Corvex: NVIDIA Blackwell GPU Deal Ignites an AI Infrastructure Micro-Cap

Corvex is an Arlington, Virginia-based AI cloud computing company specializing in GPU-accelerated infrastructure for AI workloads. The company provides GPU clusters, confidential computing, token factory services, and inference-as-a-service, purpose-built for AI model builders, enterprises, and federal/sovereantic clients.

The core business model is leasing GPU compute capacity — primarily NVIDIA HGX B200 (Blackwell) systems connected with NVIDIA Quantum-2 InfiniBand networking — to AI companies on a multi-year contract basis. Revenue is recognized as clusters are deployed and delivered. The company differentiates through rapid deployment: it installed and commissioned high-density, liquid-cooled B200 capacity in approximately two weeks after equipment arrival, avoiding the facility rebuilds that typically delay AI data center deployments.

Corvex generates revenue through infrastructure-as-a-service contracts, with customers paying for dedicated GPU clusters plus associated storage and CPU resources. The business requires significant upfront capital for GPU acquisition but generates recurring revenue over multi-year contract terms. Expansion is funded through a combination of debt financing, customer pre-payments, and existing cash — the Blackwell deal was specifically structured without new equity issuance.

Reported Results (Pre-Merger / Stub Period)

Q1 2026 Results (Post-Merger)

Note: Reported Q1 2026 includes only 12 days of Corvex OpCo operations post-merger close (March 19, 2026). Pro forma combines full-quarter Corvex OpCo results.

Balance Sheet

Key Valuation Metrics

Guidance: No formal forward guidance issued. Management stated full run-rate revenue from the Blackwell deployment begins mid-Q3 2026.

Corvex competes in the AI GPU infrastructure market — one of the fastest-growing and most capital-intensive segments in technology. The market is dominated by well-funded private and public companies with access to billions in capital.

• CoreWeave (Private): The leading AI cloud provider, valued at $80B+ with reported revenue exceeding $2B. Has raised billions in debt secured by GPU collateral. Sets the pricing and capability benchmark for the industry.

• Nebius Group (NBIS): Publicly traded AI cloud infrastructure company with significant GPU capacity. Better capitalized than Corvex with a broader European and U.S. footprint.

• Applied Digital (APLD): AI data center operator focused on HPC and GPU cloud services. Larger scale with institutional backing but different deployment model focused on owned real estate.

• Crusoe Energy (Private): Low-cost AI infrastructure powered by stranded natural gas. Has raised over $1B for GPU deployments. Competes on cost efficiency for compute-intensive workloads.

• Core Scientific (CORZ): Bitcoin miner pivoting to AI/HPC infrastructure. Has existing data center capacity being retrofitted for GPU cloud services.

Competitive moats for Corvex include: (1) rapid deployment capability — 2-week commissioning of liquid-cooled B200s in existing facilities versus industry norms of 6-12 months for new builds, (2) the Amplified AI Cloud platform's focus on confidential computing and security for sensitive AI workloads, and (3) NVIDIA partnership validated by the Blackwell deployment. However, the company's capital constraints ($68M total funding versus billions at competitors) limit the number of simultaneous deployments and create a scale disadvantage.

1. Q2 2026 earnings (expected August 2026): The first full quarter of combined Corvex operations. Investors will scrutinize revenue ramp, cash burn, and any quantitative details on the NVIDIA Blackwell deal economics.

2. Full run-rate revenue commencement (mid-Q3 2026): Management stated full run-rate revenue from the Blackwell deployment begins midway through the current quarter. When this materializes in reported financials, it provides the first real data point on deal economics.

3. Series D Preferred conversion dynamics: The pace and price at which Series D Preferred holders convert and sell will determine the effective share count and dilution trajectory. An orderly conversion supports price stability; a rush for the exits crushes it.

4. Additional customer contracts: The Blackwell deal validates the platform. Each new named customer contract de-risks the model and supports a higher revenue multiple. Watch for press releases announcing new GPU cluster deployments.

5. Capital raising and debt financing terms: Additional GPU deployments require capital. The terms of future debt or equity raises — particularly whether the company can continue avoiding equity issuance — will signal market confidence and determine the dilution path.