The “Cheaper Datadog” Illusion in the AI Era

Many companies are transitioning from Datadog to alternative observability platforms to save costs, but these savings are often temporary and do not address long-term economic challenges as telemetry data grows exponentially, particularly with AI developments. Traditional volume-based pricing models can lead to increased costs as the amount of data ingested and retained grows, causing engineering teams to limit data collection to manage expenses. This approach is misaligned with the needs of AI-driven systems, which generate substantial telemetry data. The text advocates for a shift to a Bring Your Own Cloud (BYOC) model, which decouples costs from data volume by allowing companies to pay for actual infrastructure use instead of per-unit data fees. This model encourages deeper data collection without financial penalties, aligning better with AI and cloud-native architectures. The author suggests that while traditional SaaS models may still work for smaller organizations with predictable growth, companies expecting significant telemetry expansion should reconsider their pricing strategies to avoid stifling innovation.