AI capital spending is approaching a historic turning point, and the fate of US economic growth may be deeply tied to it.

According to a Jefferies report, strategist Christopher Wood warned that financial markets have already sent a clear signal that the AI capex cycle may be topping out through widening credit spreads and sustained declines in semiconductor stocks. The macroeconomic implications of this judgment extend far beyond the technology sector itself.

According to the Jefferies report, in the four quarters through the third quarter of 2026, US AI-related real capital spending contributed 1.39 percentage points to year-on-year real GDP growth, accounting for 67% of the overall 2.1% expansion. At the same time, investment in information processing equipment and software as a share of nominal GDP has risen to 5.35%, surpassing the 4.46% peak during the 2000 internet bubble and setting a record high. This means that once AI capital spending contracts materially, US economic growth will face a direct and significant downside shock.

The market has already reacted first. Semiconductor stocks remain under pressure, and clear cracks have appeared in the credit market: the 10-year corporate bond spreads of Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) and Meta (NASDAQ: META) have widened by 16 basis points, 16 basis points and 35 basis points, respectively, from recent lows. The 5-year CDS spreads of Oracle (NYSE: ORCL) and CoreWeave (NASDAQ: CRWV) have surged from early-June lows to 329 basis points and 1,163 basis points, respectively. Meanwhile, after the US midterm elections, Democrats regained control of Congress, and the market worries that their regulatory stance on data centers and AI models will further suppress AI spending expectations.

Credit markets are flashing warnings, and consensus AI spending forecasts face downward revision pressure. Turmoil in the credit market is the most direct market evidence behind the view that this round of AI capex is peaking. The Jefferies report noted that the best way to assess AI capex risk is to continuously track AI-related credit spreads, and current spread movements have already sent a clear tightening signal.

Specific data show that the 10-year corporate bond yield spreads of Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) and Meta (NASDAQ: META) have risen from recent lows of 73 basis points, 65 basis points and 94 basis points on October 12 to 89 basis points, 81 basis points and 129 basis points, respectively. The 5-year CDS spreads of Oracle (NYSE: ORCL) and CoreWeave (NASDAQ: CRWV) surged from early-June lows of 153 basis points and 452 basis points to highs of 329 basis points and 1,163 basis points in late October. The 5-year CDS spread of SoftBank (TSE: 9984) also rose from a late-June low of 278 basis points to 442 basis points.

The core implication of this signal is that the market is showing an increasingly obvious decline in willingness to finance AI-related investment. The Jefferies report argues that this means the market consensus forecast for combined capital spending by the four major hyperscale cloud service providers—Microsoft (NASDAQ: MSFT), Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) and Meta (NASDAQ: META)—of about $1.04 trillion in 2026 faces a material risk of downward revision.

Hyperscaler free cash flow is under strain, and AI is eroding traditional businesses. The latest quarterly earnings have further intensified market doubts about the sustainability of AI capital spending. Combined third-quarter capital spending by the four major hyperscale cloud providers rose 108% year on year to $202 billion, but at the same time, free cash flow at Amazon (NASDAQ: AMZN) and Alphabet (NASDAQ: GOOGL) turned negative at -$35 billion and -$9 billion, respectively.

What has made the market even more alert is that AI is beginning to cannibalize the traditional core businesses of hyperscale cloud providers. Growth in Google Search revenue at Alphabet (NASDAQ: GOOGL) slowed sharply from 19% year on year in the first quarter of 2026 to 12% in the third quarter, and search revenue as a share of Alphabet's total revenue fell below 50% for the first time to 49.7%. At the same time, Google Cloud revenue surged 103% year on year to $30.7 billion, with its share rising from 20.7% to 24.1%. Online store revenue growth at Amazon (NASDAQ: AMZN) plunged from 15% year on year in the second quarter to 2% in the third quarter, while AWS revenue rose 39% year on year to $46 billion.

The Jefferies report argues that the core motivation for hyperscale cloud providers to invest heavily in AI is precisely fear that AI will disrupt their own traditional businesses. This is a defensive rather than offensive spending logic, trapping them in a prisoner's dilemma of "wait and die without spending, but spending does not necessarily guarantee survival."

The circular financing system is expanding, and "too big to fail" risk is emerging. The Jefferies report specifically highlighted an increasingly large structural risk in the US AI ecosystem: the circular financing system built around AI capital spending.

The remaining performance obligations (RPO), or contracted revenue backlog, of the four major hyperscale cloud providers—Microsoft (NASDAQ: MSFT), Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) and Oracle (NYSE: ORCL)—rose to $2.5 trillion at the end of last quarter. At the same time, data center financing is increasingly being conducted off balance sheet, with hyperscalers partnering with private credit to set up "special purpose" investment vehicles.

CoreWeave (NASDAQ: CRWV) completed a $3.1 billion delayed draw term loan (DDTL 5.0) financing in May this year, the first known GPU-collateralized financing tool publicly syndicated through a financing subsidiary. Its yield to maturity has risen from 7.62% to 8.36%. It is estimated that annual new issuance of data center asset-backed securities reached $27 billion in 2025 and is expected to rise to $40 billion in 2026.

The Jefferies report stated bluntly that the above system is financially extremely risky unless it is assumed that some parts of the US AI ecosystem are already regarded as "too big to fail."

Microsoft layoffs and China's AI cost reduction point to an industry paradigm shift. The potential turning point in the AI capex cycle has already left its mark at the corporate behavior level. Microsoft (NASDAQ: MSFT) announced voluntary exit packages for about 60,000 eligible employees, a number equivalent to more than a quarter of its global workforce. The Jefferies report argues that the direct motivation for this move is cost cutting under pressure from an ever-expanding capital spending bill, even though Microsoft's third-quarter free cash flow was still positive at $18.6 billion.

At the same time, Desh Peramunetilleke, Jefferies' global head of quantitative strategy, noted in a mid-October report that consensus earnings growth expectations for global AI stocks are expected to slow sharply from about 100% in 2026 to 30% in 2027 and 22% in 2028, and advised investors to begin diversifying beyond AI.

In China, the logic of cost reduction has run through the AI boom from the very beginning. Edison Lee, Jefferies' head of China technology and telecom research, pointed out that Chinese AI labs have always focused on computing efficiency, driven both by compute scarcity and by the fact that Chinese AI has been oriented toward commercial deployment from the start rather than pursuing AGI, Silicon Valley's ultimate goal. This pragmatic path may become the direction the US AI industry is forced to emulate under pressure from tighter financing.

If AI spending slams on the brakes, US Treasuries may fail as a "safe haven." The Jefferies report issued a clear warning about the macroeconomic consequences of a sharp contraction in AI capital spending: it would deal a direct negative shock to US economic growth and trigger a rapid reversal in Federal Reserve policy expectations from tightening to easing.

However, the report also noted that under this scenario, US Treasuries would be unable to play their traditional role as a "safe-haven asset." The reason lies in the "fiscal dominance" effect—the US fiscal deficit is already 6.2% of GDP, and historical patterns show that recessions typically widen fiscal deficits by about 6 percentage points. Once an economic downturn occurs, the market will quickly shift its attention to the prospect of a sharp deterioration in fiscal revenue, and the US Treasury market will then behave more like a troubled emerging market than a traditional safe asset.

The report also cited Amara's Law to remind investors that the short-term impact of new technology is often overestimated, while the long-term impact is underestimated. Current market trends suggest that global short-term enthusiasm for AI may have peaked. For investors, as uncertainty around the AI capex cycle intensifies, assets with lower correlation to AI capital spending, such as gold and biotechnology, may deserve more attention.