Connectivity is the most consistently underassessed element of UK data centre development and acquisition. Power, planning, and construction receive the most attention and the most resource. Connectivity is frequently treated as a downstream procurement decision — something to be sorted out once the facility is built.
That sequencing produces two problems. The first is that connectivity infrastructure — particularly dark fibre routes, wayleave agreements, and internet exchange access — can take as long to arrange as a grid connection in the wrong location. The second is that the connectivity choices made at development stage determine the commercial options available to the operator for the life of the asset. A facility that was built with a single carrier and a small meet-me room cannot become a carrier-neutral interconnection hub without significant capital expenditure and disruption to live operations.
This article covers what operators and developers need to understand about connectivity at development and operational stage: carrier diversity, dark fibre, internet exchange access, cloud on-ramps, meet-me room specification, and the contractual considerations that shape long-term connectivity flexibility.
Why connectivity decisions made at development stage matter
The connectivity infrastructure of a data centre — the fibre routes entering the building, the carriers present in the meet-me room, the cross-connect capacity and pricing, the internet exchange access points — determines what the facility can offer its customers for the next 20 to 25 years. Customers who need a specific carrier, a direct connection to a cloud on-ramp, or low-latency access to a financial exchange cannot be accommodated by a facility that does not have the right connectivity infrastructure in place.
The commercial implications are direct. A carrier-neutral data centre — one where customers can connect to any carrier present in the meet-me room — commands a premium over a facility with restricted carrier access. Customers who need carrier choice will not locate in a single-carrier facility regardless of how competitive the power and space pricing is. The decision to build a large, well-specified meet-me room and to invest in diverse dark fibre routes from the outset is a decision about the customer universe the facility can serve.
Carrier diversity and the meet-me room
What carrier diversity means in practice
A data centre with a single carrier provides that carrier’s network reach to its customers. If the carrier has a network failure, the customer has no alternative. If the customer’s counterparties are on a carrier not present in the facility, the customer either pays for a cross-network service with the latency and cost implications that entails, or they locate elsewhere.
Carrier diversity means having multiple independent carriers present in the meet-me room, with physically separate entry points to the building. The entry points are the critical element — a facility with four carriers all entering through a single conduit route has carrier diversity on paper but not in practice. A single excavation event can take out all four simultaneously. Best practice for carrier diversity requires separate building entry points for each carrier, diverse routing through the local network, and ideally connection to more than one local exchange.
Meet-me room specification
The meet-me room (MMR) is the physical space where carriers terminate their networks and customers interconnect. The specification of the MMR — its size, power provision, cross-connect density, and the technical standards it supports — determines the carrier ecosystem the facility can accommodate.
A small MMR that was adequate for a 10MW enterprise colocation facility is not adequate for a 100MW AI campus. The cross-connect density required to support a large number of customers each with multiple carrier connections, plus cloud on-ramps and internet exchange access points, requires space, power, and technical infrastructure that needs to be planned at design stage rather than retrofitted.
Cross-connect pricing is a significant revenue line for colocation operators and a significant cost line for customers. Cross-connect prices in UK data centres range from £50 to £500 per month per circuit depending on the facility and the connection type. Exclusivity arrangements — where a carrier pays for preferential access or exclusivity in the MMR — can generate revenue for the operator but constrain the carrier ecosystem and reduce the facility’s attractiveness to customers who need carrier choice. These arrangements need to be assessed carefully against the long-term commercial interest of the facility.
Dark fibre and wayleave agreements
What dark fibre provides
Dark fibre — unlit optical fibre that the customer operates themselves — provides complete control over the network protocol, speed, and latency of the connection. Customers who need a dedicated, unshared connection between two points — between a data centre and a financial exchange, between two data centre campuses, between a facility and a private cloud environment — typically require dark fibre rather than a managed service.
The availability of dark fibre routes to and from a data centre location is a site qualification criterion for a significant subset of customers, particularly those in financial services, media, and high-performance computing. A site that does not have dark fibre available, or where dark fibre routes require significant new infrastructure to reach the site, will not be competitive for these customers regardless of its power and space offering.
Wayleave agreements
Wayleave agreements — the legal permissions that allow fibre infrastructure to cross or be installed on third-party land — are the most frequently underestimated element of dark fibre access. A dark fibre route that appears straightforward on a map may require wayleave agreements with multiple landowners, utility providers, highway authorities, and network rail if the route crosses rail infrastructure.
Securing wayleave agreements can take months to years depending on the route, the landowners involved, and whether any element of the route requires compulsory powers. Developers who leave wayleave negotiation until after the facility is built risk a situation where dark fibre access to the site is physically possible but legally unavailable on a timescale that meets their customers’ requirements. Wayleave assessment should be part of the site qualification process, not the post-construction phase.
Internet exchange access
An internet exchange (IX) is a physical infrastructure through which internet service providers and networks exchange internet traffic. Connection to an internet exchange gives a data centre’s customers access to the peering ecosystem — the direct connections between networks that reduce latency and cost compared with transit routing.
The UK internet exchange landscape is concentrated in London, where LINX (the London Internet Exchange) operates multiple platforms across several data centres. LINX is one of the world’s largest internet exchanges, with over 1,000 member networks and aggregate traffic exceeding 10Tbps. Connectivity to LINX, or to one of the smaller regional UK exchanges including LONAP, MANAP, or NWIX, is a requirement for operators whose customers include internet service providers, content delivery networks, or enterprises with significant peering requirements.
For data centres outside London, IX connectivity requires either a physical presence at a facility connected to the exchange, or a remote peering arrangement. Remote peering — connecting to an exchange platform via a dedicated circuit rather than a physical presence — has become increasingly common as the data centre market has expanded beyond established IX locations. Developers in Scotland, Wales, and the English regions need to assess IX connectivity as part of their customer proposition and ensure that remote peering options are available and commercially viable for their target customer base.
Cloud on-ramps
Cloud on-ramps are dedicated, private connectivity services offered by major cloud providers — AWS Direct Connect, Microsoft Azure ExpressRoute, Google Cloud Interconnect — that provide low-latency, high-bandwidth access to cloud platforms without routing traffic over the public internet. Enterprise and AI customers who are deploying hybrid infrastructure — compute in a colocation facility, storage or managed services in the cloud — require cloud on-ramp access from their data centre location.
Cloud on-ramp availability is increasingly a procurement criterion for enterprise and hyperscale customers. A facility without cloud on-ramp access requires customers to route cloud traffic over the public internet or to establish their own dedicated circuits to cloud access points, adding cost and latency. For AI workloads that move large datasets between on-premises GPU clusters and cloud storage, the bandwidth and latency of the cloud connection are operationally significant.
Operators who want to offer cloud on-ramp access need to establish commercial relationships with cloud providers and ensure that the physical infrastructure — the cross-connects and dedicated circuits required to support the on-ramp services — is in place. This typically requires the facility to already have a critical mass of relevant customers before the cloud provider will invest in an on-ramp deployment. It is a chicken-and-egg problem that is more easily solved from a position of an established carrier-neutral ecosystem than from a greenfield start.
Connectivity in acquisition due diligence
Connectivity infrastructure is underassessed in the majority of data centre acquisitions, and is one of the most common sources of post-acquisition disappointment. The assessment should cover: carrier diversity and the physical entry point configuration; MMR size, specification, and cross-connect capacity and pricing; any exclusivity arrangements with carriers; dark fibre route availability and the wayleave position; internet exchange access; cloud on-ramp presence; and the contractual terms of all existing connectivity agreements.
Exclusivity arrangements are a specific area of risk. A carrier that has paid for preferential MMR access will resist the introduction of competing carriers. An acquirer who does not understand the carrier exclusivity position before completion may find that the facility’s connectivity ecosystem is more constrained than the customer base and its requirements suggest.
Armason: connectivity advisory and MSP coordination for UK data centre operators
Armason works with operators and developers to assess, structure, and coordinate connectivity requirements across the full development lifecycle. We hold relationships with carrier networks, dark fibre providers, and internet exchange operators, and we coordinate the connectivity workstream as part of our broader MSP delivery model.
If you are developing or acquiring a UK data centre asset and want to understand the connectivity position, contact the Armason team.
Related reading
· Data centre construction management UK: how development programmes are delivered and where they fail
· Data centre due diligence in the UK: what buyers, operators, and investors need to assess
· Data centre colocation procurement UK: how enterprise organisations actually source capacity in 2026
· Data centre FM and operations UK: what facility management looks like for a live colocation asset
Sources referenced in this article
· Why connectivity is the most underrated data centre metric — DCD
· LINX: London Internet Exchange — in numbers
· Colocation data center pricing: a 2026 guide — DataCenterHawk
· Data centre tiers explained: Tier 1 to Tier 4 — Cagelab
· Unlocking value in UK data centre M&A transactions — BCLP
· Data Centre Connectivity: carrier neutral, carrier specific and dark fibre — Techopedia
