Corvex Lands $103 Million Nvidia Blackwell Deal, Lifting AI Capacity Outlook

Corvex secured a three-year, take-or-pay agreement worth $103 million to supply Nvidia Blackwell computing capacity. The deal materially expands contracted revenue at its Mid-Atlantic 1 data center and reinforces its AI infrastructure strategy.

Corvex has signed a three-year, $103 million take-or-pay agreement to provide Nvidia Blackwell computing capacity to a large-scale AI inference platform customer, a deal that marks a significant step-up in the company’s contracted revenue base.

The agreement is expected to more than double annual recurring revenue tied to the Mid-Atlantic 1 facility, rising from $22 million to about $56 million once the deployment is fully operational. Investors focused on AI infrastructure are likely to view that jump as the central takeaway.

Shares of Corvex, which trades on Nasdaq under the ticker MOVE, moved higher in premarket trading after the announcement, reflecting market interest in smaller public companies building out high-demand GPU infrastructure around Nvidia’s latest Blackwell architecture.

Key Facts

  • Corvex signed a three-year take-or-pay contract worth $103 million for Nvidia Blackwell computing capacity.
  • The deployment is expected to lift contracted annual recurring revenue at Mid-Atlantic 1 from $22 million to approximately $56 million.
  • Corvex plans to deliver more than 1,000 Nvidia HGX B200 systems connected with Nvidia Quantum-2 InfiniBand networking.
  • The project will use 1.6 megawatts of IT capacity and carries an expected equipment cost of about $78 million.
  • Funding includes a $26 million customer prepayment and $38 million in committed debt financing, covering more than 80% of the deployment cost.

Corvex Nvidia Blackwell Deal

The Corvex Nvidia Blackwell deal underscores how demand for AI inference infrastructure is moving beyond the largest hyperscalers and into a broader ecosystem of specialist compute providers. By securing a take-or-pay structure, Corvex has locked in revenue visibility that many early-stage AI infrastructure companies still lack. That matters because the sector is capital intensive, and predictable utilization is critical when funding expensive GPU clusters.

This marks the company’s third Nvidia Blackwell deployment at its Mid-Atlantic 1 site, strengthening its positioning in a market where access to next-generation GPUs, power capacity and networking can be more important than scale alone. The Blackwell platform is widely viewed as a major upgrade cycle for AI workloads, particularly for inference and model serving, where efficiency and throughput can directly influence customer economics.

The customer commitment also has broader implications for Corvex’s operating model. Management has emphasized an expansion strategy built around multi-year contracts, customer prepayments and non-recourse or project-aligned financing. That approach can help reduce dilution risk and limit pressure on the corporate balance sheet, which is especially relevant for a micro-cap company pursuing rapid infrastructure growth.

Corvex’s $103 million take-or-pay agreement turns AI demand into contracted cash flow, a key distinction in a sector where growth often arrives before revenue certainty.

Why the financing structure matters

The deployment is expected to cost roughly $78 million, including GPU servers, storage, networking equipment and spare components. Corvex said it has secured financing for more than 80% of that amount, notably through a $26 million customer prepayment and $38 million in committed debt financing from USD.AI. For investors, that is a meaningful detail because it suggests the company can add capacity without bearing the full upfront burden from internal resources.

The timing is also important. Revenue from the deployment is expected to begin by the end of the first quarter of 2027, indicating that the market should view the contract as a medium-term earnings catalyst rather than an immediate top-line event. Until then, execution milestones, equipment procurement, installation and customer ramp-up will remain critical watch-points.

Implications for Investors

For investors, the clearest positive is improved revenue visibility. A take-or-pay contract means the customer pays for reserved computing capacity whether or not actual utilization reaches expected levels. In practical terms, that can reduce one of the biggest risks in GPU infrastructure investing: the gap between installed capacity and monetized capacity. If Corvex executes on schedule, the increase from $22 million to $56 million in contracted annual recurring revenue at one facility would be a substantial re-rating factor for the business.

The second major point is capital efficiency. Corvex reported approximately $21.7 million in contracted annualized revenue on live computing capacity as of Nov. 12, and third-quarter revenue of $5 million, up 31% from $3.8 million in the second quarter of 2026. Against that backdrop, a $103 million contract is large relative to the company’s current revenue base. That creates upside if management can convert bookings into operations smoothly, but it also magnifies execution risk. Delays in deployment, vendor lead times, financing terms or customer onboarding could affect the timing of anticipated returns.

Investors should also keep an eye on capacity expansion beyond this single contract. Corvex has said it plans to grow critical IT capacity from roughly 1.5 MW to 8 MW across Mid-Atlantic 1 and Midwest-1 by the end of 2026. The company also holds a right of first refusal on an additional 12.5 MW at the Midwest site, a move that could push total capacity above 20 MW. If demand for AI inference remains robust, that pipeline could support further contract wins. If demand normalizes or pricing becomes more competitive, the market may scrutinize how fast those assets can be filled at attractive margins.

Corvex’s relatively small market capitalization adds another layer to the story. At about $174.9 million based on the prior Friday close, the company is operating in a niche where a single contract can materially alter growth expectations. That dynamic can create sharp upside when momentum is strong, but it can also produce elevated volatility around financing, customer concentration and deployment milestones.

Looking ahead, the market will be watching whether Corvex can translate Blackwell-related demand into repeatable, contracted growth across multiple sites. The $103 million agreement gives the company a stronger foundation, but long-term investor confidence will depend on disciplined expansion, timely execution and sustained demand for AI inference capacity.

Ultima Markets