Datadog Enhances GPU Monitoring Amid Rising AI Costs

Datadog has announced the addition of GPU monitoring to its observability stack, addressing the growing importance of GPU instances, which currently represent 14% of cloud computing costs in the AI sector. With worldwide spending on AI infrastructure reaching $89.9 billion in Q4 2025 – a 62% increase from the previous year – the focus on optimizing GPU usage is critical for organizations investing heavily in AI technologies.

The introduction of unified visibility for AI stacks aims to help businesses better manage their GPU-related workflows by linking fleet health, costs, and performance. However, the challenge persists as companies struggle to directly charge back GPU expenses across different business units. This development highlights a strategic shift towards efficiency in AI investments, while also revealing the potential increase in reliance on cloud-based AI solutions, which may limit national computing autonomy.