Why Texas?
Texas has emerged as the preferred destination for hyperscale AI data centre construction for several converging reasons: a deregulated energy market that allows direct power purchasing agreements, land availability at scale, a business-friendly regulatory environment, no state income tax for corporations, and proximity to existing tech talent clusters in Austin, Dallas, and Houston.
The result is an unprecedented build-out of AI computing infrastructure in a single US state — and the implications stretch from power grids to real estate to global chip supply chains.
The Build-Out in Numbers
Several of the world's largest AI investments are concentrated in Texas:
- xAI Colossus (Memphis → Texas expansion): 100,000+ GPU training cluster targeting H100 and next-generation Blackwell chips.
- Tesla Dojo: Custom AI training supercomputer infrastructure at Giga Texas in Austin.
- Microsoft: Multiple hyperscale data centre campuses across the Dallas-Fort Worth corridor.
- Google: Major investment in Midlothian and other Texas sites for AI workloads.
- Amazon Web Services: Significant expansion of Texas cloud regions.
The Power Constraint
The critical bottleneck is power. A single large AI training cluster can consume 100–500 megawatts of electricity — equivalent to powering tens of thousands of homes. Texas runs its own grid (ERCOT), which already faced near-collapse during the February 2021 winter storm and faces growing summer peak demand from data centres and population growth.
Texas data centre demand is projected to grow by 20–30 GW over the next five years. ERCOT's current total capacity is approximately 100 GW. AI alone could add 20–30% to peak demand.
This power constraint is creating investable opportunities in the energy sector. Natural gas peaker plants, nuclear small modular reactors (SMRs), and grid-scale battery storage (particularly Tesla Megapack) are all benefiting from the AI electricity boom.
Investment Implications
Direct Beneficiaries
- NVIDIA (NVDA): GPU demand driven by hyperscale build-out
- Vistra Energy (VST): Texas power generation at scale
- NRG Energy: Texas grid power provider
- Vertiv (VRT): Data centre cooling and power management
- Eaton Corporation: Electrical equipment for data centres
- Tesla (TSLA): Megapack energy storage for AI campuses
Indirect Beneficiaries
- Industrial REITs with Texas exposure
- Commercial construction contractors
- Copper wire and cable manufacturers (copper demand surges with electrification)
- Water utilities (data centres use significant water for cooling)
"Power is the new oil. Every GPU running needs electrons behind it — and Texas is building the grid of the AI era."
Strategic Context
The Texas AI capacity build-out is part of a broader US strategy to dominate global AI infrastructure. With China effectively barred from buying the most advanced Nvidia chips, the US is accelerating domestic AI compute capacity to create an insurmountable technological lead.
Texas is at the centre of this strategy — and investors who understand the energy, infrastructure, and semiconductor layers of this build-out are positioned to benefit from one of the largest capital investment cycles in modern history.
- Texas is becoming the AI compute capital of the US — driven by land, power access, and regulatory environment.
- Power availability is the binding constraint — creating investable opportunities in energy generation and storage.
- NVIDIA, Vistra, Vertiv, and Tesla Megapack are direct beneficiaries of the infrastructure build-out.
- ERCOT grid stress creates risk — but also investment opportunity in grid stabilisation technologies.
- This is a multi-year capital cycle — the Texas AI build-out is a decade-long structural trade, not a near-term momentum play.