Grafana wins AI customers as observability shifts cloud

Grafana Labs said 7AI, TeamSystem and Zama are standardising on Grafana Cloud, reflecting demand from AI and technology businesses for managed observability tools.

The three customers are using the cloud platform in different ways as they expand infrastructure linked to AI models, data systems and performance testing. Grafana Labs also pointed to existing use by Anthropic, Lovable and Harrison.ai as part of a broader shift toward hosted observability services.

Its latest observability survey found that operational complexity and overhead have become the main challenge for users. Half of organisations said they use managed observability solutions in some form, up from 43% a year earlier, while exclusive use of software-as-a-service tools rose from 10% in 2024 to 17% in 2026.

The findings suggest companies are reassessing the cost and effort of running self-hosted monitoring systems as AI workloads generate more telemetry data. According to Grafana Labs, that pressure is pushing teams away from fragmented open-source deployments and toward centralised managed services.

Cost focus

TeamSystem is using Grafana Cloud's Adaptive Telemetry suite to reduce unnecessary data ingestion and tie observability spending more closely to operational value. The approach is intended to preserve visibility into important systems while lowering the volume of lower-priority data collected.

That focus comes as observability budgets remain under scrutiny. The survey found that 90% of respondents expect spending on observability to stay the same or rise in the coming year, while 65% said cost is now a key factor in choosing tools.

For companies operating AI systems at scale, the issue is not only total spend but how that spend is distributed across logs, metrics and traces. Vendors offering managed services argue that filtering and prioritising telemetry can reduce waste without removing useful operational signals.

Benchmarking work

Zama is using Grafana Cloud as the main data source for its automated performance testing suites. The company said this allows engineers to track regressions in real time across 54 fully homomorphic encryption operations and blockchain transaction types.

Grafana Labs said Zama correlates telemetry with algorithm changes and hardware trade-offs as it works on implementation efficiency. It added that Zama uses Grafana Assistant to support investigations and reduce the time needed to resolve issues across engineering teams.

The use case highlights a growing overlap between observability and software development workflows, particularly in businesses where system performance is tied directly to product design. In those cases, monitoring data is being used not only to detect faults in production systems but also to measure the effects of code and infrastructure changes during testing.

Open-source shift

7AI is moving from self-managed open-source observability tools to Grafana Cloud. Grafana Labs said the migration is intended to reduce the engineering effort spent maintaining bespoke internal systems.

That trade-off has become more visible as AI companies add infrastructure at speed. Teams that once built their own observability stacks to retain flexibility are now weighing the staffing cost of maintaining them against the appeal of outsourcing that work to cloud providers.

Grafana Labs also cited Harrison.ai as an example of how AI-focused companies are approaching the issue. The company uses Grafana Cloud to monitor systems running AI in production.

"Running AI in production means there's no margin for blind spots," said Graham Bucknell, Platforms Engineering Manager, Harrison.ai. "Grafana Cloud gives us end-to-end visibility across our systems, so our teams can understand what's happening, act quickly, and scale safely. That transparency lets us move fast without compromising reliability or trust."

Anthony Woods, Co-founder, Grafana Labs, commented on the customer wins and the wider market shift. "AI companies need observability that keeps pace with them," said Woods. "These organizations are running some of the most complex and fast-moving infrastructure in the world, and they've chosen Grafana Cloud because it gives them a single platform that unifies their telemetry, controls their costs, and puts AI to work in the places that matter, from root cause analysis to onboarding new engineers. This is what the next generation of observability looks like."