I Squared Capital buys $225M data center portfolio from Cogent Fiber to build AI inference platform

I Squared Capital just wrote a $225 million check for 10 data center facilities from Cogent Fiber, LLC. The acquisition gives the Miami-based infrastructure investment firm roughly 53 megawatts of installed power capacity and 259,000 square feet of colocation space spread across nine major US markets.

The deal in detail

The portfolio spans facilities in Chicago, Atlanta, Phoenix, and multiple locations across California and Texas. These are legacy, purpose-built assets sitting in dense urban areas.

I Squared Capital has committed to investing up to $1 billion total for upgrades, customer-led expansions, and additional acquisitions as part of this new platform. The new platform will focus on three areas: colocation services, high-density deployments, and AI inference infrastructure.

On the seller’s side, Cogent Communications plans to use the proceeds primarily for debt reduction. The transaction requires regulatory approvals and is expected to close in Q3 2026, with the earliest potential closing date set for June 12, 2026.

Why private equity is flooding into data centers

Cogent Communications, renowned for its expansive fiber-optic network, has historically managed various data center assets, some of which trace back to infrastructure acquired from Sprint. This strategic sale allows Cogent to monetize non-core real estate assets at a roughly $4.25 million per MW multiple, freeing up capital for its primary connectivity and telecom business.

The choice to focus on edge and urban facilities rather than massive hyperscale campuses is also telling. Hyperscale builds, the kind that Microsoft, Google, and Amazon develop, can cost billions per site and take years to construct. Urban colocation facilities can be upgraded and brought to market faster, serving the growing cohort of enterprises that need AI inference but don’t want to build their own infrastructure.

What this means for investors

As AI workloads become more lucrative for data center operators, facilities that once hosted crypto mining operations face economic pressure to reallocate rack space toward higher-margin AI tenants. Bitcoin miners in particular have already experienced this dynamic, with several publicly traded mining companies pivoting portions of their capacity toward AI and high-performance computing.

The risk, as always with infrastructure plays, is execution. Retrofitting legacy data centers for high-density AI workloads is expensive and technically complex. Power availability in urban markets is constrained, with some cities imposing moratoriums on new data center construction due to grid capacity concerns.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.