A Border-Adjacent Bet on Renewable Power
Cuasar Capital, part of the Spanish energy and infrastructure group EOSOL, and Chile-experienced developer KA URBE IT are building what would rank among Mexico's largest data center developments: a campus in Durango anchored by DC Norte Uno, a 300 megawatt facility with 250 megawatts available for IT load on a 20-hectare site, built to Tier III standards. A smaller companion facility, DC Norte Dos, adds another 7.74 megawatts. Combined investment across both facilities runs approximately $3.4 billion.
What makes Durango specifically attractive is the power behind it. EOSOL already operates a photovoltaic portfolio exceeding 1 gigawatt in the region, with another 4.5 gigawatts in development, meaning the data center campus is being built next to renewable generation the developer already controls rather than waiting on a utility interconnection queue the way most US projects now have to. That is the core structural advantage this deal is selling, and it is a direct answer to the exact bottleneck slowing projects north of the border.
Solving the Connectivity Problem That Killed Nearshoring Before
Nearshoring data center capacity to Mexico has been discussed for years without much follow-through, largely because connectivity and reliability concerns made it a hard sell for latency-sensitive AI workloads. This project addresses that directly: the campus connects through the Phoenix to Queretaro corridor with access to three separate carriers, giving it a real fiber path back into US network infrastructure rather than relying on a single, thinner connection that would make it a liability during any outage.
That connectivity detail is why this reads as a genuine AI and edge infrastructure play rather than a low-cost backup site for archival workloads. High-density AI inference in particular is sensitive to round-trip latency in a way that older enterprise workloads were not, and a campus explicitly marketed toward AI and edge, industrial loads with credible North American connectivity is targeting exactly the workload category currently struggling to find capacity anywhere in the southwestern US at a reasonable price and timeline. That positioning is a deliberate choice by the developers, not an afterthought, and it explains why the connectivity corridor gets equal billing with the power story in how the project is being marketed to prospective tenants.
Timing That Is Not a Coincidence
This campus is being built at the exact moment Texas has paused new data center interconnection approvals statewide and Pennsylvania has layered new local-approval requirements onto its state permitting process. Developers and their capital partners read regulatory friction in the largest US markets as a business opportunity elsewhere, and a border-adjacent site with its own renewable power supply and established US fiber connectivity is a rational response to that friction, not a speculative long shot.
It is worth being direct about what this represents for the broader capacity market. As US interconnection queues, permitting reviews, and community opposition all add months or years to domestic projects, capital is going to keep flowing toward sites that can promise renewable power without a queue and construction timelines without a moratorium risk attached. Durango is one instance of a pattern that enterprise capacity planners should expect to see repeated across multiple border-adjacent and Latin American markets over the next several years.
What Enterprise Buyers Actually Get, and Give Up
For a US enterprise evaluating whether to route workloads through a facility like this one, the upside is real: renewable power without the wait, Tier III reliability standards, and a fiber path that keeps latency workable for many AI inference and edge use cases. The tradeoffs are equally real and need explicit sign-off from legal and compliance before any production workload moves there, and that review should happen early in the vendor evaluation process rather than as a late-stage checkbox after commercial terms are already agreed.
Cross-border data flows into Mexico raise data sovereignty and regulatory questions that vary significantly by industry and by the country your end customers sit in, and Mexico's own regulatory environment for data centers and cross-border data handling is younger and less tested than what US enterprises are used to navigating domestically. Currency exposure, longer-term political risk around US-Mexico trade relations, and the practical question of who has boots on the ground for incident response all belong on the same evaluation sheet as the power and latency numbers. None of that makes the site a poor choice, it simply means the total cost of ownership calculation has to include line items that a familiar domestic colocation deal would never require in the first place.
The Site Selection Map Just Got Bigger
The practical takeaway for CTOs building a multi-year capacity roadmap is that domestic US markets are no longer the default, lowest-friction option they were even two years ago, and geographic diversification into markets like Durango deserves a real seat at the table alongside traditional US and established international options, not just a passing mention in a slide about future optionality that never gets revisited when the next budget cycle comes around and capacity actually needs to be sourced on a real deadline.
Every workload does not belong on a site like this one, and workloads with strict domestic data residency requirements should stay exactly where they are today. The capacity planning conversation now needs a genuine cost and risk comparison between waiting out a US interconnection queue, paying a premium for behind-the-meter power domestically, or moving qualifying workloads to a site like Durango where the power and the queue-free timeline are the entire value proposition. Treat this as confirmation that site selection has expanded into a genuinely global exercise, spanning well beyond the short list of familiar US metros most roadmaps still default to.
