Lit fiber is a service: the provider owns the optics, sets the capacity, and carries the operational responsibility. Dark fiber is an asset lease: you take the unlit strands, attach your own equipment, and own capacity, protocol, encryption, upgrade timing, and fault isolation. Dark fiber is cheaper at very large scale and over long horizons; lit fiber is cheaper for everything else once you count the equipment and the staff.
The decision between dark and lit fiber is rarely about the fiber. It is about how much of the network you want to own — the equipment, the monitoring, the spares, the expertise — and whether owning it buys you something you actually need.
Organizations that already run a network operations function usually find dark fiber liberating. Organizations that do not usually find it expensive in ways that were not on the quote.
| Criterion | Dark fiber | Lit fiber |
|---|---|---|
| What you receive | Unlit optical strands between two points | A working circuit at an agreed capacity |
| Who provides the optics | You do, at both ends | The provider does |
| Capacity ceiling | Whatever your equipment supports | The capacity you purchased |
| Upgrading | Change your optics; no new contract | Order more capacity from the provider |
| Protocol and encryption | Entirely your choice, end to end | What the service supports |
| Monitoring and fault isolation | Yours above the fiber; the provider handles the fiber itself | The provider's, end to end |
| Commercial structure | Multi-year lease or IRU, priced on route, strands, and term | Monthly recurring charge per circuit |
| Best economic fit | Very high capacity, many links, long horizon | Moderate capacity, few links, or no network operations staff |
| Latency | Set by route length plus your own optics | Set by route length plus provider transport equipment |
| Lead time | Longer — strand availability plus your equipment procurement | Shorter — the provider lights a route that already exists |
| Route diversity | A second physically distinct route must be leased and lit separately | Protection can be built into the service where diverse routes exist |
| Spares and lifecycle | Yours — optics, spares, and a refresh cycle every few years | The provider's, absorbed into the monthly rate |
Dark fiber removes a provider from the path of every decision. If you want to run a specific optical platform, encrypt at layer 1, carry a protocol nobody sells as a service, or move from 10G to 400G on a Tuesday, dark fiber lets you. That matters enormously to carriers, research networks, and utilities, and matters not at all to an organization that just wants a working circuit.
Comparing a per-mile lease against a monthly circuit price is the wrong comparison. Include the optics at both ends, spares, the amplification or regeneration required over distance, the monitoring platform, and the engineering time to run it. For a single moderate-capacity link, lit service usually wins comfortably. For many high-capacity links over a long term, dark fiber usually wins decisively. The crossover is real and worth calculating rather than assuming.
Over a short metro span, dark fiber needs little more than a pair of optics. Over long distances it needs amplification, dispersion management, and possibly regeneration, which adds cost and operational complexity quickly. This is why dark fiber economics look very different for a two-mile campus link than for a hundred-mile regional route.
Even on dark fiber, the physical plant remains the provider's responsibility — locates, repairs after a cut, and maintenance of the route. What transfers to you is everything above the glass. Understanding that boundary precisely, and what the repair commitment is, matters more in a dark fiber agreement than in a lit one, because you will be the one who notices the fault.
Dark fiber is not a service with a lighter invoice; it is a different division of labour. You specify, buy, install, and maintain the optics at both ends. You hold the spares, you own the monitoring, and you carry the upgrade cycle when the equipment reaches end of support in five to seven years. You also own first-line diagnosis: when a link goes down, establishing whether the problem is the fiber or your own optics is your job before it becomes anyone else's. For a team that already does this work, none of it is new. For a team that does not, it is a hiring decision wearing a procurement disguise.
Neither option is inherently redundant. Two strands in the same sheath share every backhoe, every bridge and every conduit collapse on that path, and buying dark fiber does not change that. Real diversity means two routes that do not share structure, verified on a map rather than asserted in a proposal, with the divergence documented at the points where paths are most likely to be forced together — bridge crossings, rail corridors, and the entrance to the building itself. Ask for the route on both products. The honest answer sometimes is that a second distinct path does not exist to a given address yet, which is useful to know before designing around one.
Lit service is faster in almost every case, because the provider is turning up equipment on infrastructure that is already in the ground. Dark fiber adds two queues to the same project: whether unlit strands exist on the route you want, and how long your own optics take to procure, stage and configure. Where the route needs construction, both products inherit the same permitting timeline, and permits are usually the longest pole. Sequencing matters more than either product choice — organizations that light a shorter interim path while a longer one is built rarely regret it.
We sell both, and the recommendation follows the staffing question more than the traffic question. Organizations with network engineers who already run optical or transport equipment, and who want the capacity ceiling and upgrade timing under their own control, are well served by dark fiber. Organizations without that team are usually better off with a lit service, because dark fiber hands you an asset and a set of responsibilities at the same time. Buying dark fiber for control you have no intention of exercising is the most common way this decision goes wrong.
Six questions that settle this choice faster than a feature table. The answer to each one tells you something the specifications do not.
If no one on staff has configured a transponder or held a spare optic, dark fiber is a hiring decision as much as a procurement one.
Dark fiber pays back when capacity grows unpredictably or steeply, because upgrades become equipment decisions rather than contract renegotiations.
Layer-1 encryption, unusual framing, or timing protocols push toward dark fiber. If standard Ethernet handoffs suffice, lit is simpler.
An IRU or long lease amortises over years. If the requirement might change inside three, the flexibility of a lit service is worth more than the control you are buying.
On dark fiber, you will. Confirm you have the monitoring to distinguish a cut from your own optic failing before the first outage teaches you the difference.
If resilience is the objective, settle this before choosing a product — a single dark route is not more resilient than a single lit one.
That no light is being transmitted on the strands — the fiber is installed but unlit. Leasing it means you supply the transmission equipment that lights it, which is why capacity becomes your decision rather than a purchased tier.
It is more isolated, which is not the same thing. Your traffic does not share a provider platform, and you can encrypt at the optical layer with your own equipment. The fiber itself is not inherently secure — physical access to it still matters, which is why encryption remains your responsibility either way.
An indefeasible right of use: a long-term, effectively ownership-like right to use specific strands, usually paid substantially up front. It is the common structure for dark fiber because the value of the arrangement is long-horizon control.
Often, subject to strand availability on the route at that time. It is a reasonable path for an organization growing into its own network operations capability, and worth raising at the outset so the commercial terms anticipate it.
Describe what the connection has to carry and where. An engineer will tell you which service actually fits — including when the cheaper option is the right one.
Service availability depends on location, network proximity, capacity, and engineering review. Share an address and we will confirm what can be delivered there.