Modular AI Data Centers: Energy Vault and Crusoe Spark Agree on 25 MW Framework

On February 11, 2026, Energy Vault and Crusoe announced a multi-year strategic framework agreement to deploy up to 25 MW of Crusoe Spark Modular AI Factory Units. The first site will be at Energy Vault’s powered shell facility in Snyder, Texas. This deal marks Energy Vault’s pivot into the AI infrastructure market and aims to accelerate Crusoe Cloud capacity while leveraging onsite solar and storage assets for high-density compute deployments.

Context: Energy Storage Meets AI Compute Demand

Since its founding in 2017, Energy Vault has focused on grid-scale energy storage solutions, serving utilities and large power consumers with short-term to ultra-long duration systems. Publicly traded on the NYSE under ticker NRGV, the company shifted in 2024 to an “Own & Operate” model, targeting recurring revenue streams. Meanwhile, Crusoe emerged during the AI boom of 2022–2025 as a vertically integrated provider of energy-optimized AI infrastructure. Crusoe’s modular approach addresses power constraints and interconnection delays by delivering pre-manufactured, plug-and-play AI data center units.

Rapid growth in AI training and inference since 2023 has strained existing data center capacity and utility grids. Hyperscale operators are racing to add high-density compute clusters, but long lead times for grid connections and rising energy costs have created bottlenecks. Against this backdrop, storage specialists like Energy Vault see an opportunity to repurpose their digital control platforms and renewable assets to support AI workloads.

Deal Details: Multi-Year Framework Agreement

Under the agreement, Energy Vault will deliver powered shell infrastructure—electrical distribution, cooling systems, and controls—for Crusoe’s Spark units. Deployments begin in 2026 at the Snyder, Texas technology center, which already integrates solar generation with Energy Vault’s storage systems. The modular design allows rapid plug-and-play installation of AI factory units, each delivering high-density compute capacity for Crusoe Cloud services.

The framework covers up to 25 MW of capacity and can scale repeatably across additional powered sites. Both companies emphasize fast deployment: pre-manufactured Spark units are trucked to site and connected to Energy Vault’s standardized shell, reducing setup time compared to conventional builds. Energy Vault’s Asset Vault Platform will manage these powered-and-operated facilities under its recurring-revenue model.

Analysis: Revenue Potential and Operational KPIs

From an investor’s angle, the key metric is revenue per megawatt. Energy Vault indicates this AI deal could unlock up to 20x the revenue-per-MW of its traditional battery energy storage systems (BESS), though this figure comes from company projections and isn’t independently verified. Even at the lower end of a 10x multiple, the margin profile on AI infrastructure could significantly boost EBITDA versus pure-storage projects.

Operational KPIs to watch include: deployment speed (days/weeks from delivery to live compute), capacity utilization (percentage of AI modules actively serving workloads), and reliability metrics tied to solar-plus-storage uptime. Energy Vault’s platform already monitors power flows and systems health across its storage fleet; extending this to powered shells means investors get visibility on AI data center performance too.

Crusoe’s Spark units support both inference and training workloads. For customers, that translates into a turnkey managed service with predictable SLAs. From a financial perspective, Crusoe Cloud subscription and usage fees offer recurring topline, while Energy Vault’s contracted power-infrastructure fees add predictable, long-term revenue streams.

Risks and Considerations

No deal is risk-free. Key hazards include execution delays—onsite civil works, permitting, and grid upgrades can still slow down even plug-and-play components. Although powered shells reduce lead times, interconnection agreements with utilities often stretch beyond initial estimates.

Another factor is AI demand volatility. While large language models and enterprise inference needs continue climbing, macroeconomic headwinds could temper corporate IT spending. If AI compute growth slows, Crusoe and Energy Vault may face under-utilized capacity or pricing pressure.

On the technical side, integrating high-density compute with on-site solar and storage requires close coordination. Energy Vault’s platform is proven in energy storage but less so in data center operations. Any hiccup in reliability could trigger SLA credits or reputational costs.

Implications for Investors

For retail and institutional investors, this partnership highlights a few actionable points:

• Revenue Diversification: Energy Vault’s move into AI infrastructure helps reduce its reliance on battery-only contracts and taps into a market with higher per-MW returns.
• EBITDA Upside: Even conservative assumptions on the 10x revenue multiple versus BESS could buoy margins. Watch quarterly deployments and contract commencements for real-time pickup in fee income.
• Deployment Metrics: Track the timeline from framework signing to first MW online. Delays can signal under-estimated grid or permitting challenges, while on-time roll-outs validate Energy Vault’s powered shell concept.
• Order Book Growth: Beyond Snyder, additional powered shell sites or similar agreements with other AI providers would reinforce the strategy and de-risk revenue forecasts.
• Partnership Execution: Monitoring joint operational updates—module deliveries, commissioning milestones, uptime statistics—will help investors gauge whether Energy Vault’s platform scales as promised.

Looking Ahead

By mid-2026, the first Crusoe Spark units should be online at Snyder, offering a real-world test of modular AI data center economics. Investors should watch reported utilization rates, service uptimes, and any commentary on margin accretion versus pure-storage projects. If Energy Vault can replicate Snyder’s setup across multiple sites and secure additional framework agreements, the company’s AI infrastructure venture could become a significant earnings driver in 2027 and beyond.

In short, the Energy Vault–Crusoe partnership is a clear example of how energy specialists are leveraging renewable assets and digital platforms to meet surging AI compute demand. For investors, the key will be converting framework potential into contracted, deployed capacity—and turning modular shells into a steady revenue stream.