I Squared’s $1 Billion Edge Push Turns AI Inference Into a Location-Power-Connectivity Trade

I Squared Capital’s May 26 announcement is worth publishing because the useful signal is not just that another investor bought data center assets. The stronger signal is that private infrastructure capital is starting to underwrite AI inference as a geography problem. If training demand concentrated value in giant campuses with massive power blocks, I Squared is arguing that the next layer of value may sit in smaller facilities where location, immediate power, and fiber density matter more than raw megawatt spectacle.

The primary-source facts are specific enough to matter. I Squared says it agreed to buy ten data center facilities from Cogent Fiber for $225 million in cash, then commit up to $1 billion more to build a new U.S. platform focused on colocation, high-density deployments, and AI inference infrastructure. The seed portfolio includes roughly 53 megawatts of installed power capacity and about 259,000 square feet of available colocation space across nine U.S. markets: Chicago, Atlanta, Phoenix, Los Angeles, Kansas City, Baltimore, Houston, Nashville, and Stockton.

The useful I Squared signal is not that more money went into data centers. It is that AI inference is starting to be priced as a location, power, and connectivity trade.

That is the original Grid Report angle. The real product being purchased here is not just data center square footage. It is a combination of location, power, and connectivity in supply-constrained metros where low-latency inference can matter. I Squared’s own framing is explicit: “location, power, and connectivity” determine long-term value, and as AI shifts from training to everyday use, demand for high-density, low-latency facilities should grow.

This clears the duplicate block against the site’s recent markets and infrastructure coverage. The CME compute-futures article was about pricing and hedging GPU capacity. The Big Fiber financing piece was about dark fiber as a campus bottleneck. NVIDIA’s cloud ecosystem story was about regional compute distribution through partner clouds. I Squared is different. It is about a private-infrastructure investor assembling an asset platform around inference geography rather than training concentration.

The market implication is that AI infrastructure capital may be segmenting faster than many investors still assume. Not every dollar will chase 500-megawatt campuses. Some of it will chase facilities that are already powered, already connected, close to internet exchanges, and capable of supporting liquid-cooled high-density deployments without waiting years for a greenfield build. That changes the investment lens from “who has the biggest site?” to “who has the fastest useful footprint?”

There is still a caveat worth stating. The transaction has not yet closed and remains subject to regulatory approvals, including Hart-Scott-Rodino review. So this is not proof that edge inference economics are already settled. But the fact that a $60 billion infrastructure manager is committing platform capital around that thesis is itself a useful market signal, especially because the company is not buying one trophy campus. It is building a repeatable portfolio.

For operators, the practical takeaway is that latency-sensitive inference demand may support more regional and metro-adjacent capacity than the current hyperscale narrative implies. For investors, the sharper question is which assets combine immediate power, dense connectivity, and expansion room in markets where new supply is constrained. That bundle may start to command a premium as AI traffic moves closer to end users and enterprise workloads.

For search performance, the article is strong because it answers a specific, emerging question: why would an infrastructure investor commit $1 billion to edge and inference colocation right now? Readers searching for I Squared Capital AI inference, Cogent Fiber data center sale, or edge colocation AI infrastructure get a clear capital-markets thesis rather than a bare acquisition recap.

Business Wire, “I Squared Capital Launches U.S. AI Inference and Edge Colocation Data Center Platform With $1BN Commitment,” published May 26, 2026: https://www.businesswire.com/news/home/20260525403366/en/I-Squared-Capital-Launches-U.S.-AI-Inference-and-Edge-Colocation-Data-Center-Platform-With-%241BN-Commitment

I Squared Capital company background as described in the same release, accessed June 3, 2026: https://www.isquaredcapital.com/