yet the bank of questions remains: specifically, how quickly enterprises will adopt these AI capabilities at scale, and at what point the technology will be capable of replacing a broader share of the workforce. Those uncertainties, the analysts said, continue to shape investor judgement about the sustainability of returns from heavy AI-related capex.
Contextual note - Aubrey's commentary centers on recent earnings reactions, capital allocation strategies among large cloud providers, and the implications of funding choices for investor sentiment. The analysis focuses on operational performance metrics and projected capex commitments from the largest hyperscalers without extending beyond the data and statements provided by the firm.