Capacity decisions tend to get made under deadline pressure, a customer commitment coming due, a network segment approaching saturation, and under pressure the default answer is usually whatever solved the problem last time. That's a reasonable instinct in a crisis. It's not a good basis for a repeatable capacity strategy.
There are three real paths to additional capacity: build it yourself, buy it from a carrier directly, or broker it through a partner who sources across multiple carriers. Each has a different cost structure, timeline, and risk profile, and the right answer changes depending on the situation.
Build: right when the capacity need is permanent and large enough to amortize
Building your own infrastructure, whether that's dark fiber, your own backhaul, or expanded colocation footprint, makes sense when the capacity need is durable enough to justify the capital outlay and the timeline to build doesn't conflict with when you need the capacity online. It also makes sense when the route or location is strategically important enough that owning it outright is worth the premium over leasing.
The trap with build decisions is underestimating the time-to-service. Construction timelines, especially for fiber, are measured in months, and a build decision made under near-term pressure often arrives too late to solve the problem that prompted it.
Buy: right when you have a clear, stable need from a specific carrier
Buying capacity directly from a single carrier is the fastest path when that carrier already serves the locations you need and the commercial terms are competitive. It's a reasonable default for straightforward, single-location needs where there's no ambiguity about which carrier is the right fit.
The limitation is that buying direct ties your capacity to one carrier's footprint, pricing, and availability. If that carrier doesn't serve a location well, or their pricing isn't competitive for that specific route, buying direct from them anyway because the relationship is already established is a cost you're choosing to absorb.
Broker: right when the need spans multiple locations, carriers, or unknowns
Brokering capacity, sourcing across a wide carrier base to match each location to its best-fit provider, is the right approach when a single carrier doesn't have the strongest footprint everywhere you need it, or when speed matters and running parallel RFPs across multiple carriers isn't realistic on your own timeline.
This is also the right model when you don't yet know which carrier is best for a given location and don't have the internal bandwidth to find out. A broker with existing relationships across 500-plus carriers can answer that question in days rather than weeks of independent research.
A simple way to decide
Single location, known carrier, durable need: buy direct.
Strategic, high-volume, long-term route: build, if the timeline allows it.
Multi-location, unclear best-fit carrier, or time-constrained: broker.
Why this matters before the deadline, not during it
The framework is most useful when it's applied before a capacity crunch forces a decision, because under pressure, brokering often gets skipped in favor of whatever's fastest to set up, even if it's not the best long-term fit. Building this evaluation into your planning cycle, rather than your emergency response, gets you the right answer more often.