Business
Corvex Among the First Companies to Achieve Verified Production Deployment of Confidential Computing for AI on NVIDIA HGX™ B200 Systems
Corvex, Inc., an AI cloud computing company specializing in GPU-accelerated infrastructure for AI workloads, today announced it has completed and verified the production deployment of confidential computing on NVIDIA HGX B200 systems.
About this update from Corvex, Inc.
[{"type":"text","content":"Encrypted NVIDIA NVSwitch and NVIDIA NVLink fabric protect data in use at runtime, with CPU and GPU remote attestation","length":123,"tagName":"p"},{"type":"text","content":"Near native performance enables secure AI without compromising production scale","length":79,"tagName":"p"},{"type":"text","content":"ARLINGTON, Va., March 3, 2026 /PRNewswire/ -- Corvex, Inc., an AI cloud computing company specializing in GPU-accelerated infrastructure for AI workloads, today announced it has completed and verified the production deployment of confidential computing on NVIDIA HGX B200 systems.","length":285,"tagName":"p"},{"type":"text","content":"The deployment validates encrypted GPU-to-GPU communication across NVIDIA NVSwitch and NVIDIA NVLink fabrics utilizing Intel Trust Domain Extensions and NVIDIA Confidential Computing to provide end-to-end AI security inclusive of CPU and GPU remote attestation using Intel® Trust Authority (ITA). Together, these capabilities establish confidential computing as an operational, production-ready infrastructure primitive for AI workloads on the NVIDIA HGX B200 system.","length":492,"tagName":"p"},{"type":"text","content":"As AI systems move from experimentation to mission-critical production infrastructure, customers increasingly require runtime assurances for workloads where sensitive data and models must be decrypted in memory to execute. Confidential computing addresses "data in use" risk by combining hardware enforced isolation with cryptographic attestation that provides verifiable evidence of platform integrity while workloads are running.","length":441,"tagName":"p"},{"type":"text","content":"Remote attestation delivers cryptographic proof that the underlying hardware, firmware, drivers, and system configuration remain in a known and uncompromised state at runtime. This verification is foundational for highly regulated industries and protecting valuable IP in AI workloads like model weights, enabling continuous compliance, defensible auditability, and deployment models where trust must be proven rather than assumed.","length":431,"tagName":"p"},{"type":"text","content":""In production AI, security is only trustworthy if it can be independently verified," said Seth Demsey, co-CEO at Corvex. "Confidential computing makes trust at runtime ...