AI agent expansion drives demand for enterprise data infrastructure

Enterprise AI usage is shifting from simple query-response interactions toward task-oriented execution, driving significant demand for AI infrastructure. According to OpenAI, Codex accounted for 64% of total output tokens from Codex and ChatGPT combined as of June 2026. This shift is characterized by AI agents performing complex tasks such as file editing, information retrieval, and multi-step workflows. This trend is expanding beyond engineering teams into legal, sales, recruiting, and marketing departments.

As AI agents increase the volume of logs and event data, companies are re-evaluating their data infrastructure. ClickHouse has seen its Annual Recurring Revenue (ARR) grow to over $350 million, fueled by the need for real-time analysis of massive datasets. Notably, OpenAI has reportedly increased its use of ClickHouse tenfold over the past year, processing more than 30 petabytes of data daily. This has led to a shift in workloads away from traditional observability platforms like Datadog, as enterprises seek more cost-effective and high-performance solutions for the massive data volumes generated by AI operations.