⚠ Lock-In Is Real but Cyclical, Not PermanentModerate threat

The switching costs that make Bloom's installed base sticky are genuine, but they are cyclical rather than permanent, and mistaking one for the other would overstate the moat. Within a system's operating life, the cost and disruption of removing installed Energy Servers hold the customer in place, and the service relationship reinforces the hold. But at the boundaries — service-contract renewals, and the end of a system's useful life when replacement is required anyway — the lock-in dissolves and the customer freely re-evaluates Bloom against every alternative then available. The recurring revenue is thus a series of medium-term commitments that must be periodically re-won, not a permanent claim on the customer, and each renewal is a fresh competition on possibly-shifted terms.

This cyclicality of the lock-in matters more because the competitive landscape is not static. The alternatives to Bloom — grid power, gas turbines, batteries, rival fuel cells, and eventually small modular nuclear — are improving and cheapening over time, and the specific conditions that make Bloom compelling today (a power-starved grid, a premium on speed) may ease. So a customer locked in now might, at renewal or replacement years hence, find the alternatives more attractive and choose differently, eroding the base that looks so durable in the current surge. The switching costs are real and valuable, and a satisfied customer has genuine reasons to stay; the installed base is Bloom's best moat asset. But an investor should model it honestly as a moat that must be defended and re-won at intervals against an improving field, not one that holds customers automatically and forever — a source of real but cyclical stickiness that supports a thin, potentially-thickening moat rather than the permanent lock-in of a wide one — stickiness measured in contract terms, not decades¹.