AINEWS
The 12 year old software monitoring vendor, now past $600M a year in subscription revenue, trades top line growth for usage based pricing as AI moves from pilot to operations.
Grafana Labs, the 12-year-old company that sells software-monitoring tools to enterprises, said this month it has crossed $600 million in annual recurring revenue, the standard SaaS top-line number, with more than 10,000 customers on its books. That is at least a 50% jump from the $400 million ARR and 7,000 customers the company reported in September 2025.
The company sells a translation layer between complex production systems and the engineers who have to keep them running. In plain terms: companies pay Grafana to watch their software and AI systems for problems, slowdowns, and unexpected agent behavior. As AI workloads ship out of R&D and into production, the bill for monitoring them is moving from project budgets to operations line items, where the spend does not go away.
"AI is a major factor" behind the acceleration, Grafana cofounder and CEO Raj Dutt told Business Insider. The company shipped six new AI features in July alone, including Agent Observability, used by hundreds of customers to track AI agents for speed and reliability. More than 18,000 organizations, including Alter Domus and Deutsche Telekom, now run Grafana Assistant, the conversational analytics product Dutt calls the fastest-growing in the company's history, with the majority of users on a paid tier.
The more interesting texture in the announcement is what the company is leaving on the table. Adaptive Telemetry, a feature that lets customers reduce the volume of data they pay Grafana to monitor, has cost the company close to $100 million in potential revenue, Dutt said. The product strips out redundant and low-value metrics, so a customer's bill tracks what they actually use rather than the raw data their systems emit.
"The revenue quality is better," Dutt told Business Insider. It is an unusual pitch for a SaaS executive to make on the record: most would rather hide the foregone dollars. Grafana is publishing them.
Datadog, the larger competitor in the same category, is hearing the same demand from customers even if it is not reporting the same numbers. "Customers increasingly deploy AI and use Datadog to monitor and secure it," Datadog cofounder and CEO Olivier Pomel said this month. Two CEOs in the same narrow category describing the same pull from buyers is the closest thing to independent corroboration the thesis gets without third-party spend data.
That shared signal is the structural read. If AI monitoring is genuinely a category-expanding line item rather than a one-off pilot cycle, the customers Grafana gives a discount to today are the customers whose AI bill will keep growing tomorrow, and the Adaptive Telemetry bet pays for itself. If AI monitoring is concentrated in a handful of large enterprises and the rest of the customer base is flat, the $100 million looks more like a discount the company wrote itself.
Watch item: Grafana's next funding round, which the September 2025 press release said would go toward global expansion, will price the trade in public. So will Datadog's next quarter, which will show whether the AI-monitoring demand Pomel described is broad-based or concentrated in a few large customers.