Splunk Finds Most Teams Cannot Control AI Costs
Techstrong TV Interviews
•
18m
More than half of IT teams say they cannot accurately control their AI costs. Alan Shimel talks with Greg Leffler, director of developer evangelism at Splunk, about the 2026 Splunk State of Observability report and what it reveals about AI, predictability and the future of operations.
From the NOC to developer evangelism
Leffler started his career in the NOC at eBay Advertising, triaging alerts and calling engineers. Those early lessons about cost and reliability shape how he thinks about AI costs today. Later, he became one of LinkedIn's early SREs and rose to senior manager of SRE. After a year on LinkedIn's editorial team, he joined Splunk, where he now speaks and writes about observability from a decade of hands-on experience.
Observability and security come together
Splunk has long argued that observability and security teams should share tools and data. Leffler says AI is finally making that real. With Cisco Cloud Control, teams can ask what caused a problem and let AI search across security and observability data at once. As a result, they can quickly tell whether a login spike is an attack or simply a marketing campaign. That shared view also helps teams see where AI costs are coming from.
What the report found about AI costs
This year's survey of more than 2,000 practitioners and executives focuses on restoring predictability, since AI is not deterministic. Leffler notes that unpredictable usage makes AI costs hard to forecast, and many leaders feel the pressure. Over half of respondents cannot accurately control AI costs, and 54% lack clear business outcomes for their AI. In addition, 62% often use AI agents to find root causes, yet only 6% use AI to write new code all the time, while 26% never do. Leffler sees that gap as a sign that trust in AI is still being built. Teams want proof before they let agents change production code. Meanwhile, the OpenTelemetry project will celebrate its graduation at KubeCon in Salt Lake City, a milestone that gives teams an open standard for the data behind every decision about AI costs.
Getting a handle on AI costs is now an observability problem. Explore more DevOps and observability videos and the latest Techstrong TV interviews.
For more information please visit splunk.com/state-of-observability
Up Next in Techstrong TV Interviews
-
IT Is Becoming an HR Department for A...
NVIDIA CEO Jensen Huang once said the IT department will become the HR department for AI agents. Alan Shimel talks with John Capobianco, head of AI and developer relations at Itential, about what managing AI agents looks like in practice for network and infrastructure teams.
From the factory...
-
What's The One Piece Of Infra You'd N...
Where should organizations draw the line when giving AI control over critical infrastructure? Security leaders discuss why identity and access management, privileged credentials, key management and recovery systems require human oversight. Explore the risks of autonomous AI, the importance of pro...
-
AI Agents Aren't Rogue. They're Relen...
Doing Exactly What They're Told
AI agents are often described as going rogue. In reality, most are doing exactly what they were asked to do. They just pursue the goal more creatively than people would, and they never get tired.Bitdefender Director of Cybersecurity Services Nick Jackson examines...