How I Made A Fortune, Lost Most Of It, And Built It Back

EQIX is still the cleanest way to own the scarcity premium in digital infrastructure, but at this stage the upside is more about revision durability than multiple expansion. The key mechanism is switching costs: once network density and interconnection are embedded in workflows, pricing tends to be sticky, so the main risk is not demand disappearing but growth normalizing faster than the market expects.

Second-order beneficiaries are the picks-and-shovels around the AI network stack: opticals, interconnect, and high-speed networking vendors such as ANET and CRDO should keep seeing incremental demand even if compute capex becomes lumpier. The likely relative loser is lower-quality colo capacity with less network density and weaker balance sheets, where customers can negotiate harder if new supply comes online; that argues for EQIX over less differentiated peers rather than a blanket sector bet.

The contrarian concern is that "AI demand" is already doing a lot of work in the valuation. If hyperscalers increasingly bypass third-party colo for direct campus-to-campus connectivity, or if new capacity additions in 2H only partially absorb existing pricing, the market may discover that scarcity is cyclical, not permanent. Rising rates are the cleanest near-term catalyst against the bull case because duration-like REIT multiples can compress even when fundamentals remain healthy.

The thesis is most vulnerable over the next 1-2 quarters if bookings, churn, or same-store revenue fail to accelerate while capex stays elevated. Over 6-18 months, the question is whether EQIX converts AI-driven demand into sustained FFO per share growth faster than peers; if not, the premium multiple is the first thing to come under pressure.