Axe Compute

We see nascent and execution‑dependent advantages rather than durable moats.

Components: switching costs 35/100 (multi‑year take‑or‑pay Build contracts can create friction to switch once clusters go live), cost advantage 25/100 (prepayments and project‑level financing can reduce equity needs but are not structural cost advantages versus hyperscalers), efficient scale 30/100 (local power and site capacity can deter new entrants in specific regions, but competitors are abundant), intangible assets 20/100 (early brand and relationships are unproven), network effects 10/100 (no clear user‑to‑user value accrual).

Weighting these by importance yields roughly 30/100. Post‑quarter 2.8 billion dollars of signed TCV and a 1.5 billion dollar B300 cluster highlight opportunity, but these are forward‑looking and do not constitute a defensible network or platform lock‑in.

Moat erosion risks include GPU supply easing, hyperscaler pricing responses, and customer preference for integrated cloud stacks.