Stealthium Joins the Vultr Cloud Alliance to Bring GPU Observability and Runtime Security to AI Infrastructure

Stealthium has joined the Vultr Cloud Alliance, expanding the ecosystem of technology partners available to organizations that build and run AI workloads on Vultr. Stealthium, a member of the NVIDIA Inception Program, is the runtime observability and security company for GPUs, making AI accelerated compute observable, secure, and controllable.

As organizations put more AI workloads into production, access to high-performance GPU infrastructure is only part of the equation. Teams also need to understand how those resources are being used, identify performance and health issues, and protect workloads at runtime.

Together, Vultr Cloud GPU and Stealthium combine high-performance AI compute with deeper GPU-level intelligence, helping organizations make better use of GPU resources, identify issues faster, strengthen runtime security, and operate AI infrastructure more effectively.

"By adding Stealthium to Vultr's composable infrastructure, organizations can gain deeper insight into GPU workloads, helping them optimize utilization, troubleshoot issues faster, and strengthen runtime security." Kevin Cochrane, CMO, Vultr.

Bringing visibility and control to GPU infrastructure

While GPU infrastructure adoption is rapidly growing and represents a significant investment, traditional infrastructure monitoring can leave teams with limited insight into what is actually happening inside the GPU. Security tooling has a similar blind spot: instrumentation built for the CPU is unable to detect and prevent attacks executing on the GPU.

Stealthium can be deployed by organizations running workloads on Vultr Cloud GPU to add GPU-level telemetry, workload correlation and context, health monitoring, and runtime threat detection. Infrastructure, platform, and security teams can now access the critical signal they require to maintain and maximise their AI accelerated compute environments.

For organizations running AI workloads on Vultr Cloud GPU, this can help them:

Strengthen runtime security by detecting unusual or potentially malicious activity within GPU workloads.
Troubleshoot faster with deeper insight into GPU health, workload behavior, and real-time activity.
Operate AI infrastructure more effectively at scale with a clearer understanding of workload activity, GPU health and utilization, and security across their environment.

Stealthium deploys as an optional capability, a single agent per host on Vultr Cloud GPU instances, with no changes to the Vultr platform. Teams can add it to the environments that need it, when they need it.

"Whether it's their first deployment or their fiftieth, Vultr customers can now add runtime protection and real-time observability to their Cloud GPU environments. When teams can see what is actually running on their GPUs, they can catch threats sooner, fix problems faster and get more from every GPU they pay for as they scale." Sherif Elnabawi, CBO & Co-Founder, Stealthium.

Meet Vultr and Stealthium in Singapore

The partnership is being announced today, September 30th, at Cloud & AI Infrastructure Singapore.

Attending the event? Visit the Vultr booth at #55-F25 to meet the teams and learn how Vultr Cloud GPU and Stealthium can help organizations run AI workloads with greater operational insight, security, and efficiency.