The Capital Cycle is the official podcast that Marathon Asset Management (the London firm founded in 1986) launched in 2024, hosted by financial historian Edward Chancellor, who interviews Marathon's investors about each Global Investment Review letter — applying the firm's long-term, contrarian "capital cycle" supply-side approach.
In an October 2026 podcast, Jefferies global equity strategy head Chris Wood argued that massive spending on AI infrastructure is driving a classic capital cycle. He warned that such cycles often end in overcapacity and falling returns. Citing more than $200 billion in combined AI-related capital expenditure by four technology giants in 2024–2025, Wood questioned whether the investments could generate adequate returns within a reasonable timeframe. He cautioned investors about the risks of industry consolidation and value destruction after the cycle peaks.
At a Glance In a presentation at Jefferies, Chris Wood discussed the artificial intelligence (AI) capital cycle. He argues that the massive capital expenditures (capex) currently underway in AI are driving a typical capital cycle, and historical experience suggests such cycles often end in overcapac
This section is hosted by Edward Chancellor, who invites Chris Wood, Head of Global Equity Strategy at Jefferies, to discuss the capital cycle in the artificial intelligence (AI) sector. Wood's core argument is that the massive capital expenditure currently underway in AI infrastructure is driving a typical capital cycle, and historical patterns suggest such cycles often end with overcapacity and declining returns. He warns investors to be wary of industry consolidation and value destruction after the peak of the capital cycle.
Edward Chancellor notes that the AI sector is currently experiencing an unprecedented wave of capital expenditure, which bears similarities to several historical capital cycles, such as the railroad boom and the internet bubble. He invites Chris Wood to analyze the sustainability of AI investment from a capital cycle perspective.
Chris Wood believes that AI is in the early stages of a typical capital cycle, characterized by the following core features:
1. Massive Capital Expenditure: Technology giants (such as Microsoft, Google, Amazon, and Meta) have significantly increased their investments in AI infrastructure (data centers, GPU chips, power facilities), and this spending is still accelerating.
2. Supply Expansion Outpacing Demand: Historical experience shows that in a capital cycle, supply often expands excessively, while demand growth may fail to keep pace, leading to overcapacity.
3. Risk of Declining Returns: As competition intensifies and capital floods in, the average industry return on investment will tend to decline, potentially triggering consolidation and value destruction.
Wood cites several historical capital cycle cases to support his assessment:
| Historical Cycle | Core Characteristics | Outcome |
|---|---|---|
| 19th Century Railroad Bubble | Overbuilding, duplicative investment | Overcapacity, wave of bankruptcies, heavy investor losses |
| 2000 Internet Bubble | Overinvestment in telecom and fiber optic infrastructure | Bubble burst, massive company failures, capital destruction |
| 2010s Shale Oil Boom | Capital influx, surge in production | Price collapse, industry consolidation, poor investor returns |
Wood argues that the AI capital cycle is likely to follow a similar pattern: the current investment frenzy is driven by optimistic expectations of AI's future potential, but history has repeatedly shown that when a large amount of capital pours into an emerging field, it often ends with supply gluts and deteriorating returns.
Wood acknowledges that AI technology itself may be transformative, but he emphasizes that the laws of the capital cycle do not change due to the novelty of the technology. He distinguishes between "technological potential" and "investment returns"—even if AI ultimately changes the world, it does not mean that all current investors will profit.
Key Uncertainties:
This section does not discuss specific companies or positions; it only uses the overall industry and technology giants as case studies for the capital cycle analysis.