Vantage’s 1.4GW Texas campus makes grid contracts the real data center schedule

Vantage Data Centers says construction is underway on a 1.4-gigawatt campus in Shackelford County, Texas, and it expects the first building to deliver in the second half of 2026. The company pegged the total investment at more than $25 billion and framed the site, branded “Frontier,” as an AI-driven hyperscale hub. That is the real operational headline: when a single project is measured in gigawatts, the grid agreement and the cooling architecture stop being background engineering and become the schedule.

Business Wire reported the plan calls for 10 data centers on roughly 1,200 acres, totaling 3.7 million square feet, and designed for “ultra high-density” racks of 250kW-plus supported by liquid cooling. Those specs pull procurement forward. Switchgear lead times, busway strategy, CDU and heat rejection selections, and commissioning labor all start to look like critical-path items, not line items that can be finalized later.

PwC’s $31.6T projection is a procurement signal, not a forecast to admire

Data Center Frontier’s coverage of PwC’s inaugural Global Data Centre Outlook put a number on what operators have been living for two years: PwC projects $31.6 trillion in AI data center buildout through 2050. The report’s framing, as relayed by Data Center Frontier, emphasizes a multi-decade cycle where recurring GPU and server investment increasingly dominates, and where power availability is the gating factor.

In practice, that projection changes behavior well before 2050. It gives CFOs and procurement leaders permission to treat power, cooling, and hardware refresh as recurring capacity subscriptions rather than one-off capex events. The facilities team is left with a blunt reality: if AI compute is the business plan, the utility is now a strategic supplier.

ERCOT is stress-testing how Texas will actually serve AI megawatts

Texas is where that supplier relationship gets tested in public. Data Center Frontier reported on ERCOT’s efforts to put “Texas AI megawatts to the test,” signaling a more operationally hands-on posture toward very large loads. For operators, ERCOT’s scrutiny matters less as a policy story than as a design requirement: large campuses may need to plan for curtailment, staged energization, and more explicit load flexibility terms, depending on how interconnection and reliability requirements are written.

That lands directly on the people writing specifications today. A campus designed around 250kW racks will likely require phased commissioning and careful load growth planning, because the electrical and mechanical plants must be right-sized without stranding capital. ERCOT’s posture, as described by Data Center Frontier, suggests that “build it and they will power it” is not a strategy. It is an assumption to be negotiated.

‘Frontier’ now means two different contracts operators will have to manage

Adding to the naming confusion for search results and procurement files, “Frontier” is also a carbon-removal purchasing consortium. DataCenterDynamics reported June 18, 2026 that the consortium plans to put another $915 million into carbon removal technologies, raising its total commitment to $1.8 billion, and that Anthropic joined as a member.

These specifics matter to enterprise sustainability teams because the Growth AMC is built for commercial scale, beyond pilot programs. DataCenterDynamics reported the initiative will concentrate on a tighter set of roughly 10 to 15 companies and will use 8 to 10-year offtake agreements intended to move projects to a final investment decision, with some contracts running out to 2040. Frontier also outlined five priority technology pathways, and it has reviewed more than 500 companies across more than 20 pathways, backing 52 projects since launch in 2022, according to DataCenterDynamics.

This is where carbon removal stops being a brand statement and becomes a sourcing motion. Long-duration offtake agreements create obligations that look more like energy PPAs than like annual credit buys. For data center operators under customer pressure to document emissions strategy, the operational question becomes: can reporting, verification, and delivery schedules for removals be managed with the same rigor as fuel, power, and water metrics?

Where this lands for data center ops, facilities, and sustainability procurement

Taken together, the Vantage build plan, PwC’s long-horizon spend projection, ERCOT’s grid scrutiny, and Frontier’s larger carbon-removal commitments point to the same operational shift: the build is no longer just a construction project. It is an ongoing set of interdependent supply contracts across power, heat rejection, hardware refresh, and carbon accounting.

For operators with highly volatile AI training demand or customers that reserve the right to swing workloads across regions, this matters even more. High-density AI halls create bigger penalties for downtime and slower penalties for “good enough” cooling. That drives a different posture toward redundancy, spare parts, and service-level agreements with cooling and electrical OEMs.

Questions to take into design reviews and contract negotiations now

• Grid and interconnect: What are the explicit curtailment or load-shed expectations embedded in the utility or ISO agreement, and how will they be tested during commissioning? Data Center Frontier’s reporting on ERCOT’s AI megawatt scrutiny is a prompt to get those terms in writing early.
• Cooling and density: If the target rack density is 250kW-plus with liquid cooling, what is the failure mode for a CDU or heat rejection constraint, and how is it isolated without taking down an entire row? Business Wire reported Vantage designed Frontier for 250kW-plus racks and liquid cooling, which should be treated as a spec baseline, not a science project.
• Sustainability procurement: If the organization is considering carbon-removal offtake, what is the verification standard, delivery schedule, and revocation or replacement mechanism if a project misses milestones? DataCenterDynamics reported Frontier’s Growth AMC centers on 8, 10 year offtake agreements extending as far as 2040, and that length changes risk management.
• Capex timing: How much of the site’s spend must be committed before power is contractually firm, and what triggers each phase gate? PwC’s $31.6T through-2050 framing, as covered by Data Center Frontier, implies more recurring build cycles where phase discipline is the only way to avoid stranded infrastructure.