Edward Chancellor and Alex Duffy discuss how the AI-driven capital expenditure cycle has become a dominant and risky thematic, with US economic growth heavily dependent on AI capex and emerging markets now overexposed to the tech supply chain. They draw parallels to the mining supercycle, warning of a fundamental earnings bubble and advocating for contrarian investments in neglected sectors like Chinese real estate and the Philippines.
Summarized by Podsumo
AI capex now accounts for an estimated 40-100% of US economic growth, with non-AI capex turning negative year-on-year, signaling a crowding out of real economy investment.
The AI theme has shifted to emerging markets, where Korea and Taiwan now represent 45% of MSCI EM index (up from 25% in 18 months), leading to increased correlation and diversification risk for global portfolios.
Despite seemingly low P/E ratios (e.g., Micron at 9x earnings), the tech supply chain shows a fundamental earnings bubble with price-to-book multiples surging 4-6x, comparable to the 2000s mining supercycle.
Hyperscalers' returns on incremental capital are collapsing (negative returns per some analysts), requiring $5 trillion in AI revenue within a few years to rationalize spending, creating vulnerability for Asian tech suppliers.
Alex Duffy favors contrarian investments in lower-halation sectors like Chinese real estate (trading at 75% of book value) and the Philippines (at 20-year lows), where capital cycle rationalization offers asymmetric upside.
"This is one trade with a market that's fixated on one singular thematic, and it's all in everywhere. So as a somewhat contrarian capital cyclist, this is quite a painful experience."
"The mistake that markets make is to say this is not bubble territory because valuations are like extrapolated. Today, the tech supply chain... it's an earnings bubble, not a valuation bubble."
"If you're wrong on the positive outcome, the drawdown will not be 10 to 15%, it'll be 50 to 60%."