Energy Vault Starts Construction on Texas AI Power Campus

Energy Vault Starts Construction on Texas AI Power Campus

2026-07-28 companies

Snyder, Tuesday, 28 July 2026.
Energy Vault has broken ground on a Texas campus deploying modular data centers that can be set up in weeks rather than years, eventually scaling to 500 megawatts.

Groundbreaking on the Snyder AI Campus

Energy Vault, Inc. (NYSE: NRGV) has officially commenced construction on its pioneering “Snyder AI” infrastructure campus in Snyder, Texas, marking a significant milestone in the integration of grid-scale energy management and high-density computing [1][2][5]. The physical groundbreaking, which occurred in late July 2026, initiates the development of a highly specialized facility designed to host Crusoe Energy Systems’ “Crusoe Spark” modular data centers [1][5]. This project represents a rapid transition from a commercial agreement signed between Energy Vault and Crusoe earlier in 2026 to active construction on the ground [1][4]. The site, located at 9301 Camp Springs Rd in Hermleigh, Texas, is currently undergoing extensive site preparation, fiber installation, and the deployment of critical power-related infrastructure to support the initial phase of the buildout [2][5][7].

The Phase 1 Rollout and the Powered Shell Model

The initial phase of the Snyder AI campus is engineered to deliver 8 megawatts (MW) of powered shell capacity, with a target commercial operation date set for the first quarter of 2027 [1][2][5]. Under the structured commercial agreement, Energy Vault is purchasing 10 Crusoe Spark modular data centers directly from Crusoe [1][2][5]. These units will then be leased back to Crusoe, functioning as “powered shell” infrastructure that supports the Crusoe Cloud platform [1][5]. This collaborative framework allows Crusoe to leverage Energy Vault’s expertise in land acquisition, power delivery, and regulatory navigation, while enabling Crusoe Cloud customers to run high-performance artificial intelligence model training, fine-tuning, and inference workloads [1][2].

Accelerating Deployment Timelines

In the rapidly evolving AI landscape, traditional data center builds often face multi-year delays due to power procurement bottlenecks and complex grid interconnection queues [4][5][GPT]. The Snyder AI campus seeks to bypass these hurdles through modular design, deploying Crusoe Spark units that can be installed and commissioned in a matter of weeks rather than years [4][5]. By combining pre-engineered modular data centers with Energy Vault’s specialized power management capabilities—including N+1 electrical redundancy and integrated battery energy storage systems (BESS)—the project establishes a repeatable blueprint for rapid computing expansion [4][5][7]. Marco Terruzzin, Chief Commercial and Development Officer of Energy Vault, emphasized this execution speed, noting that the combination of land, power, redundancy, and controls is designed to help AI compute customers transition from concept to capacity far faster than traditional methodologies allow [1][2].

Strategic Vertical Integration and Scalability

The partnership highlights a growing trend of vertical integration within the technology and energy sectors [1][5]. Cully Cavness, President, Co-founder, and Chief Strategy Officer of Crusoe, described the Crusoe Spark program as a strategic capacity accelerant [1][7]. By pairing modular physical capacity with direct Crusoe Cloud offtake, the collaboration creates a highly scalable business model that can be replicated at other sites where power is readily available [1][2]. This approach mitigates the risk of power constraints and provides a predictable path for scaling high-performance compute clusters [1][2][GPT].

Scaling Up to Meet Exponential AI Demand

While the immediate focus remains on the initial 8 MW deployment, the Snyder AI campus has been master-planned to accommodate massive future scaling [1][2][5]. A planned Phase 2 expansion is designed to boost capacity to 25 MW, representing an incremental increase of 17 MW over the initial phase [1][2][5]. Over the longer term, the broader site is designed as a sustainable, modular platform capable of expanding up to a total capacity of 500 MW, which is 62.5 times the capacity of the first phase [1][2][5]. Energy Vault has indicated that the annual revenue generated per megawatt at this site is expected to align with its previously issued corporate financial guidance [4][6].

As Energy Vault executes its broader AI infrastructure strategy, the company is also positioning itself to navigate a complex regulatory and macroeconomic environment [1]. According to corporate filings, the long-term economics of such infrastructure projects are influenced by regulatory frameworks, including the Inflation Reduction Act of 2022 and potential local tax structures [1]. To support its transition into owned and operated projects and to mitigate external financial pressures, the company has outlined strategic goals to diversify its supply chain, aiming to lessen the impact of international tariffs and ensure a steady pipeline of critical electrical components for its expanding Texas footprint [1].

Key Takeaways for the Energy-AI Nexus

The groundbreaking at Snyder, Texas, serves as a concrete manifestation of the energy-AI nexus, where energy infrastructure providers are no longer passive utility partners but active developers of high-density computing real estate [1][GPT]. By combining grid-scale battery storage, robust electrical redundancy, and modular server shells, Energy Vault and Crusoe are addressing the critical bottleneck of power availability that threatens to slow down the AI revolution [1][5][7][GPT]. As commercial operations target a Q1 2027 launch, the success of this 8 MW starting phase will likely dictate how quickly the partners can scale toward the ultimate 500 MW potential, offering a valuable case study for the entire infrastructure sector [1][2][5].

Sources


AI infrastructure Energy storage