Datadog teams with Google Cloud for GPUs
- Datadog said April 22 its GPU Monitoring product is now generally available, alongside a Google Cloud push to monitor AI agents, models, and accelerators together. - Datadog says GPU instances now make up 14% of compute costs, and its Google Cloud tooling auto-instruments Agent Development Kit agents. - The launch extends Datadog’s AI observability push at Google Cloud Next after new partner-award wins. (cloud.google.com)
Datadog is tying together AI agent tracing and GPU monitoring for companies building on Google Cloud, with new product availability announced on April 22. (datadoghq.com) (markets.businessinsider.com) The company said its GPU Monitoring product is now available to all customers. Datadog framed it as a way to link GPU fleet health, cost, and performance directly to the workloads and teams using them. (markets.businessinsider.com) (datadoghq.com) On Google Cloud, Datadog said teams can use the same platform to watch AI applications and agents, including prompts, tool calls, step-by-step latency, token usage, and costs. It said its LLM Observability product supports auto-instrumentation for agents built with Google Cloud’s Agent Development Kit on the Gemini Enterprise Agent Platform. (datadoghq.com) A graphics processing unit, or GPU, is the expensive chip that trains and runs many artificial intelligence models. Datadog said those instances now account for 14% of compute costs, which it cited as a budgeting and capacity-planning problem for customers. (markets.businessinsider.com) Datadog’s pitch is that companies should not have to use one tool for model outputs, another for cloud infrastructure, and another for cost tracking. Its setup documents say customers can add cloud cost and instance-type information by enabling integrations including Google Cloud. (datadoghq.com) (cloud.google.com) The Google Cloud angle is not new, but the April 2026 announcement sharpens it around AI operations. Google Cloud’s Ritika Suri said Datadog’s observability combined with Google Cloud’s Agent Development Kit lets teams move faster on AI without giving up reliability or security. (datadoghq.com) The backdrop is Google Cloud Next, where Datadog said it won two 2026 Google Cloud Partner of the Year awards: AIOps Technology and Infrastructure Modernization Marketplace. Google announced those awards on April 21. (datadoghq.com) (cloud.google.com) The immediate change for customers is practical: one dashboard is becoming a place to inspect an agent’s prompt path and the GPU machines underneath it. Datadog is betting that as AI spending shifts from experiments to production, that view will sell. (datadoghq.com) (markets.businessinsider.com)
