NVIDIA (NVDA) shares rose 2.83% to $212.48 after Corvex announced a multi-year Blackwell GPU infrastructure agreement with a leading AI company. The contract expands an existing customer relationship and adds storage, processors, networking, and dedicated computing capacity. Corvex also completed the liquid-cooled deployment much faster than standard data center conversion projects.

Corvex Expands Blackwell GPU Capacity

Corvex, which trades on Nasdaq under the ticker MOVE, will supply NVIDIA Blackwell GPU clusters under the expanded agreement. Blackwell is NVIDIA's current-generation architecture for high-performance AI and accelerated computing, succeeding the widely deployed Hopper platform. The infrastructure links NVIDIA HGX B200 systems through Quantum-2 InfiniBand networking, enabling fast communication between computing units. The deal also includes dedicated high-speed storage and central processing units designed for large-scale workloads.

Corvex delivered the first portion of the cluster during the first quarter of 2026. The company completed the remaining capacity during the second and third quarters, extending the customer's available computing resources. Revenue was recognized as each portion entered service under the expanded customer commitment.

Full run-rate revenue is expected to begin midway through the current quarter, when the complete cluster supports customer operations. The longer-term agreement provides Corvex with a steadier revenue base and strengthens its position in GPU infrastructure services. For NVIDIA, it represents another commercial deployment of Blackwell processors paired with Quantum-2 networking, adding to a growing list of partners building around the architecture as enterprises and cloud providers race to scale AI training and inference capacity.

Rapid Deployment Cuts Data Center Conversion Time

Corvex installed liquid-cooled HGX B200 systems inside a facility that previously relied on air cooling. The company finished installation and commissioning approximately two weeks after the equipment arrived on site. Notably, Corvex completed the project without rebuilding the facility, relocating equipment, or waiting for new data center construction.

Liquid-cooling conversions often take several quarters because operators typically need to upgrade power systems, cooling equipment, and building infrastructure. Liquid cooling has become increasingly important for next-generation GPU clusters, which draw substantially more power per rack than earlier systems and generate heat loads that conventional air cooling struggles to manage efficiently. Corvex instead adapted the operating site quickly, bringing dense Blackwell computing capacity online in a fraction of the usual timeframe. The shorter schedule allows customers to access advanced computing resources without lengthy construction work or facility migrations.

The deployment highlights Corvex's engineering model, which focuses on speed, density, and commercially competitive infrastructure. This approach directly addresses the delays facing companies that require immediate access to modern computing systems. Faster delivery also enables Corvex to recognize revenue sooner after receiving equipment and completing customer requirements.

Funding Structure Supports Capacity Expansion

Corvex financed the expansion through a combination of debt, customer prepayments, and cash already available on its balance sheet. The company avoided issuing new shares, thereby preserving existing ownership levels during the capacity buildout. Customer funding also reduced Corvex's immediate cash burden while supporting the purchase and installation of equipment.

The financing mix does, however, increase the importance of stable payments, strong utilization, and disciplined debt management. Corvex must keep the cluster operating efficiently to support cash generation and meet its financing obligations. The multi-year customer commitment provides operating visibility as the company scales its infrastructure footprint.

For NVIDIA, the agreement adds another channel for Blackwell adoption as specialized infrastructure providers expand high-density computing services. The deal also reflects a broader industry pattern in which AI infrastructure providers are securing multi-year, pre-funded commitments from large customers to finance GPU deployments. Corvex gains contracted capacity revenue along with a deployment example that could support future customer discussions. Taken together, the agreement and its rapid rollout underscore continued demand for Blackwell systems across large-scale computing projects.