How to Profit When Cycles Turn – Django Davidson on Capital Cycle Investing

Capital-cycle investing starts with measurable supply, not demand forecasts, which Django Davidson considers prone to storytelling. When capacity is withdrawn and new investment stays scarce, competition can ease and returns improve; when capital floods into an opportunity, returns tend to fall. Because companies and investors respond slowly, cycles take time to turn. The framework can improve the odds of investing well, but deciding when to leave an upswing or buy after capacity exits remains an art, not a precise science. Davidson sees higher interest rates as reviving creative destruction after years of near-zero rates, though how high rates may go is uncertain.

Davidson is cautious about US technology as some businesses move from capital-light, high-return models toward capital-heavy investment. Hyperscalers have shifted from buybacks to share issuance and from little debt to more borrowing. They went from about 1% of investment-grade debt issuance in 2023 to a fifth of all credit issuance, while moving from $200–$300 billion in annual free cash flow to negative free cash flow as they built AI data centers. Davidson compared AI investment as a share of the economy with the US railroad boom of the mid-1800s; projected investment of $3–$4 trillion in his “AI Paradox” piece a year earlier had more than doubled in subsequent estimates.

The key uncertainty is whether AI computing remains an oligopoly or becomes a commoditized, capital-intensive, low-return business. Davidson sees data centers as factories, making factory economics relevant: China processes three quarters of global AI tokens at about one fifth of the total cost. If customers choose the lowest-cost provider, US capacity built to cover fixed costs could intensify competition with China and pressure returns. A low valuation alone is not a reason to buy when excess capacity and fixed costs threaten profits. In commodities, low valuations can be more compelling when capacity has been removed and future supply is constrained.

The firm’s early memory-semiconductor investments were based on consolidation, higher fabrication costs and rising entry barriers, not a forecast that memory would become an AI bottleneck. The industry narrowed from about 20–22 players in the 1990s to three main manufacturers, creating the potential for supply limits to benefit from unexpected demand such as AI. The firm added as the cycle ebbed, trimmed as it advanced and sold more aggressively over roughly the past six months. Memory companies now trade at three or four times the estimated cost to rebuild their infrastructure; Chinese IPO capacity and announced US reshoring make Davidson judge the cycle closer to its top than its bottom. European chemicals look cheap relative to replacement cost, but the firm underestimated Chinese capacity, which has overwhelmed much European commodity-chemical production. Without serious European anti-dumping measures and tariffs, he expects many producers to struggle to make money soon. The firm shifted a small chemicals position toward specialty and distribution businesses with stronger barriers and valuations below replication cost.

Davidson sees commodities and energy, after prolonged underinvestment, in the early innings of their cycle. Platinum is a supply-driven opportunity: it is about 120 times less prevalent in the earth’s crust than gold, and supply is concentrated among about four dominant South African companies. Their capital spending fell from about $600 per ounce roughly 10 years ago to about $300, while demand remains substantially linked to internal-combustion engines and the pace of electric-vehicle adoption is uncertain. Copper supply is also constrained, with difficult projects and disappointing Chilean mine output. Davidson expects prices to rise enough to bring higher-cost, smaller mines online, while recognizing that higher prices eventually attract supply and cure high prices. He is less focused on gold because its case depends more on demand, including dollar debasement, than on acute supply bottlenecks.

The portfolio holds about 400 stocks expressing roughly 20–25 themes, allowing cycles to mature at different times and small positions in areas concentrated investors may avoid. Davidson said the average holding period is just under a decade and that a shared capital-cycle approach among multiple portfolio counsellors provides second opinions, including help adding to a position after a fall. In Japan, he sees scope to improve returns by reducing invested capital tied up in excess cash, cross-holdings, unused plants and multiple head offices. Returns on capital for corporate Japan have been rising, and he believes the trend has further to run; the firm works with or alongside activists supporting buybacks, asset sales and other changes, with stock-exchange rules an important driver.

Davidson cautions that high profits and returns can mean-revert, although some companies resist the decline. He argued that high profits, price increases and competition amplified by social media weakened the customer proposition at Kraft Heinz and other consumer-goods companies. He described Nike as a 25-times-earnings, high-return business and argued that its valuation and profitability helped foster a culture detached from customers and poor decisions. Higher discount rates also favor companies earning money sooner over those promising distant profits. Davidson calls the capital-cycle framework a way to make contrarian decisions more rationally, not a reason to oppose the crowd automatically. Investors should set guardrails around their own tolerance for drawdowns: conservative investors should avoid positions likely to cause painful losses, while aggressive investors should be cautious about concentrated single-stock bets.