Why a fiber route near a property does not necessarily make it ready for AI infrastructure.

Fiber maps are very convincing.

A colored line runs past a development site. Another line appears a few miles away. The marketing package adds “fiber-ready,” and suddenly the property looks prepared for a data center.

There is only one problem.

The line may tell you almost nothing about the connectivity a customer can actually purchase.

The fiber may belong to a carrier that does not serve the property. It may not contain available strands. The route may be dark, privately controlled or already committed to another customer. The nearest access point could be miles away.

Those two lines suggesting redundancy may even share the same conduit, bridge or railroad crossing.

A fiber map can show where a network travels.

It does not necessarily show where a customer can enter it.

That difference is becoming increasingly important as communities compete for AI infrastructure.

The Short Answer

A fiber route near a property does not guarantee that an AI data center can connect to it.

True connectivity readiness depends on whether the route has available capacity, whether the owner is willing and able to serve the site, how the property will physically connect to the network and whether genuinely diverse backup routes are available.

Developers must also evaluate latency, cloud access, interconnection locations, construction costs and expansion timelines.

“Fiber nearby” is a geographic observation.

“Connectivity available” is a commercial and engineering fact.

The Signal

The AI buildout is creating a new wave of fiber construction.

In August 2026, Zayo announced it was working with Nvidia to expand network capacity for AI infrastructure. The company said it was building more than 8,000 miles of new long-haul fiber across rapidly growing AI corridors while increasing capacity on existing routes.

Zayo reported that its current build and overbuild projects span more than 15,000 route miles across North America. Its acquisition of Crown Castle’s fiber business added another 90,000 metro route miles and 40,000 on-net enterprise locations. Zayo

This is not ordinary broadband expansion.

It is the construction of a new AI transportation system.

AI facilities increasingly operate as distributed computing environments. Processors may run in one location, data may be stored in another and applications may be delivered through several cloud regions.

That requires large volumes of information to move between facilities.

A Ciena survey of more than 1,300 data-center decision-makers found that organizations expected data-center interconnection bandwidth requirements to grow at least sixfold over five years. Ciena

The demand is real.

But a growing national fiber network does not mean every individual property is ready to connect.

What Does a Fiber Map Actually Show?

Most publicly available fiber maps show the approximate path of a carrier network.

They May Identify:

• Long-haul fiber corridors

• Metro fiber rings

• Carrier points of presence

• Data centers connected to the network

• Subsea cable landing locations

• Major network exchanges

• General service areas

That information is useful for understanding regional connectivity.

It is not a site-level engineering assessment.

Public network maps may simplify routes, omit street-level details or combine several network products into one visual. Some maps show where a carrier owns infrastructure. Others show locations the carrier can serve through partners.

The line may represent a fiber corridor with hundreds of strands.

Or it may represent access to one managed service.

Those are very different assets.

A serious evaluation begins after the map is opened.

Who Owns the Fiber Route?

Fiber ownership determines who controls access.

The Route Might Belong To:

• A telecommunications carrier

• A cable company

• A utility

• A municipality

• A railroad

• A state transportation agency

• A data-center operator

• A private enterprise

• A regional fiber consortium

Not every owner operates the route as an open commercial network.

A utility may use fiber primarily for grid communications. A railroad may have leased capacity under long-term agreements. A private company may control strands for its own operations. A municipal network may face policy restrictions on who it can serve.

Even when capacity exists, the owner may not sell or lease it to the proposed project.

That is why the first site-level connectivity question should not be, “Is there fiber?”

It should be:

Who owns it, and can we buy access?

Lit Fiber vs. Dark Fiber

Not all fiber service is delivered the same way.

What Is Lit Fiber?

Lit fiber is a fiber-optic connection that already includes the electronic equipment required to transmit information.

The service provider operates the network, maintains the equipment and sells the customer a defined amount of bandwidth. The product may be offered as Ethernet, internet access or a dedicated wavelength.

Lit service can be easier and faster to activate because the provider manages the infrastructure.

The customer purchases capacity.

It does not control the entire optical system.

What Is Dark Fiber?

Dark fiber is installed fiber-optic cable that is not currently carrying an active optical signal.

A customer may lease individual strands and install its own network equipment. This provides greater control over capacity, security, technology upgrades and network architecture.

Lumen defines dark fiber as installed optical cable that is not being used, with the leasing customer responsible for the switching, routing and optical equipment required to operate it. Lumen

Dark fiber can be particularly valuable for hyperscalers, cloud providers and large data-center operators moving enormous volumes of information between facilities.

But unused cable does not automatically mean available cable.

The strands may already be reserved. The owner may not offer dark-fiber leases in that market. The route may require new regeneration facilities, colocation space or optical equipment before it becomes usable.

“Dark” describes the condition of the fiber.

It does not describe whether you can lease it.

Capacity Is Not the Same as Access

A fiber cable can have significant technical capacity while offering no commercially available capacity to a new customer.

There are several reasons.

The Strands May Be Committed

A carrier may have leased available strands through long-term agreements. The physical cable remains in the ground, but its usable inventory has been allocated.

The Electronics May Be Full

A lit network may have fiber available but lack the optical equipment, ports or power required to add another high-capacity customer.

Expanding service could require new equipment or an upgrade to the carrier’s network.

The Route May Not Have an Access Point

Fiber cannot necessarily be connected at any convenient location.

A carrier may require access through a designated splice point, network hut, colocation facility or point of presence. A route could pass directly beside a property while the nearest approved connection location sits several miles away.

The Carrier May Not Serve the Property

The site might be considered near-net rather than on-net.

That difference matters.

• On-net generally means the carrier already has infrastructure connected directly to the property or building.

• Near-net means the carrier has infrastructure nearby but must extend it to reach the site.

• Off-net means the provider must use another company’s network to deliver service.

A near-net property may be a strong candidate.

But it is not connected yet.

The lateral extension could require new construction, rights-of-way, road crossings, railroad approvals and months of permitting.

The carrier may be close enough to see.

That does not mean the customer can plug in.

The Last Mile Can Be the Hardest Mile

Long-haul fiber receives most of the attention because it connects regions and major data-center markets.

The final connection to the property can determine whether the site is usable.

A Fiber Lateral May Need to Cross:

• A public road

• A railroad

• A drainage channel

• Privately owned land

• Utility easements

• Environmentally sensitive areas

• Existing underground infrastructure

Every crossing can introduce cost, permitting and construction risk.

The connection also needs a physical entrance into the campus. Large facilities may require multiple entrance points, separate conduits and network rooms positioned on different sides of the development.

A route sitting 500 feet from the property can still become a complicated infrastructure project.

Distance is only one part of the calculation.

Access is the other.

Two Carriers May Still Mean One Route

A project may appear resilient because two carriers have agreed to provide service.

That does not prove physical diversity.

One carrier may lease capacity from the other. Both services may travel through the same conduit. The routes may enter the property from opposite directions but cross the same bridge several miles away.

A construction accident, flood, fire or equipment failure at that shared point could interrupt both connections simultaneously.

This is known as shared-risk infrastructure.

True network diversity requires routes that remain physically separate across the critical path.

That May Include:

• Different street corridors

• Separate conduits and trenches

• Multiple campus entrance points

• Different carrier networks

• Separate network facilities

• Independent bridge and railroad crossings

• Distinct regional paths

• Different cloud and interconnection locations

Buying from two companies is commercial diversity.

Following two genuinely separate physical paths is infrastructure diversity.

AI campuses need both.

Route Miles and Fiber Miles Are Not the Same

Network announcements often use two different measurements.

A route mile measures the physical distance traveled by the fiber corridor.

A fiber mile generally multiplies the route distance by the number of strands installed along it. A 100-mile route containing 144 strands represents far more than 100 fiber miles.

This is why a company may announce thousands of new route miles and millions of fiber miles.

The distinction matters because the physical corridor and the potential capacity inside it answer different questions.

Route miles tell us where the network reaches.

Fiber count helps indicate how much future capacity the corridor may support.

Neither measurement alone confirms how much capacity is currently available to a specific customer.

Why Do Cloud On-Ramps Matter?

A data center can have excellent regional fiber and still sit far from the infrastructure connecting directly to major cloud providers.

A cloud on-ramp is a private connection to a cloud platform, typically available inside a carrier-neutral colocation or interconnection facility.

Equinix describes a cloud on-ramp as a direct private connection offered by a cloud provider within a vendor-neutral colocation data center. Equinix

Cloud on-ramps allow companies to connect to services such as Amazon Web Services, Microsoft Azure, Google Cloud and Oracle Cloud without moving all traffic through the public internet.

This Can Improve:

• Performance

• Security

• Reliability

• Predictability

• Data-transfer economics

• Access to multiple cloud regions

A property located near fiber but far from cloud on-ramps may need additional network hops and longer routes.

A site positioned near a major interconnection market can reach more clouds, carriers and customers with fewer intermediate steps.

That makes interconnection geography part of land value.

Carrier Hotels Are the Network Marketplace

A carrier hotel is a data center where many network providers, cloud platforms and enterprises interconnect.

These facilities function as marketplaces for connectivity.

A company inside a well-connected carrier hotel may be able to reach dozens or hundreds of networks through short physical cross-connects. A company outside that ecosystem may need long-haul transport to reach the same providers.

Carrier hotels and internet exchanges create network gravity.

As more carriers enter a facility, it becomes more valuable for other carriers and customers to connect there. That activity attracts cloud on-ramps, content providers, enterprises and additional infrastructure investment.

This is why a fiber route should not be evaluated only by where it passes.

It should be evaluated by what it connects.

A line leading nowhere useful is less valuable than a route connected to the right network ecosystem.

Latency Is Still a Geography Problem

Fiber moves information extremely quickly.

It does not move information instantly.

Every mile adds propagation time. Every router, switch, regeneration point and intermediate network can introduce additional delay.

The most direct physical route is not always the route traffic actually follows.

A network may send information through a distant carrier facility before returning it to a nearby market. This is sometimes called tromboning because the data travels outward and back instead of following the shortest path.

For general cloud storage, the delay may be difficult to notice.

For financial trading, industrial automation, autonomous systems, gaming, healthcare applications and real-time AI inference, milliseconds can influence performance.

That makes latency a site-selection issue.

Leaders Should Evaluate:

• Physical distance to priority markets

• Actual network routing

• Number of intermediate hops

• Location of cloud on-ramps

• Proximity to internet exchanges

• Access to edge-computing facilities

• Reliability during network congestion

• Performance of backup routes

The cloud may appear locationless.

The network carrying it is not.

The Connectivity Readiness Stack

A property becomes connectivity-ready when several layers align.

1. Physical Proximity

Fiber routes exist close enough to serve the property.

2. Legal and Commercial Access

The route owner is authorized and willing to provide service.

3. Available Capacity

Usable strands, wavelengths, ports and supporting equipment are available.

4. Site Connection

The lateral construction, rights-of-way and campus entrances are technically and legally achievable.

5. Network Diversity

Backup routes avoid the same trenches, crossings, facilities and regional failure points.

6. Market Reach

The network connects efficiently to cloud regions, carrier hotels, internet exchanges, data centers and priority customers.

7. Expansion Capacity

The system can grow as the campus adds computing capacity and generates additional traffic.

A weakness in any layer can reduce the value of the entire network plan.

Fiber proximity starts the conversation.

It does not finish the assessment.

Following Fiber

Data-center announcements usually arrive after months or years of infrastructure planning.

Fiber activity can provide an earlier signal.

New routes reveal where network providers expect future traffic. Dark-fiber leases indicate where hyperscalers may be preparing capacity. Carrier points of presence can show which markets are becoming more important. Cloud on-ramps and interconnection expansions reveal where digital ecosystems are gaining density.

This is the idea behind Following Fiber.

In April 2026, Zayo announced that an anchor customer was supporting 8,000 route miles of new construction and network overbuilds across high-demand AI corridors. The company said power availability was helping determine where data-center demand was concentrating and which network corridors required additional capacity. Zayo

The sequence matters.

Power availability attracts computing capacity.

Computing capacity attracts fiber.

Fiber connects the campus to other markets.

That connectivity then makes the region more attractive to additional investment.

Infrastructure begins creating its own economic gravity.

Company Watch: Zayo

Everyone is watching where the data centers are announced.

I’m watching where Zayo is digging.

Zayo owns and operates long-haul and metro fiber connecting data centers, cloud markets and enterprise locations across North America.

That gives the company visibility into where network demand is materializing—and where existing infrastructure may be insufficient.

The company originally announced plans in January 2025 to build more than 5,000 new long-haul route miles. By April 2026, it had expanded the program substantially, with projects underway across more than 15,000 route miles.

Its August 2026 Nvidia announcement adds another important signal.

Zayo is not simply increasing capacity between traditional technology hubs. It is building routes toward emerging AI corridors where power availability is attracting new computing infrastructure.

This makes the company useful to watch for two reasons.

First, its network expansion may reveal where AI demand is becoming commercially credible.

Second, its projects demonstrate that AI infrastructure cannot scale through existing fiber alone.

The next generation of AI markets will require new routes.

Zayo is helping show us where those routes are going.

Fiber Changes the Value of Land

Land near scalable, diverse and commercially accessible fiber is more useful than visually similar land without it.

The difference may not appear in an ordinary property appraisal.

It can become decisive during data-center site selection.

A Parcel May Gain Strategic Value When It Sits Near:

• Multiple long-haul routes

• Dense metro fiber

• Carrier-neutral facilities

• Cloud on-ramps

• Internet exchanges

• Data-center clusters

• Separate rights-of-way

• Expandable network corridors

The reverse is also true.

A property may lose value when its apparent connectivity depends on one carrier, one trench or one difficult railroad crossing.

This is why infrastructure diligence must occur before a community markets a property as AI-ready.

The claim should be supported by carrier conversations, route verification and preliminary construction plans.

“Fiber-ready” should mean more than a colored line on a brochure.

Connectivity Across the Eight Pillars

Connectivity creates value only when it aligns with the rest of the Infrastructure of Intelligence™.

Power allows the processors to operate. Fiber allows them to exchange information. Land determines where the systems can be built. Policy affects rights-of-way and construction timelines. Capital finances new routes and network equipment.

Workforce determines whether technicians and construction crews are available to install and maintain the network. Sustainability and water influence the larger campuses connected to it.

A fiber-rich location without power may remain undeveloped.

A power-rich location without fiber may remain isolated.

This is the Twin Infrastructure Principle™:

Power + Connectivity = AI Readiness

The strongest regions will be the ones where energy and information can move together.

The Hot Take

The most important fiber route may not be the one already on the map.

It may be the one a carrier is willing to build next.

Existing infrastructure shows where yesterday’s demand developed. New construction can reveal where companies believe tomorrow’s traffic will emerge.

That means communities should not treat fiber availability as a fixed condition.

They can influence it.

A coordinated region with power, developable land, clear rights-of-way and several credible data-center projects may give carriers a commercial reason to build. An anchor customer can transform a proposed corridor into financeable network infrastructure.

The communities that organize demand may attract connectivity before their competitors.

What Leaders Should Start Asking

Developers, investors and economic-development organizations should ask:

• Which carriers physically own fiber near the property?

• Is the site on-net, near-net or off-net?

• Where does the route actually run?

• Is the fiber lit or dark?

• Are dark-fiber strands available to lease?

• How much lit capacity can be purchased today?

• What equipment upgrades would additional capacity require?

• Where is the nearest approved splice or access point?

• What lateral construction is required?

• Which rights-of-way must be secured?

• Are road or railroad crossings involved?

• How long would permitting and construction take?

• What would the connection cost?

• Can multiple carriers serve the site directly?

• Are the proposed routes physically diverse?

• Do backup routes share conduits, bridges or network facilities?

• Where are the nearest cloud on-ramps?

• Which cloud providers can be reached privately?

• Where are the nearest carrier hotels and internet exchanges?

• What latency can the network provide to priority markets?

• How will capacity expand as the campus grows?

• What fiber investments are carriers planning over the next five years?

These questions separate general network coverage from site-level connectivity readiness.

The Golden Nugget

A fiber route is not valuable because it passes the property. It is valuable because the customer can access it, scale it and survive when part of it fails.

That is the difference between fiber proximity and usable connectivity.

Frequently Asked Questions

Does Fiber Near a Property Make It Data-Center Ready?

No. The route must have available capacity, an accessible connection point and a provider willing to serve the site. Construction, rights-of-way and network diversity must also be evaluated.

What Is the Difference Between Lit Fiber and Dark Fiber?

Lit fiber includes the optical equipment and managed service needed to transmit data. Dark fiber consists of unused strands that a customer leases and operates using its own networking equipment.

What Does On-Net Mean?

An on-net property is directly connected to a carrier’s network. A near-net property is located close to the network but requires additional construction. An off-net property generally relies on another provider’s infrastructure.

Why Does Route Diversity Matter?

Route diversity protects a facility from outages caused by damaged cables, equipment failures, severe weather or construction accidents. Backup connections must avoid shared physical failure points to provide genuine resilience.

What Is a Cloud On-Ramp?

A cloud on-ramp is a direct private connection to a major cloud provider, usually accessed through a carrier-neutral colocation or interconnection facility.

Why Does Latency Matter for AI?

Latency measures the delay as information moves across a network. Low latency can be important for real-time AI applications, industrial systems, financial platforms, healthcare and autonomous technologies.

How Can a Community Verify Fiber Readiness?

The community should engage carriers directly, confirm route ownership and capacity, identify connection points, evaluate construction requirements and document physically diverse paths to important markets.

The IOI Take

The market treats fiber as a checkbox.

The Intelligence Economy will treat it as a strategic system.

A route on a map can identify potential. It cannot establish readiness.

True connectivity depends on ownership, capacity, access, route diversity, interconnection and expansion. Those details determine whether an AI campus can communicate with other facilities, reach cloud platforms and deliver intelligence to customers.

This creates an opportunity for communities.

Most regions market fiber using the same language: high-speed, redundant and available. Few can produce the engineering and commercial evidence behind those claims.

The regions that complete that work will stand out.

They will know where the routes run, who controls them, how much capacity is available and what must happen to connect priority development sites.

That is how connectivity becomes an economic-development asset.

Do not stop at the map.

Follow the fiber.

One Last Thought

A fiber line can pass directly beside a property and still be commercially, technically or legally out of reach.

That is what makes connectivity different from ordinary real estate.

Proximity creates possibility.

Access creates value.

The map may show where the fiber goes.

The real question is whether intelligence can travel with it.

Workforce

Policy

Capital

Sustainability