Snowflake is planning to acquire AI-based site reliability engineer (SRE) platform provider Observe to strengthen observability capabilities across its offerings and help enterprises with AIOps as they accelerate AI pilots into production.

“Longer term, Snowflake is positioning itself as infrastructure for AI at scale. As AI agents generate exponentially more data, vertically integrated data and observability platforms become essential to running production AI reliably and economically,” Carl Perry, head of analytics at Snowflake, told InfoWorld.

Explaining further, Perry said that issues and bugs with AI-driven applications are harder to diagnose than in traditional software, increasing the pressure on enterprises to identify and resolve problems quickly, and Snowflake wants to tackle this by Observe’s telemetry, log, and trace analytics with Snowflake’s AI and Data Cloud.

That should give enterprises a unified view of data pipelines, model behavior, and infrastructure health — areas that are often fragmented across tools as AI systems move from experimentation to production, Perry pointed out.

The combined capabilities would help enterprises detect performance regressions, data drift, and cost anomalies earlier, while giving SRE and data teams a shared operational layer to manage reliability and governance for AI-driven applications, Perry said.

Moor Insights and Strategy principal analyst Robert Kramer agrees with Perry on the need for observability capabilities for managing AI applications at scale, terming them as “strategic” for CIOs.

Leaders who don’t adopt these capabilities for their enterprises “would be left holding a very expensive bag of science projects” as the time for pilots is over and it’s high time for them to realize tangible value in their AI investments, said The Futurum Group’s practice leader for data, AI, and infrastructure Bradley Shimmin.

The analyst also noted that Snowflake’s acquisition of Observe could significantly ease pricing pressures for CIOs by introducing a more cost-efficient approach to observability and challenging players such as Splunk and Datadog.

Unlike traditional vendors such as Datadog and Splunk, which charge premium prices because they treat telemetry data (logs, metrics, traces) as specialized, proprietary data that requires their ecosystem for storage and analysis, Snowflake plans to treat telemetry as standard data within its Data Cloud, Shimmin said.

This shift not only reduces storage and processing costs but also simplifies integration with existing enterprise data strategies, Shimmin added.

Additionally, Kramer feels that Snowflake might be able to deliver more value to its customers if it can combine Observe’s capabilities with previous acquisitions, such as TruEra.

“If Snowflake can connect system observability from Observe with model monitoring from TruEra, it could enable unified visibility from pipeline to model to production infrastructure, expanding its platform capabilities,” Kramer said.

Observe currently offers three platforms — AI SRE, o11y.ai, and LLM Observability — with capabilities like log management, application performance monitoring, and infrastructure monitoring.

The startup, which was established in 2017 by Jacob Leverich, Jonathan Trevor, and Ang Li, rolled out its initial observability platform a year later, leveraging a centralized database on Snowflake.

That, according to Perry, will help Snowflake integrate Observe’s offerings relatively faster. Snowflake has not disclosed the financial terms of the acquisition, and it is subject to regulatory approvals.