A Jefferies report said the AI capex cycle may have peaked, BW Businessworld and the Economic Times reported on November 7 (BW Businessworld; Economic Times). “All of the above suggests the AI capex cycle may have peaked, or at least commenced a mid-cycle correction”, the report said, according to the Economic Times.
The report said Microsoft, Amazon, Alphabet and Meta raised their combined capital expenditure by 108% year on year to $202 billion in the third quarter of 2026. Consensus estimates put the four companies’ combined capex at $1.04 trillion next year, and Jefferies said 2027 AI capex could fall short of current expectations.
Business Pill · Credit Spreads
An evergreen Business Engineer explainer on credit spreads. It teaches the general idea only and does not discuss the Jefferies report.
The key insight: A Jefferies report said the AI capex cycle may have peaked after Microsoft, Amazon, Alphabet and Meta spent a combined $202 billion in the third quarter, up 108%, and that 2027 capex could fall short of the $1.04 trillion consensus, BW Businessworld and the Economic Times reported.
The brokerage said financial markets were increasingly signalling reluctance to fund AI-related investments, BW Businessworld reported. It pointed to rising borrowing costs: bond yield spreads of the major hyperscalers have widened, and credit default swap spreads of Oracle and CoreWeave have increased sharply.
We reported this week that Meta’s 2031 bonds were priced at a 5.85% yield while CoreWeave’s 2031 notes traded at 16.3% (our piece), and on the record in SpaceX’s credit default swaps (our piece).
AI-related investment contributed 1.39 percentage points, or 67%, of US real GDP growth of 2.1% in the four quarters to the third quarter of 2026, according to a Jefferies report as reported by BW Businessworld and the Economic Times. The remaining ~0.71 points is our arithmetic.
The US Growth Link
Jefferies said AI-related investment contributed 1.39 percentage points, or 67%, of US real GDP growth of 2.1% in the four quarters to the third quarter of 2026. It said a credit-driven slowdown in AI spending could weigh on US growth and shift expectations for Federal Reserve policy from tightening towards easing.
The brokerage said investors should look for opportunities beyond the AI supply chain, particularly in healthcare and biotechnology, where AI applications in drug development could offer significant potential.
The Portfolio Change
Jefferies raised India’s weight in its Asia Pacific ex-Japan relative-return portfolio by two percentage points, from 12% to 14%, keeping its Overweight stance, the reports said. It cut South Korea and Taiwan by one percentage point each, keeping both at Underweight.
South Korea and Taiwan are home to the memory and foundry makers at the centre of this week’s chip sell-off. We reported on Taiwan’s short-selling curbs after the TSMC rout (our piece) and on a broker’s reading of what Samsung and SK Hynix shares price in (our piece).
Our Numbers
By our arithmetic, the four companies’ $202 billion third-quarter capex is an annualised pace of $808 billion, so the $1.04 trillion 2027 consensus is about 29% above that pace. AI investment’s 1.39 points leaves about 0.71 points of the 2.1% growth to everything else, also our arithmetic.
The same four spenders were on a $730 billion capex path when we compared them with US energy investment (our piece), and GPU-backed loans reached $48.4 billion this year in one tracker (our piece).
The Structural Read
As we read the report as described, Jefferies’ case rests on funding rather than demand: it cites widening hyperscaler bond spreads and sharply higher Oracle and CoreWeave credit default swap spreads.
As we read the portfolio change, Jefferies moved two points into India and took one point each from South Korea and Taiwan, home to the memory and foundry makers at the centre of this week’s chip sell-off.
By our arithmetic, the $1.04 trillion 2027 consensus is about 29% above the $808 billion annualised pace of the third quarter, so the consensus assumes spending above the third-quarter pace.
Jefferies report, as quoted by the Economic Times
“All of the above suggests the AI capex cycle may have peaked, or at least commenced a mid-cycle correction”
Three Implications
FOR HYPERSCALERS Jefferies said bond yield spreads of the major hyperscalers have widened, BW Businessworld reported.
FOR CHIP-HEAVY MARKETS Jefferies cut South Korea and Taiwan by one percentage point each in its Asia Pacific ex-Japan portfolio, keeping both at Underweight.
FOR THE US ECONOMY Jefferies said AI-related investment was 1.39 points of 2.1% US real GDP growth, and that a credit-driven slowdown in AI spending could weigh on growth.
The analysis says: “The job of any operator, investor, or builder is to know which layer they are operating inside and which layer the binding constraint is rotating toward next.”
As we read it, Jefferies is pointing at capital as the binding constraint: the cost of funding the build, not the demand for it. The analysis does not discuss the Jefferies report.
What Is Not Established
We have not read the Jefferies report itself; its figures and quotes here are as reported by BW Businessworld and the Economic Times. Neither report names the author or gives the report’s date, and neither gives the new South Korea or Taiwan weights. Jefferies’ view on 2027 capex is its forecast, not a result. This piece makes no forecast.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
A Jefferies report said the AI capex cycle may have peaked after Microsoft, Amazon, Alphabet and Meta spent $202 billion in the third quarter, and that 2027 capex could miss the $1.04 trillion consensus. It raised India to 14% of its Asia Pacific ex-Japan portfolio and cut South Korea and Taiwan by a point each.
A note on sourcing. The Jefferies figures, views and the quote are as reported by BW Businessworld and the Economic Times on November 7; we have not read the report itself. The annualised pace, the gap to consensus and the remaining growth points are our arithmetic. This is not investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
