The AI Spending Plateau: A Tectonic Shift in Tech’s Ecosystem

A recent surge in valuations of AI chipmakers has given way to a more cautious approach by investors. As influential players in the tech industry position themselves for slower growth, a significant shift is taking place in the ecosystem.

Hyperscalers such as Microsoft, Amazon, Alphabet, and Meta have driven AI spending with massive investments in data centers and infrastructure. However, this assumption of accelerating investment has become untenable, given UBS’s estimate that hyperscalers’ capital expenditures will rise only 25% next year and just 6% in 2028.

This shift has far-reaching implications. Investors like Alexis Bossard at Edmond de Rothschild Asset Management and Alberto Conca at LFG+ZEST are reducing their exposure to semiconductor stocks while increasing their holdings of hyperscalers themselves, indicating a fundamental change in the tech landscape. The nature of AI adoption is being reevaluated.

The Philadelphia Semiconductor Index has more than doubled over the past year, driven by expectations of continued large-scale spending on data centers and AI infrastructure. However, with major players hinting at a slowdown, this rally is losing steam. Bank of America’s July fund manager survey found that 82% of respondents viewed semiconductors as the most crowded trade, signaling that investors are acknowledging the reality of slowing AI spending growth.

As hyperscalers adapt to slower growth, they will need to rely more heavily on external financing for their AI ambitions. This raises questions about capital-market pressures and how they may constrain future spending growth.

The shift in AI spending is not just a slowdown; it’s a fundamental change in the way hyperscalers approach this technology. Bossard noted that once capex stops increasing, “it will definitely be a relief for hyperscalers and a negative signal for the semi industry.” This marks a move away from pouring more money into data centers or semiconductor companies and toward finding new ways to deploy AI in areas like cybersecurity, liquid cooling, and selected software firms.

The question remains: what happens if AI spending continues to grow but at a slower pace? Will hyperscalers find new cost-cutting measures, or will they need to abandon their aggressive expansion plans altogether? The answers lie in the nuances of this shift, and it’s too early to determine which direction the market will take.

One thing is certain: the tech industry stands at a crossroads. As investors position for slower AI spending growth, a profound change is underway in the dynamics between hyperscalers and semiconductor companies. This shift is not just about valuations or market sentiment; it’s about the future of AI itself – and how it will be deployed in the years to come.