⚠ Switching Costs Fade at Contract EndModerate threat

The switching costs that make Bloom's installed base sticky are real but time-bounded, and the boundaries are moments of vulnerability. Within a system's operating life, the cost and disruption of ripping out installed Energy Servers hold the customer in place. But when a service agreement comes up for renewal, or when a system nears the end of its useful life and must be replaced regardless, the lock-in dissolves: the customer is once again free to evaluate Bloom against every alternative — a now-available grid connection, cheaper gas turbines or batteries, an improved rival fuel cell, or a newer technology — with no sunk-cost barrier to leaving, since the old system is being retired anyway. The competition Bloom won years earlier is re-run, on terms that may have shifted against it.

This matters because the alternatives are improving and the circumstances that favor Bloom today may not persist. A customer that bought Bloom in the current power crunch because it needed fast, reliable on-site power might, at renewal years later, find that the grid has caught up, that competing solutions have cheapened, or that its own needs have changed — and choose differently. So the recurring revenue that looks annuity-like is really a series of medium-term commitments that must be periodically re-won, not a permanent claim. The switching costs are genuine and valuable, and a satisfied customer with a good service relationship has real reasons to stay. But an investor should not model the installed base as locked in forever: it is locked in for a cycle, and at each cycle's end the door reopens to an ever-improving field of competitors and substitutes — which is why even Bloom's best moat mechanism is sturdy rather than impregnable, and supports a thin moat that could thicken rather than a wide one already built — the gap a ~20x-sales price ignores¹.