Grafana Labs tops 10,000 customers as AI demand rises

Grafana Labs has passed 10,000 customers worldwide and exceeded USD $600 million in annual recurring revenue, linking the growth to rising demand for observability tools as businesses deploy more artificial intelligence systems.

The New York-based software group said the milestone comes as customers expand their use of its platform rather than buy single products. The average number of products used per contracted customer rose from 2.3 to 4.4 over two years. Now, 87.8% of contracted customers use at least two products and 64.9% use four or more.

Use of its AI assistant has become one of the clearest signs of that shift. More than 18,000 organisations are now actively using Grafana Assistant across paid and free plans, and a majority of self-serve and contracted new customers used it during the first half of the year to help with onboarding.

Co-Founder Anthony Woods said many companies are now trying to solve two related problems at once.

"Every organization is at a different stage of the observability journey, but AI has added new obstacles to navigate that weren't there before: Are you using AI to observe your systems, and do you also need to observe the AI itself?" Woods said.

"The honest answer, for almost everyone now, is both, and that's what's driving the growth behind these milestones," he said.

Observability software helps engineering teams track the health and behaviour of applications, infrastructure and data systems. The category has taken on added importance as companies move AI models and agents into live environments, where problems such as model drift, unexpected actions and rising token costs can sit outside traditional monitoring methods.

Grafana Labs cited findings from its own survey showing that complexity and overhead remain the top concerns in observability, while 57% of organisations are implementing large language model observability in some form. That points to a market in which buyers want tools that can both monitor AI systems and use AI to interpret operational data.

Platform expansion

Over the past year, the company has widened its product set with a focus on those two use cases. During an internal AI-focused product week, it launched six AI-related products, including Grafana Assistant Investigations, which uses multiple AI agents to examine observability data, and Grafana Agent Observability, which is designed to monitor AI agents and models in production.

Other additions included a server and command-line tools intended to give AI agents and coding tools direct access to Grafana Cloud, along with a larger workspace for the assistant and a feature that lets teams save prompts and rerun them automatically.

Co-Founder and Chief Executive Officer Raj Dutt said the increase in product usage reflects how customers are approaching AI deployments.

"What stands out this year isn't just that customers are adopting AI, it's that they're going deeper with us once they do," Dutt said.

"The average customer now runs more than four of our products, up from just over two years ago, and that's not an accident. When teams bring AI into production, they don't want another point solution bolted on; they want a platform they can grow into. Crossing 10,000 customers tells us that bet is paying off," he said.

Cloud growth

Grafana Labs also reported a sharp rise in cloud usage. Monthly active Grafana Cloud users increased from about 127,000 to more than 251,000 over two years, while more than 1.1 million distinct users logged in during that period. Self-serve monthly active organisations rose from roughly 42,000 to 66,000, and there were more than 500,000 Grafana Cloud sign-ups over the same timeframe.

The figures point to a mix of enterprise expansion and product-led growth. A free tier has helped bring in smaller teams and individual users, while larger customers have added products as their observability needs broaden.

Cost pressure

One of the central issues for buyers is the volume of telemetry generated by AI and conventional software systems. Grafana Labs said its Adaptive Telemetry suite, which now covers metrics, logs, traces and profiles, has removed 28.5 billion metric series and reduced log volumes by 26 petabytes.

The suite cuts telemetry costs by 30% to 50% on average, according to the company. That matters for customers facing higher storage and analysis bills as AI workloads create more data.

Chief Technology Officer Tom Wilkie framed the issue in economic terms.

"The economics of observability break down when every new AI workload just adds more data to store and more noise to sort through," Wilkie said.

"Adaptive Telemetry exists so cost scales with the value of your data, not the volume of it. This matters more as AI systems generate telemetry at a pace no team can review manually," he said.

Customer example

Grafana Labs also pointed to customer use in high-pressure settings. Deutsche Telekom described using the platform during a failure in backup infrastructure shortly before a major live sports event.

"When our backup infrastructure failed three days before the World Cup opener, Grafana Cloud is what got us through it," said Neha Kumar, Consultancy Team Lead, Observability, and Felix Mohnke, Observability Lead, at Deutsche Telekom. "We built an AI-powered situation room that let our team see the whole picture and keep millions of viewers online with zero downtime. That's the kind of platform we've come to rely on. Grafana Assistant already cut our dashboard builds from four months to a single day, so by the time we actually needed it under pressure, we trusted it. That's the combination that matters: AI that makes you faster day to day, and observability solid enough to hold up when it counts most."

Alongside the customer and revenue milestones, Grafana Labs said it had introduced broader platform updates including Knowledge Graph, Bring Your Own Cloud, Federal Cloud, Grafana 13 and Mimir 3.0, and had again been named a Leader in Gartner's Magic Quadrant for Observability Platforms.