Corvex surges 67% on multi-year NVIDIA Blackwell GPU deal

Corvex signed a multi-year agreement to supply clusters of NVIDIA Blackwell GPUs, sending shares up 67% after the AI cloud provider commissioned liquid-cooled capacity in about two weeks inside an existing air-cooled facility.

"Standing up liquid-cooled Blackwell capacity inside an air-cooled facility in roughly two weeks is an engineering accomplishment," Seth Demsey, Co-Founder and Co-Chief Executive Officer at Corvex, said.

The deal expands an earlier customer commitment and includes NVIDIA Quantum-2 InfiniBand networking, dedicated high-speed storage, and CPUs. Corvex delivered the first portion of the cluster in the first quarter of 2026, with remaining capacity delivered in the second and third quarters. The expansion is funded through debt financing, customer pre-payment, and cash on hand — no new equity issued. Revenue is being recognized throughout 2026 as portions of the cluster are delivered, with full run-rate revenue starting midway through the current quarter.

The deployment speed matters because liquid-cooling retrofits of air-cooled data centers typically take multiple quarters to bring online. Corvex's roughly two-week commissioning window inside an existing facility — documented in a case study produced with Lenovo — could pressure AI cloud rivals including CoreWeave and Lambda Labs to accelerate their own deployment timelines as demand for Blackwell-class GPUs continues to outpace supply.

The agreement centers on NVIDIA's HGX B200 platform, the latest Blackwell-generation GPU architecture designed for AI training and inference workloads. NVIDIA Quantum-2 InfiniBand networking provides the high-bandwidth, low-latency fabric needed to scale GPU clusters across thousands of nodes. Corvex's product suite includes AI Factories and GPU Clusters, Confidential Computing, and the Corvex Token Factory, an inference platform currently in development.

The AI cloud market has become increasingly competitive as demand for GPU compute outpaces available supply. CoreWeave, one of the largest independent GPU cloud providers, has expanded aggressively with NVIDIA hardware, while Lambda Labs has focused on serving AI researchers and startups. Hyperscalers including Microsoft Azure, Amazon Web Services, and Google Cloud have also invested heavily in Blackwell capacity, creating a crowded field where deployment speed and financing flexibility are key differentiators.

Corvex's decision to fund the expansion through debt financing, customer pre-payment, and cash on hand — rather than issuing new equity — preserves shareholder value while enabling growth. This approach contrasts with some competitors that have relied on equity raises to fund GPU acquisitions, a strategy that dilutes existing shareholders.

The stock's 67% surge reflects investor enthusiasm for AI infrastructure plays with confirmed customer demand. Corvex shares, which had surged roughly 100% in pre-market trading before settling back, now trade at a significant premium as the market prices in multi-year revenue visibility from this agreement. The deal also provides NVIDIA with another channel for Blackwell GPU deployment, reinforcing the chipmaker's position in AI data center infrastructure as it competes with AMD's Instinct line and custom silicon from hyperscalers.

For investors, the key question is whether Corvex can sustain this momentum. The multi-year agreement provides revenue visibility through 2026 and beyond, but the company faces execution risk as it scales liquid-cooled capacity across additional facilities. The two-week deployment timeline, if replicable, could give Corvex a competitive advantage in winning new customers who need GPU capacity quickly.

This article is for informational purposes only and does not constitute investment advice.