Cogent vs Lumen for IP Transit and Dedicated Internet

Why this comparison matters for WISPs

Upstream transit is the cost line that scales with your success. Every subscriber you add pushes traffic out through your upstream ports, and the difference between a well-bought commit and a badly bought one compounds monthly for the length of the term. It is also the decision with the strangest price dispersion in telecom: publicly observable transit pricing varies by multiples depending on location, volume, term and how the negotiation is run — which is exactly why a structured head-to-head between a price-aggressive specialist and an incumbent carrier is worth the effort.

The comparison matters because the two companies represent the two poles of the transit market. Cogent's publicly visible posture is volume and price: a specialist built to move large amounts of IP traffic cheaply, strongest when you are on its network. Lumen's posture is reach and incumbency: a national carrier whose backbone is one of the networks that constitutes the internet's core, with a large enterprise and wholesale organization around it. Buying well means understanding what each posture is actually selling — and what it is not.

It matters for a third reason that is easy to miss: this is rarely an either/or decision even when you sign only one contract. The strongest WISP upstream designs are multi-homed — two or more providers, BGP, physically diverse paths — and "Cogent or Lumen" is often really "Cogent and Lumen," or either of them plus a third provider. The architecture section below covers that design, because the provider question and the redundancy question are the same question asked twice.

Finally, a warning that the rest of this page will repeat: never single-home a WISP's entire upstream on one transit provider regardless of price. A cheap commit that takes your whole network down with it is the most expensive circuit you will ever buy. If a single-provider quote looks irresistible, the correct response is to buy it as one leg of a multi-homed design — not as your only leg.

Who these two companies are

Everything in this section is drawn from public statements and public reporting as of this writing. Treat it as orientation, not diligence — verify current status directly with each provider before signing anything.

Cogent Communications

Cogent Communications is a U.S.-based, publicly traded carrier publicly known as a high-volume, price-aggressive specialist in IP transit and dedicated internet access. Its model is on-net-centric: it publicly markets its most aggressive economics in the large number of commercial buildings and data centers where its network is already lit, and its product set concentrates on IP connectivity rather than a broad managed-services portfolio. As publicly reported, Cogent acquired Sprint's wireline network from T-Mobile in 2023, adding fiber assets, customers and network scope to its footprint — an integration whose details you should verify where they touch your routes.

Lumen Technologies

Lumen Technologies is an incumbent U.S. national carrier whose fiber network descends from CenturyLink and Level 3 Communications — the latter being the origin of its Tier-1-scale backbone, one of the large settlement-free networks at the internet's core. It publicly markets IP transit, dedicated internet, Ethernet and wavelength transport, plus a broad enterprise portfolio, across a large national and international footprint. Lumen has publicly announced divestitures of parts of its business in recent years; the practical consequence for you is simple: verify which legal entity you are contracting with, what network it will operate, and what your contract's assignment and change-of-control clauses say.

The scale difference cuts differently than in most carrier matchups. Both companies operate very large backbones; the difference is what sits around the backbone. Cogent is a focused connectivity seller whose identity is price and volume; Lumen is a diversified incumbent whose identity is reach and breadth. Neither is inherently better for a WISP buyer — but they produce very different quoting behavior, routing philosophies and account experiences, which the rest of this page unpacks.

Tier-1 incumbent vs transit specialist: the structural difference

The most consequential difference between these two providers is not price — it is routing philosophy, and it is structural. The internet's core is a set of large networks that exchange traffic settlement-free under peering arrangements. Lumen's backbone, inherited from Level 3, sits in that settlement-free core: per its public positioning, it reaches the rest of the internet without paying another network for transit. Cogent also publicly describes itself as settlement-free with its peers, but its peers and its peering posture are different — and that difference is the thing to understand.

Cogent's peering disputes are well documented as an industry phenomenon: over the years, various large networks have at times restricted or de-peered capacity toward Cogent, and the affected parties have been the customers on both sides of the restriction. We are not asserting any specific dispute is active as you read this — peering relationships change, and you must verify current reachability yourself. The structural lesson is what matters: when you buy transit, you are buying that provider's entire relationship graph with the rest of the internet, including its disputes. A transit provider's peering posture is a product feature, and it belongs in your evaluation next to price.

How to evaluate reachability in practice: ask both providers for a looking-glass or route-server view and test paths from their networks to the destinations your subscribers actually use — the major content networks, cloud providers and streaming sources that dominate your traffic. Better yet, run the same tests from your existing network if you have one. Then repeat the exercise periodically after you buy, because the answer changes. A provider that reaches 95% of your traffic cheaply and 5% poorly is a fine primary leg paired with a second provider that covers the gap — which is, again, the multi-homing argument.

The other structural difference is commercial posture. A transit specialist makes money when ports fill and commits grow, so expect aggressive per-megabit pricing and straightforward products. An incumbent carrier makes money across a broad portfolio, so expect more products on the table — transport, waves, managed options — and pricing that responds to competition more than to list logic. Both postures are exploitable by a prepared buyer; both are traps for an unprepared one.

Footprint reality: on-net, off-net and the cost of reaching the network

Footprint is the first filter, and in this matchup it is unusually decisive. What follows is a hedged summary of publicly marketed footprint postures as of this writing — verify every building and data center yourself, because on-net lists change and marketing pages are optimistic by design.

Cogent's economics depend heavily on being on-net, as its public positioning makes clear: in a lit building or data center, delivering your port is cheap for Cogent, and the pricing reflects it. Off-net, the same circuit may require construction or a third-party tail circuit, and the price and lead time change character entirely. The practical rule: Cogent is a candidate where Cogent is lit, and the first question to ask is not "what do you charge" but "is my address — or my nearest data-center suite — on your network, in writing."

Lumen's footprint posture is the incumbent's: a large national fiber network with lit buildings and data-center presence across the country, plus the wholesale reach to deliver in more places via local loops and partner tails where it is not lit. Broader deliverability does not mean cheaper — off-net delivery through an incumbent often carries the tail circuit's cost and its contract terms inside your quote. The same rule applies in reverse: verify on-net status per address, in writing, and ask separately what portion of the route is Lumen's own fiber versus a third-party tail.

For most WISPs the footprint question resolves at a data center, not at the tower. The standard architecture is transport from your aggregation site to a carrier-neutral facility, then transit from one or more providers inside that facility, where cross-connects make every on-net provider reachable for a modest monthly fee. If you are not yet in a facility like that, price the move before pricing the transit — a cheap off-net transit quote often loses to a slightly pricier on-net quote once you add the transport to reach it. Our IP block calculator handles the addressing side of the same planning exercise.

Product lineup comparison: IP transit, DIA and transport

Both companies publicly market products across the layers a WISP upstream buyer needs — IP transit, dedicated internet access, and transport — but the emphases differ, and comparing across layers without realizing it is the classic procurement mistake. The table below summarizes publicly marketed product categories as of this writing. Availability of any specific product at any specific location must be confirmed in the quoting process — treat this as a map of what to ask for, not a catalog of what you will get.

The reading note that matters most: IP transit and DIA are the same internet delivered under different commercial shapes. Transit assumes you are a network — your own ASN, your own address space, BGP sessions, usually delivered at a data center. DIA assumes you are a customer — the provider can assign address space, and delivery can be a building handoff. WISPs with an ASN should evaluate both providers' transit products; if you are still buying DIA at your aggregation point, the ASN and BGP guide is the prerequisite reading.

On capacity planning: buy the commit for where your traffic will be mid-term, not at turn-up, and understand the burst mechanics before you sign. If your peak is approaching the top of your port size, our guide on when to upgrade from 1G to 10G walks through the upgrade triggers — the same logic applies at 10G to 100G, with bigger numbers.

Provider Fit Score

This worksheet turns a vague preference into arithmetic. Set a weight (0–10) for each criterion based on what matters for your network, then score each provider 1–10 from your actual quotes, route tests and conversations. The weighted score is the sum of weight times score divided by the sum of weights — so the criteria you care about most drive the result.

The scores pre-filled below are editorial defaults — our rough reading of each provider's structural posture, not measurements and not recommendations. Cogent scores higher on per-megabit price posture and on-net economics; Lumen scores higher on footprint breadth and product range. Replace them with your actual quotes and route tests before drawing any conclusion.

Single-homed vs multi-homed: the architecture decision underneath this one

Here it is again, because it is the most important paragraph on this page: do not single-home your entire upstream on either of these providers, whatever the price. Single-homing means one provider's backbone between your entire subscriber base and the internet. Every one of that provider's bad days — a core outage, a routing incident, a peering degradation, a fiber cut into your facility — becomes your entire network's bad day, with SLA credits as the only compensation. Credits do not retain subscribers.

Multi-homing — two or more transit providers, BGP sessions to each, physically diverse paths out of your facility — converts provider failures from outages into routing events. Done properly, a provider going down costs you some degraded paths and a re-convergence, not your subscriber base's evening. The design is standard, well understood, and documented in our guide to route failover between two providers; the prerequisites — ASN, address space, BGP policy — are in the ASN and BGP requirements guide.

Multi-homing also changes this comparison's meaning. In a two-provider design, the question is not "which single provider is better" but "which two providers cover each other's weaknesses." A price-aggressive specialist as the volume leg plus an incumbent as the reachability leg is a classic pairing; so is either of these two plus a regional provider with strong local peering. Evaluate providers in pairs: test each candidate's routes to your traffic's actual destinations and look for complementary coverage, not duplicated strength.

One design trap to avoid: two providers in the same facility on the same conduit are one failure domain wearing two logos. Diversity means separate physical entries, separate transport paths to your aggregation site, and providers whose core failures you do not expect to correlate. Ask both providers to certify path diversity in writing, and walk the facility risers yourself.

Contract and billing worksheet: commit, burst and the 95th percentile

Transit billing mechanics decide what you actually pay, and they are fully negotiable at contract time. The dominant model is a committed rate — you pay for a committed information rate every month whether you use it or not — with usage above the commit billed as burst, most commonly measured by 95th-percentile sampling: your traffic is sampled every few minutes all month, the top five percent of samples are discarded, and the highest remaining sample is your billable usage. Understand that sentence before you sign anything, because it determines your bill more than the port size does.

Two rows deserve emphasis. The overage rate is where transit contracts hide their teeth: a low commit price paired with a punishing overage rate is a bet that you will not grow — and you are a WISP, so you will. Negotiate the overage rate before the commit rate. And auto-renewal deserves distrust in a falling-price market: the transit market's long-term price trend means the contract you signed two years ago is probably above market today, and a renewal you slept through forfeits that decline. Calendar every notice window the day you sign.

Get the final language in the executed agreement, not in an email from a sales rep — and if either provider resists putting billing mechanics in writing, treat the resistance as data about how disputes will go later.

Commit and Burst Cost Estimator

This estimator makes billing mechanics concrete. Enter your commit, your measured 95th-percentile usage, and the per-Mbps prices from a real quote; it shows the monthly bill split between commit and overage. The defaults are illustrative placeholders — not market prices, not a quote, and not a claim about what either provider charges. Replace them with actual quoted numbers.

Use this estimator on every quote, with your projected usage at month 12 and month 24, not just turn-up. Then run the scenario that actually matters: your growth month, where usage lands well above commit. The provider with the lower commit price and the punitive overage rate wins the quiet months and loses the busy ones — and the busy months are the ones you remember. The effective per-Mbps line is the number to compare across providers once transport and cross-connects are added in the next section.

Cost normalization: the all-in comparison worksheet

Two transit quotes are never directly comparable as received. One is on-net with a cross-connect, the other is off-net with a transport tail folded in; one bills 95th percentile outbound, the other bills the max of both directions. Normalize every quote to an all-in monthly and term-total number using the worksheet below.

The arithmetic: all-in monthly equals commit cost, plus projected overage at your forecast usage, plus cross-connect MRC, plus the monthly share of transport to reach the provider, plus anything else recurring. Term-total multiplies that out with escalators and adds NRC. Run both providers through the same projected usage curve — the same curve, not each provider's suggested one — and the comparison usually resolves itself.

The transport row is where this matchup gets interesting. Cogent's price aggression is concentrated on-net; if reaching Cogent requires a transport circuit from your aggregation site to a Cogent-lit facility, that circuit is part of Cogent's real price to you. Lumen's broader footprint may put its PoP closer to you — or it may not, at your address. Normalize both to the same delivery point: your router.

Decision matrix: who should start with which

If the framework above has a bias, it is toward matching provider structure to buyer situation rather than crowning a universal winner. The matrix below is a starting hypothesis — deliberately generic, because your quotes, route tests and weights should make the final call.

Treat any row that matches your situation as a reason to start the conversation there — and then run the head-to-head anyway. The provider who loses the hypothesis often wins the quote, because procurement pressure concentrates minds.

Running a head-to-head transit procurement

A real transit head-to-head is not "collect two quotes and compare PDFs." It is a structured process where both providers bid the same specification, knowing there is competition, and you evaluate on normalized cost plus measured reachability. Here is the compressed playbook.

Step one: write one spec. A single document for both providers: delivery location (facility and suite, or building address), port speed, initial commit and projected commit at months 12 and 24, BGP session requirements, your ASN and address space, SLA requirements drawn from the worksheet above, term preferences and your normalization format. State that this is a competitive bid with a decision date. Identical input is what makes output comparable.

Step two: test routes before you sign. Get looking-glass or route-server access from both providers and test paths to the destinations that dominate your traffic — the content and cloud networks your subscribers actually use. Note anything reached through congested or indirect paths, and ask each provider directly about its current peering posture toward those networks. Document the answers; they are part of the product you are buying.

Step three: normalize and score. Drop both offers into the cost normalization worksheet and the Provider Fit Score, replacing editorial defaults with quote- and test-derived numbers. Include the transport and cross-connect costs to reach each provider. Where a provider is vague on billing mechanics or SLA language, score the vagueness as risk, not as neutral.

Step four: negotiate once, then contract for resilience. Take the stronger normalized offer to the runner-up once — both providers move on price and terms when a deal is credibly competitive, and transit pricing has more room than the first quote suggests. Then sign the design, not just the circuit: the winning provider as one leg, a second provider on a physically diverse path as the other, BGP failover tested before you carry production traffic, and renewal notice windows on a calendar you will actually check.

20 questions to ask both providers

Print this list and bring it to both sales calls. The quality and specificity of the answers — not just the answers themselves — will tell you most of what the Fit Score needs.

Frequently asked questions

Is Cogent or Lumen cheaper for IP transit?

Neither publishes useful list pricing for your situation, and transit prices vary by multiples with location, volume, term and negotiation. Cogent is publicly known for aggressive pricing, especially on-net; Lumen's pricing responds to competition. The only honest answer is procedural: get both quotes, normalize them with the worksheets above at your projected usage, and the answer for your locations will be obvious.

Should I worry about Cogent's peering disputes?

Understand the concept, then test it rather than assuming it. Settlement-free peering disputes between large networks are a documented industry phenomenon, and Cogent has historically been a party to well-publicized ones — but specific disputes come and go, and we assert nothing about what is active as you read this. Test routes to your traffic's actual destinations before signing, re-test periodically, and multi-home so any single provider's reachability problem is a routing event, not an outage.

Do I really need two transit providers?

If your subscribers' connectivity is revenue, yes — full stop. A single provider's outage, routing incident or peering degradation becomes your entire network's outage, and SLA credits do not retain subscribers. Multi-homing with BGP failover is the standard design for exactly this reason; our route failover guide covers the mechanics.

What about Lumen's divestitures — does it change the network I am buying?

It can change the entity and the assets, which is why you verify before signing rather than after. Lumen has publicly announced divestitures of parts of its business in recent years. Ask which legal entity will contract with you, whether the routes and facilities your service uses are part of any announced transaction, and what your assignment and change-of-control clauses say. The same question, for the same reason, applies to Cogent's publicly reported Sprint wireline acquisition: know which assets your routes depend on.

Methodology and disclosure

This page is an informational decision framework, not an endorsement, ranking or performance claim. Cogent, Lumen, CenturyLink, Level 3, Sprint, T-Mobile and all other company and product names mentioned are trademarks of their respective owners; their use here is nominative and does not imply affiliation with or endorsement by those companies.

All factual statements about either company are drawn from public sources — the companies' public reporting, public statements and public marketing — and are hedged accordingly, with an "as of this writing" time reference. Network footprints, peering relationships and corporate structures change; verify current facts directly with each provider. We deliberately publish no pricing, route-performance measurements or reachability statistics, because we have no independent basis for them. The widget scores and pre-filled estimator numbers are editorial defaults and illustrative placeholders — not market data — and the widgets exist precisely so you can replace them with your own figures.

SmashByte is a connectivity advisory and may have commercial relationships with providers in this market, including providers discussed on this page or their competitors. Those relationships do not change the methodology above: every recommendation on this page is a framework you apply to your own verified data. Before signing any agreement, verify current offerings, footprints and contract terms directly with each provider, and have your counsel review the executed documents.