Key Market Debate: Has AI Capex Peaked or Is This Just a Mid-Cycle Pause?

Deep News21 minutes ago

Financial markets are behaving more and more as though the AI capital expenditure cycle has already topped out, even though there is still no convincing "real-economy evidence" to support that conclusion.

According to Jefferies, the central disagreement in the market right now is whether this marks the genuine peak of the AI capex cycle or simply a mid-cycle correction within an ongoing upswing. In a research note dated November 19, the firm said multiple risk signals are now building up at the same time around this core question.

Nvidia's role as the "central bank" of the AI ecosystem is facing pressure from a governance shift, and any contraction in its enormous off-balance-sheet financial exposure would tighten the entire AI financing chain in a systemic way. At the same time, the earnings quality of hyperscalers is coming under stricter market scrutiny: after stripping out non-cash paper gains tied to AI-related investments, their core profit growth is far below headline figures. In addition, the U.S. Commerce Department's Bureau of Industry and Security (BIS) has announced licensing requirements for AI chip exports to several Southeast Asian countries and Brazil, further reshaping the geographic distribution and pace of global computing capex expansion.

The firm believes these three threads are intertwined and together form the core dispute over how to characterize the current AI capex cycle.

Market signals versus cycle characterization: pricing moves ahead of hard evidence

Jefferies' report states plainly that financial market behavior increasingly resembles an AI capex peak, yet convincing "ground-level evidence" to confirm that view is still lacking. The report follows the analytical approach of its earlier November 5 note, tentatively taking market signals as the guide and framing the current situation as a contest between two scenarios. The first is that this is the true peak of the AI capex cycle; the second is that it is merely a mid-cycle correction within an upswing. Until more fundamental evidence emerges, the report argues, this judgment remains unresolved and is the single most important characterization question facing investors today.

Nvidia's "central bank" role: huge off-balance-sheet exposure and governance shift risk

Over the past two years, Nvidia has acted deeply as both financier and guarantor for the entire AI ecosystem, with a financial exposure far larger than the market had generally recognized. According to a November 12 breakdown by Tomasz Tunguz, Nvidia's related commitments include: US$42 billion in technology acquisition commitments; US$61 billion in equity investments covering 229 startups and 13 listed companies; US$164.5 billion in lease guarantees and cloud computing capacity backstops, requiring corresponding capital reserves; and US$63.1 billion in customer receivables, with payment terms stretching as long as one year.

Governance changes are drawing market attention. Colette Kress, who served as Nvidia's chief financial officer for 13 years and now also serves as interim CEO, has taken a reserved stance on these large-scale financial risk exposures. The report argues that if Nvidia shifts from an aggressive matchmaking model toward more prudent financial discipline, it would, as the leader of AI trades, transmit a clear tightening effect across the entire AI capex and financing chain.

Neoclouds are under particularly acute pressure. Nvidia previously viewed neoclouds, meaning data center startups, as an important source of alternative computing demand beyond hyperscalers. Notably, major hyperscalers, including Google parent Alphabet's TPU, are all advancing their own in-house chips. As the related turmoil has intensified, CoreWeave's share price has pulled back 25% since the news emerged, while the yield on its dollar bonds has risen to around 15.7%, reflecting strong credit and viability pressure.

The report argues that if the most speculative slice of computing demand shrinks, it could push down computing prices for other participants, in turn creating repricing pressure across the entire AI computing supply-demand landscape.

Breaking down hyperscaler profits: a "reflexivity" warning on capex returns

The Jefferies report conducts a deep breakdown of the profit structure of four hyperscalers, Microsoft, Amazon, Alphabet and Meta, revealing that earlier earnings growth relied heavily on non-cash paper gains from AI-related investments rather than substantive improvement in core operating profit.

The "other non-operating gains and losses" line tracked in the report mainly includes unrealized paper gains and losses on unlisted AI-related investments and other non-cash items. Its quarterly aggregate swung abruptly from +US$155 billion in 2Q26 to -US$9 billion in 3Q26, directly reflecting the paper giveback caused by recent declines in AI-related asset prices.

On an annualized basis, the data is equally telling: in the twelve months through 3Q26, the combined annualized net profit of the four rose 73% compared with the twelve months through 1Q25, an increase of US$248 billion (from US$340 billion to US$588 billion). However, within that increase, annualized "other non-operating income" surged 18-fold from US$12 billion to US$209 billion, contributing the vast majority of the gain (US$198 billion); excluding that item, the remaining annualized net profit grew only 15%, an increase of US$51 billion (from US$329 billion to US$379 billion).

The report notes that this highlights the amplifying effect of the prior AI bull market's "reflexivity" on book profits, and makes the market increasingly sensitive to the actual return on enormous capital expenditure. As AI-related asset prices adjust, that "reflexivity" is now working in the opposite direction.

Meanwhile, the U.S. Commerce Department's Bureau of Industry and Security (BIS) has announced export licensing requirements for AI chip exports to Southeast Asian countries and others. The combination of tighter export controls, escalated regulatory scrutiny and a tighter financing environment means cross-border computing expansion and some data center projects that rely on external funding will face higher compliance and funding hurdles. The report argues this will affect the geographic distribution and expansion slope of global AI capex.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment