Corvex deploys NVIDIA H200 GPUs to support production AI
Corvex has signed a long-term GPU lease to deploy a dedicated cluster of NVIDIA H200 accelerators for an AI-driven provider of high-performance battery technologies. The deal supports production AI workloads as well as core R&D tied to proprietary algorithms.
The news highlights how specialized AI infrastructure providers are competing with hyperscalers on cost, security, and flexibility. It also shows how GPU-heavy industries such as advanced energy storage are pushing AI into production under strict IP and data-sovereignty constraints.
H200 GPUs for production AI
Under the agreement, Corvex will deliver a dedicated NVIDIA H200 GPU cluster to underpin the customer’s AI development and research initiatives. The battery technology company selected Corvex for its overall price-performance, confidential AI enablement, and hyperscaler-class operations without the typical cost or lock-in.
Corvex said its GPU cluster architecture maximizes compute density while maintaining elasticity, allowing customers to burst capacity during peak demand. This approach targets continuous training and inference workloads where utilization and cost-per-compute are critical.
“This deployment reflects how leading AI innovators are scaling production AI without compromising economics, market access, or operational velocity,” said Jay Crystal, Co-Chief Executive Officer of Corvex. “Corvex delivered leading cost-per-compute for continuous workloads, hyperscaler-class managed operations, and a custom-designed and confidential on-premise GPU solution that enables our customer to expand into the most security-conscious environments.”
Security-first infrastructure without hyperscalers
Beyond cloud-based GPU clusters, Corvex is enabling secure, managed on-premise deployments designed for highly regulated environments. The architecture includes hardware-enforced encryption, remote attestation, and payload-free telemetry to protect sensitive training data and models.
Corvex is also supplying a fully managed Kubernetes platform designed to automate orchestration and boost GPU utilization. The goal is to reduce idle capacity and infrastructure overhead while keeping the ability to scale for large training runs.
“We believe this transaction is reflective of the growing demand from AI model builders and sophisticated enterprises for secure GPU infrastructure that is easy to use and more cost-effective than what has been offered to date,” said Seth Demsey, Co-Chief Executive Officer of Corvex.“Corvex is quickly becoming a go-to platform for scaling secure AI workloads at enterprise scale.”
The deal underscores Corvex’s growth strategy: delivering dedicated, secure AI infrastructure that meets enterprise governance requirements while keeping costs below traditional hyperscaler offerings.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
