CloudHQ Raises Its Mexico Data Center Bet to $6B as Power Capacity Quadruples
CloudHQ raised its Querétaro campus investment to almost US$6B — up $1.2B — while quadrupling planned power capacity to ~1,000 MW. The six-facility layout is unchanged: the extra money buys electricity transmission, not buildings.
CloudHQ has raised its planned investment in its Querétaro, Mexico campus to almost US$6 billion — up $1.2 billion from the $4.8 billion announced in September 2025 — and quadrupled the campus's planned power capacity from 250 MW to roughly 1,000 MW. The six-facility configuration is unchanged. The extra money is not buying buildings; it is buying electricity transmission capacity for cloud and AI workloads. The announcement came October 19 during a meeting between CloudHQ executives and Querétaro governor Mauricio Kuri González. Construction is on schedule, and the CloudHQ Mexico data center expansion is now officially a power project.
The CloudHQ Mexico data center is a power project in disguise
Say it plainly: a $1.2B increase with zero new buildings means the constraint was never the buildings. The CloudHQ Mexico data center campus is now a 1,000 MW electrical infrastructure project that happens to include server halls. This is the story of the whole AI buildout in one campus — the real bottleneck is power, chillers, and permits, not GPUs. CloudHQ is just writing the check before the shortage writes it for them. Anyone still treating data center announcements as real-estate stories is reading the wrong spreadsheet.
Why Querétaro
The 55-hectare site sits on State Highway 100 (El Colorado–Higuerillas), and the critical asset is the dedicated ~2 GW maneuvering substation being built with CFE and Cenace — Mexico's grid utility and energy regulator. That is the kind of infrastructure that takes years to permit, which is exactly why it is the moat. The campus uses waterless cooling, which matters because Querétaro is water-stressed: a data center that drinks the local aquifer would be a political liability, not an asset. Target: ready around Q2 2027 with roughly 900 permanent jobs. Mexico's nearshoring pitch — US-adjacent compute with lower land and labor costs — only works if the electrons are guaranteed. CloudHQ just bought the guarantee.
Querétaro's rise as a data center hub is no accident. The state has courted cloud investment for years with tax incentives and fast-tracked permits, betting that powered industrial land would become the scarcest asset in the AI era. That bet is paying off: operators have stacked announcements across central Mexico through 2026, and the corridor between Querétaro and Mexico City is starting to look like a northern-hemisphere answer to Virginia's Data Center Alley. The difference is that Querétaro's grid is being built for AI from day one — no retrofits, no substation fights in suburbs that never wanted them.
One campus, one gigawatt, one strategy
CloudHQ runs about 56 data centers, operational or in development, totaling roughly 1.1 GW across the US, EMEA, and APAC. That puts Querétaro in perspective: a single campus approaching a gigawatt is one of the largest single bets in the company's portfolio. Demand is not the question — data center demand doubled in 2026, per JLL — so the fight is over who can deliver powered, permitted capacity first. Hyperscalers can build their own; everyone else rents from operators like CloudHQ. The winners of this cycle are the ones holding interconnection agreements, not the ones holding architectural renders.
What this means if you run infrastructure
Three practical takeaways. First: size the power before you size the building. If you are planning capacity for 2027, ask providers for power delivery timelines and substation status, not just rack counts. Second: watch the grid partnerships — deals with utilities and regulators are now as strategic as deals with chip vendors. Third: the ops tooling around AI data centers is its own growth story — NetBox's Nvidia and Grafana integrations show the monitoring layer racing to keep up with denser, hotter racks. The AI buildout's binding constraint is electrical infrastructure. CloudHQ's extra $1.2B just said so out loud.


