Zayo & Nvidia boost US fibre for AI infrastructure

Zayo has teamed up with Nvidia to expand long-haul fibre capacity for AI infrastructure, including more than 8,000 miles of new fibre across US AI corridors.

The project is intended to support growing data traffic tied to AI workloads as demand spreads beyond established data centre hubs. Zayo plans to add six new long-haul routes and overbuild existing network infrastructure in 10 high-demand markets.

The agreement reflects a broader shift in the AI market, where network links are drawing more attention alongside chips and data centres. Industry executives and analysts increasingly see long-haul connectivity as a constraint as GPU clusters, AI factories, and hyperscale deployments become more geographically distributed.

Zayo is focusing its build-out on areas where AI demand is emerging, rather than only adding infrastructure on established routes. The company has been modelling where AI-related traffic is likely to rise across the US while expanding bandwidth on existing assets and constructing new corridors.

Steve Smith, Chief Executive Officer at Zayo, said the company had been planning for changes in where network infrastructure would be needed.

"AI is fundamentally reshaping where and how network infrastructure needs to be built across the U.S. Zayo has invested significantly in modelling where AI-driven demand will emerge and is actively expanding infrastructure ahead of that demand," said Steve Smith, Chief Executive Officer at Zayo.

"As AI adoption accelerates and compute demand grows across the ecosystem, the need for new network corridors and scalable connectivity is increasing. While much of the market remains focused on overbuilding existing routes, Zayo is one of the few providers adding capacity to existing infrastructure while building in the locations and at the scale AI requires for long-term growth. Zayo's work combines decades of experience building and operating large-scale network infrastructure with NVIDIA accelerated computing to support AI connectivity across North America," Smith added.

Network constraint

The collaboration highlights how the AI infrastructure discussion is expanding beyond processors to include the systems that move data between facilities. As AI computing becomes more distributed, operators need connections between training sites, inference locations, and interconnection points.

Dylan Patel, Chief Executive Officer at SemiAnalysis, said the next bottleneck was increasingly connectivity between sites rather than only the availability of compute.

"The next constraint for AI is not just compute - it is the ability to connect massive, distributed AI infrastructure at scale. As GPU clusters, AI factories, and hyperscaler deployments expand beyond the traditional data center hubs, long-haul fiber becomes a critical layer of the AI supply chain," said Dylan Patel, Chief Executive Officer at SemiAnalysis.

"Zayo's work with NVIDIA directly addresses one of the most important infrastructure gaps in the market: building new high-capacity corridors where AI demand is emerging, not just adding capacity where networks already exist," Patel added.

Nvidia also framed the effort as part of a broader need to match advances in AI computing with more network capacity. The chip designer has been extending its reach beyond semiconductors into systems and infrastructure tied to AI deployment.

"AI is moving faster than ever, and the network is quickly becoming just as critical to that progress as compute itself," said Vladimir Troy, Vice President of Engineering, AI Infrastructure at NVIDIA. "Zayo's work with NVIDIA is helping us get ahead of that curve, pairing Zayo's expertise in large-scale network infrastructure with NVIDIA's AI leadership to build the connectivity backbone the entire ecosystem needs to keep innovating."

Wider expansion

Zayo said the latest project is part of a broader AI-focused network programme. Over the past 18 months, its build and overbuild work has reached more than 15,000 route miles across North America.

The company also pointed to its acquisition of Crown Castle's Fibre Solutions business, which added 90,000 metro route miles and 40,000 on-net enterprise locations. The deal strengthened its metro footprint, which is relevant for inference workloads that often need dense local connectivity as well as long-haul links.

Zayo argued that better connectivity is not only an issue for the largest cloud operators. Neocloud providers, frontier model developers, and businesses in sectors such as healthcare, finance, and manufacturing also need access to the same underlying network infrastructure to bring new AI systems online.

Smith said network availability could influence how quickly newer cloud providers deploy compute resources for customers.

"If we look at neoclouds specifically, we know their customers depend on them to make compute infrastructure accessible as soon as they need it, and network capacity is imperative to making that happen," Smith said.

"As AI infrastructure becomes more distributed, access to high-capacity connectivity in the right markets is becoming critical to how quickly providers like neoclouds can bring new GPU capacity online and support customer demand. That's what makes our work with NVIDIA so important. Building new AI corridors where infrastructure is actually scaling helps remove a major bottleneck for the broader AI ecosystem," he added.

Bob Laliberte, Principal Analyst, Networking & Observability at theCUBE Research, said the move reflected a growing view that AI roll-outs depend on coordinated planning across several layers of infrastructure.

"This is more than a network expansion; it's a strategic move to enable distributed environments and, more specifically, the AI ecosystem. By working with NVIDIA and building additional capacity in key AI corridors, Zayo is enabling not just hyperscalers, but also neocloud providers and enterprise AI deployments to have the WAN connectivity required to scale. This kind of forward-looking strategy reflects a growing recognition that AI success depends on coordinated infrastructure across compute, storage, and network," said Bob Laliberte, Principal Analyst, Networking & Observability at theCUBE Research.