Tsuga Raises $35 Million Series A as AI Pushes Observability Infrastructure to Its Limits

Paris-based observability startup Tsuga has secured $35 million in Series A funding, less than a year after emerging from stealth, as enterprises grapple with the growing complexity and cost of monitoring AI-powered systems.

The round was led by Singular, with participation from existing investors General Catalyst and new backers DST Global and QuantumLight, alongside support from Picus and Databricks Ventures. The financing comes during a period of rapid growth for the company, which says it has already signed millions of dollars in annual recurring revenue and secured customers including Le Monde, Camunda, Buk, and Black Forest Labs.

The new capital will be used to expand the team to approximately 100 employees and accelerate platform development.

Why Observability Is Facing a New Challenge

Observability platforms help organizations monitor applications, infrastructure, and increasingly AI systems by collecting telemetry data such as logs, metrics, and traces. As enterprises deploy more AI agents and autonomous systems, the amount of telemetry generated has grown dramatically.

According to Tsuga, traditional observability architectures were built around a model where customers send telemetry data to a vendor-operated cloud platform. While that approach worked when data volumes were smaller, AI workloads are producing significantly more telemetry, driving up ingestion costs and creating new governance challenges around sensitive data.

The company argues that many organizations are now being forced to choose between controlling costs through sampling, which reduces visibility, or accepting rapidly increasing observability bills.

A Different Approach: Bring Your Own Cloud

Founded in 2024 by former Datadog executives Gabriel-James Safar and Sébastien Deprez, Tsuga was built around a Bring Your Own Cloud (BYOC) architecture.

Rather than sending telemetry to a third-party service, the platform is deployed directly within a customer’s cloud environment. The company’s platform runs across major cloud providers including AWS, Microsoft Azure, Google Cloud, and sovereign cloud environments. Telemetry data remains inside the customer’s own infrastructure, stored in their object storage and encrypted using their own security controls.

This architecture is designed to address several challenges simultaneously: data sovereignty requirements, compliance concerns, infrastructure costs, and vendor lock-in. Tsuga supports open standards such as OpenTelemetry and uses open data formats, allowing customers to maintain control over their telemetry data.

Building Observability for AI Systems

The rise of AI applications is also changing what organizations need to monitor.

In addition to traditional logs, metrics, and application traces, enterprises increasingly need visibility into AI-specific workflows such as prompt execution, token usage, agent interactions, confidence scores, and multi-agent call chains. Tsuga’s platform combines these capabilities into a single environment while keeping data inside the customer’s cloud boundary.

The company has positioned itself as an “AI-Native Resilient Observability” platform, offering observability tools designed specifically for AI workloads. Features include AI-powered anomaly detection, automated root cause analysis, deployment monitoring, and agent observability capabilities.

Tsuga also provides an MCP server and command-line tools that allow engineering teams to build their own AI agents on top of observability data without moving that data outside their security perimeter.

A Services Layer Alongside the Software

Unlike many SaaS vendors, Tsuga combines software with a forward-deployed engineering model.

The company works directly with customers to optimize telemetry environments, reduce unnecessary data collection, and improve observability efficiency over time. This approach reflects a broader trend in enterprise software where vendors increasingly provide operational expertise alongside technology.

As organizations continue to expand AI deployments, managing the resulting flood of telemetry data is becoming an increasingly important operational challenge. Tsuga’s approach focuses not only on collecting and analyzing that data, but also on helping customers reduce the volume of information that needs to be processed and retained.

A more neutral version would be:

A Sign of Changing Priorities in Observability

The funding round comes at a time when infrastructure teams are reassessing how observability systems fit into increasingly AI-driven environments.

As organizations deploy more AI applications and autonomous agents, telemetry volumes continue to grow, placing additional pressure on architectures that were designed before these workloads became commonplace. At the same time, data governance, compliance, and cost management have become more prominent considerations for enterprise buyers.

Tsuga’s approach reflects one response to these challenges: keeping observability data within customer-controlled cloud environments rather than routing it through a vendor-operated platform. Whether this model gains broader adoption remains to be seen, but it highlights how the observability market is evolving as AI workloads become a larger part of enterprise infrastructure.

With its latest funding, Tsuga joins a number of startups exploring new approaches to monitoring and managing modern software systems. The company’s growth suggests that some organizations are actively evaluating alternatives to traditional observability architectures as they adapt to the operational demands of AI-native applications.