The largest cloud providers don’t buy subsea capacity the way most enterprise customers do — as a reaction to an immediate bandwidth shortage. They plan against a rolling five-to-seven-year regional capacity forecast, and by the time a purchase order reaches our sales team, the underlying demand model has usually already gone through multiple rounds of internal review.
That long planning horizon exists because a new subsea cable system takes two to four years from final route decision to ready-for-service — permitting, survey, manufacturing, and lay all take real calendar time that can’t be meaningfully compressed under pressure. A hyperscaler that waits until existing capacity is nearly exhausted to start the procurement conversation has already lost the option of a new-build system as a solution; at that point they’re bidding for whatever lit or dark capacity already exists on the routes they need, which is a fundamentally weaker negotiating position.
In practice, this means our most sophisticated customers engage us early and iteratively: an initial regional-growth conversation two to three years before they need service, often before they’ve picked a specific route or even a specific coastal region, followed by route and capacity refinement as their internal forecast firms up, and finally a formal RFP once the requirement is concrete enough to specify fiber-pair count, wavelength count, and required ready-for-service date.
We structure our own new-system planning around exactly this cadence — Boreal Span’s diversity route and the expansion capacity we’ve reserved on Amberline both exist because of demand signals from hyperscale customers eighteen to thirty-six months before either system’s formal RFP process began.
