Feasibility File · Vantage Data Centers

Frontier

Shackelford County · TX

Real enough to take seriously and too under-documented to score as delivery-ready. Physically plausible, commercially and regulatorily unproven.

Verdict

4.1/ 10

Bearing Labs feasibility score

Frontier belongs in the weak-but-plausible band until Vantage, Oracle/OpenAI, VoltaGrid, ERCOT-facing counterparties, or Texas regulators place the power, permit, customer, and cost-allocation record in public view. The practical posture is gated confidence: do not treat Frontier as delivery-ready until the power stack is public; do not treat the water claim as fully proven until the WUE and water-source record is public; do not treat Stargate demand as fully durable until phase MW commitments and customer protections are public; do not treat the 2028 buildout as substantiated until permits, energization milestones, heavy-haul logistics, and long-lead procurement are public. Score should remain 4.1 until specific named evidence is public: executed power and interconnection records for material portions of the 1.4 GW load; TCEQ air and emissions records for any onsite gas generation; phase-by-phase (source on file) customer commitments; a source-specific water balance; and county/Texas permit records showing that first-building delivery and full-campus buildout are on an evidenced path.

Where the story diverges

What the operator says

Vantage Data Centers' largest campus globally and its first operating Texas facility: a 1.4 GW hyperscale AI campus in Shackelford County, near Albany and the Abilene corridor, planned across roughly 1,200 acres with 10 single-story data centers totaling about 3.7 million square feet. Announced as a more-than-$25 billion investment, construction beginning December 2025, Kiewit as general contractor, first building delivery targeted for H2 2026, broader buildout running toward 2028.

Tied publicly to the Stargate expansion associated with OpenAI and Oracle. Public materials emphasize ultra-high rack densities, carrier-neutral connectivity, a closed-loop chiller and combined air/liquid cooling design with near-zero water utilization, LEED pursuit, local scholarships, and supportive statements from Texas Gov. Greg Abbott and Shackelford County Judge John Viertel.

What's actually there

Frontier has the right sponsor profile, land scale, regional power-market logic, and low-water cooling concept to be plausible. Vantage is backed by DigitalBridge and Silver Lake; Shackelford County sits in a deep West Texas power corridor with dozens of in-service lines within 25 miles, including multiple 345 kV-class lines owned by AEP Texas North, Big Country Electric Coop, and Oncor. ERCOT is a flexible bilateral market with natural gas, renewable resources, and a regulatory structure that can move faster than vertically integrated utility markets when commercial terms are clear.

The current record has three materially different power narratives that have not been reconciled: Vantage's public-facing 'multiple, diverse power feeds' language with N+1 redundancy; Stargate-linked material identifying Shackelford County and naming Vantage and VoltaGrid as development partners with an on-site, behind-the-meter, gas-powered microgrid using Jenbacher reciprocating engines; and a utility/regulatory record that does not yet show an executed utility-service agreement, ERCOT/TSP study milestone, transmission-upgrade package, power-supply contract, PPA, gas-generation permit, emissions profile, fuel-supply plan, or customer cost-protection structure for the full 1.4 GW. ERCOT was tracking ~226 GW of large loads seeking interconnection and just over 7 GW had submitted Full Interconnection Study applications, which makes large-load interconnection a systemwide gating issue. Frontier is not impossible — but until the power stack is public, it remains a strategy outline, not a proven delivery path.

Risk 01

Single most decision-relevant exposure

Firm 1.4 GW power deliverability and cost allocation

Frontier's other claims — H2 2026 first-building delivery, 2028 buildout trajectory, grid coexistence, sustainability, customer durability, and economic viability — all depend on whether 1.4 GW can be delivered in phases at a known cost without creating unacceptable reliability, emissions, ratepayer, or permitting exposure. A credible solution could combine staged grid service, behind-the-meter gas, PPAs, bridge power, batteries, and later transmission upgrades. But until the public record shows that stack, the power story is a strategy outline, not a proven delivery path.

Claim Scorecard

Every announced claim, scored.

Each claim assessed independently against supporting and contradicting evidence in the full report.

01
4.0/ 10
Frontier can reliably deliver 1.4 GW through 'multiple, diverse power feeds'
Preliminary
Nearby 345 kV infrastructure and ERCOT market depth make the claim physically plausible; no public utility, interconnection, generation, upgrade, or cost-allocation record substantiates it.
4.0/ 10
02
4.0/ 10
First building in H2 2026 and broader 1.4 GW campus toward 2028
Preliminary
Kiewit, land scale, and phasing support buildability; schedule proof is missing for power, permits, civil readiness, heavy-haul logistics, and long-lead electrical equipment.
4.0/ 10
03
5.0/ 10
Near-zero-WUE cooling can support 1.4 GW without material municipal water impact
Meaningful but qualified
Closed-loop/liquid cooling is directionally credible — Frontier's strongest claim — but no public annual/peak water balance, source agreement, drought plan, or municipal no-impact finding validates the claim at scale.
5.0/ 10
04
3.5/ 10
Coexistence with adjacent GW-scale Abilene/Shackelford corridor campuses without destabilizing reliability
Preliminary
ERCOT's large-load evidence points to system stress; no coordinated cluster deliverability or reliability plan is public.
3.5/ 10
05
5.0/ 10
OpenAI/Oracle Stargate demand is durable enough to underpin Frontier
Preliminary
Stargate linkage is real; executed take-or-pay terms, phase MW commitments, customer credit support, and cancellation protections are not public.
5.0/ 10
06
4.0/ 10
Project remains economically viable at more than $25 billion
Preliminary
Sponsor strength and AI demand support plausibility; illustrative economics require durable contracted net cashflow and known power costs not yet visible — ~$3.5B/yr of net operating cashflow needed to break even before residual value at $25B capex and 8.75% WACC.
4.0/ 10
07
3.5/ 10
Rural Shackelford/Abilene labor market can support 5,000+ workers
Preliminary
Regional access and Kiewit help; Shackelford County (~3,197 residents, ~1,591 housing units) and competing corridor projects make the public workforce claim weak.
3.5/ 10
08
3.5/ 10
Net-zero-by-2030 / 'renewable whenever possible' commitments for 1.4 GW 24/7 AI load
Preliminary
Water-side design helps; undisclosed power supply, possible gas generation, and no public carbon-free procurement plan keep the energy-sustainability claim weak.
3.5/ 10

Cross-Cutting Risk Factors

The interfaces that decide the project.

01
Power schedule drives everything else
The H2 2026 first-building target is not primarily a shell-construction question. It is a phase-one energization question. If the first building requires grid upgrades, a gas microgrid, bridge power, or unusual interconnection sequencing, the schedule becomes a power-delivery and permitting schedule.
02
Behind-the-meter gas can solve one problem and create three more
Onsite gas generation could reduce grid-interconnection delay and improve controllability. It also creates air-permitting, emissions, fuel-supply, noise, community-acceptance, and sustainability issues. The project's score would improve if the public record showed a clean, permitted, time-bounded, cost-allocated gas strategy. It would weaken if gas emerges late as an undisclosed workaround.
03
Water is the bright spot, but only if quantified
Closed-loop/liquid cooling is the strongest technical claim in the package. The unresolved question is whether Frontier's 1.4 GW implementation remains near-zero-WUE under West Texas summer conditions, commissioning, maintenance, fire/life-safety, and emergency operations. At 1.4 GW operating continuously, every 0.01 L/kWh-IT equals ~32 million gallons per year.
04
Workforce and local services can become the political bottleneck
A rural county can welcome a major project and still struggle with worker housing, road damage, emergency response, inspections, and public-service capacity. The current record shows benefit messaging; it does not show a full burden-mitigation plan.
05
Incentives and ratepayer protections will matter
A project above $25 billion is likely to implicate state tax incentives, local abatements or infrastructure arrangements, and utility cost-allocation debates. The public will eventually ask who pays if load ramps slip, customer demand changes, or transmission/generation investments are stranded. Meta Richland Parish shows the protective terms that can answer that question; Frontier has not yet shown them publicly.
06
Conflicting public scale descriptions should be reconciled
The main Vantage project profile describes 1,200 acres, while Stargate-linked public material refers to a 700-acre Shackelford County campus. That discrepancy should be reconciled because acreage affects grading, water, road, buffering, tax, and environmental analysis.

Project Profile

Reference data from the public record.

Project: Vantage Data Centers 'Frontier'
Stage tag: Under Construction
Location: Shackelford County, TX, near Albany · Abilene corridor / ERCOT West
Developer / operator: Vantage Data Centers (DigitalBridge, Silver Lake-backed)
Commercial anchor: OpenAI / Oracle Stargate expansion · Vantage and VoltaGrid named as development partners
Scale: 1.4 GW critical IT load · 10 single-story data centers · ~3.7M sq ft
Site size: Vantage profile: ~1,200 acres · Stargate-linked material: ~700 acres — discrepancy
Investment: >$25B announced
Schedule: Groundbreaking Dec 2025 · first building H2 2026 · broader buildout aimed at 2028
General contractor: Kiewit
Cooling: Closed-loop chiller / combined air-cooled + liquid-to-liquid · claimed near-zero WUE
Power narrative: Vantage public: 'multiple, diverse power feeds.' Stargate-linked: on-site, behind-the-meter Jenbacher gas microgrid
ERCOT context: ~226 GW of large loads seeking interconnection; ~7 GW had submitted Full Interconnection Study applications
Connectivity: AT&T, FiberLight, Windstream, Zayo within a few miles
Scoring arithmetic: Claim mean 4.0625 → 4.1 · controlling claim 4.0 · cap 5.5 (does not bind)
Evidence as of: May 29, 2026

Bearing Labs · Feasibility Brief

Frontier Feasibility

Full claim-by-claim assessment with supporting and contradicting evidence, reference-class comparisons, cross-cutting risks, key data gaps, and recommendation.

PDFDownload

Companion reports — Power & Grid Path, Permitting & Litigation, Community Impact, and Schedule Forensics — are produced on request.

What the operator says

What's actually there

Frontier

What the operator says

What's actually there

Firm 1.4 GW power deliverability and cost allocation

Every announced claim, scored.

Frontier can reliably deliver 1.4 GW through 'multiple, diverse power feeds'

First building in H2 2026 and broader 1.4 GW campus toward 2028

Near-zero-WUE cooling can support 1.4 GW without material municipal water impact

Coexistence with adjacent GW-scale Abilene/Shackelford corridor campuses without destabilizing reliability

OpenAI/Oracle Stargate demand is durable enough to underpin Frontier

Project remains economically viable at more than $25 billion

Rural Shackelford/Abilene labor market can support 5,000+ workers

Net-zero-by-2030 / 'renewable whenever possible' commitments for 1.4 GW 24/7 AI load

The interfaces that decide the project.

Power schedule drives everything else

Behind-the-meter gas can solve one problem and create three more

Water is the bright spot, but only if quantified

Workforce and local services can become the political bottleneck

Incentives and ratepayer protections will matter

Conflicting public scale descriptions should be reconciled

Reference data from the public record.

Frontier Feasibility

Frontier

What the operator says

What's actually there

Firm 1.4 GW power deliverability and cost allocation

Every announced claim, scored.

Frontier can reliably deliver 1.4 GW through 'multiple, diverse power feeds'

First building in H2 2026 and broader 1.4 GW campus toward 2028

Near-zero-WUE cooling can support 1.4 GW without material municipal water impact

Coexistence with adjacent GW-scale Abilene/Shackelford corridor campuses without destabilizing reliability

OpenAI/Oracle Stargate demand is durable enough to underpin Frontier

Project remains economically viable at more than $25 billion

Rural Shackelford/Abilene labor market can support 5,000+ workers

Net-zero-by-2030 / 'renewable whenever possible' commitments for 1.4 GW 24/7 AI load

The interfaces that decide the project.

Power schedule drives everything else

Behind-the-meter gas can solve one problem and create three more

Water is the bright spot, but only if quantified

Workforce and local services can become the political bottleneck

Incentives and ratepayer protections will matter

Conflicting public scale descriptions should be reconciled

Reference data from the public record.

Frontier Feasibility