- The author sees a valuation divergence: infrastructure beneficiaries already price in AI applications generating enough returns, while application businesses may retain unrealized AI-driven value.
- Slower infrastructure-spending growth need not mean slower adoption; the thesis favors companies whose customers buy AI-enabled products for measurable economic benefits, including in software, biotech, and defense.
- If applications fail to create value, infrastructure expectations may also weaken, making select infrastructure shorts a potential hedge. Palantir remains a watchlist example, not an obvious bargain; the author prefers gradual accumulation as valuations reprice.
Investment pitches
filter:
63