AI Chip Rally Hits Turbulence as Investors Eye Hyperscaler S

Investors are recalibrating their bets on the AI infrastructure boom, rotating out of semiconductor stocks and into the hyperscalers themselves as the hyperscaler spending slowdown comes into view. The Philadelphia Semiconductor Index has fallen from its June peak, evidence that the market is pricing in a tempering of the breakneck spending pace that defined the AI buildout over the past two years. The rotation is one of the most significant capital allocation shifts in the technology sector since the AI spending cycle began.

UBS estimates that the hyperscaler spending slowdown will bring capital expenditure growth down from 76% this year ($673 billion) to 25% in 2027 and 6% in 2028. The trajectory suggests the near-trillion-dollar investment cycle is approaching an inflection point where deployment scales meet demand realities.

Moody's Ratings has sized the combined capital spending of the six largest US hyperscalers (Microsoft, Amazon, Alphabet, Meta, Apple, and Oracle) at roughly $700 billion this year, nearly six times the level recorded in 2022. The ratings agency noted that companies are pacing data center construction and chip orders to limit the risk of overbuilding, a signal that even the biggest spenders are looking for guardrails.

Bond market data reinforces the caution around the hyperscaler spending slowdown. Apollo Global Management calculated that bond cover ratios for hyperscaler debt dropped from 5 times in February to under 2 times in July, indicating that investor appetite for financing further expansion is narrowing. New York's imposition of a one-year moratorium on large new data centers adds a regulatory dimension to the supply-side risks that hyperscalers must manage.

The shifting investor sentiment becomes clearer when examining fund flow data. Morningstar figures show that chip-focused funds drew a record $10 billion in net inflows through May, capturing the euphoria that surrounded AI semiconductor names earlier this year. But the rotation now underway suggests that institutional money is moving downstream. Investors are increasingly favoring the hyperscalers themselves, the companies that ultimately foot the AI infrastructure bill, over the chip suppliers that have reaped the most dramatic revenue gains from the cycle.

The logic behind this rotation is straightforward. If hyperscaler capex growth slows from 76% to single digits in two years, the chipmakers that have been selling into that buildout face a demand cliff. The hyperscalers, by contrast, will own the infrastructure assets capable of generating long-term returns from AI services, even if the pace of new construction moderates. For portfolio managers, this shifts the risk-reward calculation from a growth-at-any-price thesis toward one that weights capital efficiency and free cash flow generation.

This recalibration is happening against a broader backdrop of regulatory and fiscal scrutiny. New York's data center moratorium, while localized, signals that municipal and state governments are beginning to question the environmental and grid capacity implications of massive AI infrastructure projects. Similar conversations are emerging in Europe, where energy costs and sustainability targets are forcing hyperscalers to rethink site selection and power procurement strategies.

UBS projections for a 6% capex growth rate by 2028 would be a dramatic normalization from the current trajectory, but they do not signal a collapse in AI investment. A near $700 billion annual run rate, even growing modestly, still is an enormous commitment of capital to AI infrastructure. The key question for investors is whether the semiconductor companies that have captured the bulk of AI-related revenue can sustain their margins and growth rates as the buildout transitions from acceleration to deceleration.

For the hyperscalers themselves, slower capex growth carries a silver lining. Companies like Microsoft, Amazon, and Alphabet have faced persistent questions from analysts about the return on their AI investments. A moderation in spending allows these firms to demonstrate that their capital deployment is generating measurable revenue and operating income, rather than requiring ever-larger rounds of funding. The bond market's tightening of cover ratios effectively reinforces this discipline: investors who finance hyperscaler debt are demanding clearer evidence that the spending translates into cash flow.

The data center buildout is not stopping. Moody's assessment that the six largest hyperscalers will spend roughly $700 billion this year, nearly six times 2022 levels, confirms that AI infrastructure remains a top corporate priority. But the shift from 76% growth to 25% and then 6% changes the competitive dynamics across the AI supply chain. Chipmakers that have enjoyed pricing power and allocation leverage may find those advantages eroding as hyperscalers gain confidence that their existing capacity is sufficient to meet near-term demand. The divergence between chip stock performance and hyperscaler equity performance in recent weeks suggests that markets are already pricing in this transition.

What This Means for the AI Investment Thesis

The hyperscaler spending slowdown narrative does not imply that AI infrastructure investment is contracting. A near $700 billion annual run rate still is enormous continued deployment. But the shift from accelerating growth to decelerating growth changes the calculus for which companies benefit most. Semiconductor suppliers have been priced for perpetual acceleration; the market is now rewarding the asset owners who can monetize what has already been built.

For technology leaders and investors, the key insight is that the AI buildout is maturing from a construction phase into an operations phase. The winners in the next cycle may be those who operate the infrastructure, not those who sell the picks and shovels. Companies positioned to deliver AI services at scale through existing capacity have a clearer path to profitability than those dependent on continued expansion of capital budgets.

Why This Matters

The rotation from chipmakers to hyperscalers signals that the market believes the frontier of AI value creation is shifting from hardware supply to service delivery. If capex growth continues to decelerate as projected, the semiconductor companies that rode the initial wave will need to demonstrate that their revenue streams are sustained by recurring demand rather than a one-time buildout. For hyperscalers, the moderation in spending growth offers a path to improving free cash flow and returns on invested capital, precisely the metrics that bond markets have started to scrutinize more closely. The hyperscaler spending slowdown, if it unfolds as UBS projects, will test whether the AI industry can deliver profitable returns on the largest infrastructure build in technology history.

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team.

What This Means for the AI Investment Thesis

Why This Matters

Related Articles