We raised $35 million. Here is why, and what it means.

Today we announced our Series A. The round is led by Singular, with General Catalyst also participating, both of whom backed our $10 million seed round in December 2025. They are joined by new investors DST Global and Quantumlight, with Picus and Databricks also participating. We want to explain what we are building, why AI-Native Resilient Observability is the category that makes it matter, and why we raised, in plain language.

The problem

Observability has been running the same playbook for decades. Vendors ingest your telemetry, store it in their cloud, and charge more as your infrastructure grows. The AI era has ended that tolerance. Every agent loop, every autonomous deployment, every token interaction generates data at volumes legacy platforms were never designed to handle. Ingestion costs compound. Governance gaps open as sensitive AI interaction data flows to third-party clouds that were never designed to hold it. The platforms that dominated the last era are not struggling because of a feature gap. They are struggling because their entire architecture was built on an assumption the AI era has made obsolete. That is the gap AI-Native Resilient Observability was built to close.

Most vendors are responding by adding AI features to architectures built for human workflows. That is the wrong answer to the right question.

What we built

Traditional SaaS is dead. Long live software and a service. The SaaS model promised simplicity but somewhere along the way vendors lost sight of what actually matters: delivering sustained, compounding value to the customer. Pricing became a penalty on growth. Infrastructure costs became a hidden tax. And the further the data moved from the customer's environment, the further the vendor's interests moved from the customer's outcomes. Tsuga is built on a different belief. We deploy inside your cloud, so there is no infrastructure tax and no data ever leaves your control. Our forward-deployed engineers work alongside your team continuously, tuning your observability environment, reducing noise, and identifying every opportunity to drive down the volume of data processed and retained. And because your telemetry stays with you, the AI tools we provide work for you: running on your data, inside your perimeter, focused entirely on delivering the value you need, not on maximising what flows through our platform.

Because telemetry never leaves the customer's environment, the AI built on top of it never has to either. Automated root cause analysis runs on complete, unsampled data. The Tsuga MCP server and CLI give engineering teams the building blocks to construct their own agents on that foundation, inside their own security boundary, without sending data anywhere it should not go. Your teams build the agents. We make sure they work with SRE-grade intelligence and stay under control.

Why these investors

Singular and General Catalyst backed us at seed and are back again. When investors who know the company from the inside choose to lead a Series A, that is a different signal from new money coming in cold. They have seen the architecture, the customers, the commercial model, and the team under pressure. Their decision to return says more than we could say ourselves. DST Global and Quantumlight joining them, alongside Picus and Databricks, reflects a view we hear increasingly from the market: the shift to sovereign, AI-native infrastructure is structural, not cyclical, and the window for defining the category is now.

The Databricks investment is a statement about where the market is going. We are a Databricks partner, and the integration between our platforms means customers can route their Tsuga observability data directly into Databricks, opening up capabilities that neither platform delivers alone. It reflects a shared conviction: that the data layer and the observability layer belong together, inside the customer's own environment, and that the intelligence built on top of them should run there too.

Six months in

We launched in December 2025. Six months later we have millions in ARR already contracted, an average contract value in the six figures, and customers that include Le Monde, Camunda, Buk, and Black Forest Labs. Le Monde used Tsuga to monitor its infrastructure through the French municipal elections. Camunda and Buk are running the platform across multi-cloud environments with strict data residency requirements. Black Forest Labs is using it to observe AI workloads at the frontier of what the technology can do.

What comes next

The raise will enable us to grow the team efficiently and accelerate the rollout of the platform. We are investing in forward-deployed engineering capacity, the Skills library, the MCP server, and the agent-building toolchain that turns complete, resident telemetry into genuine AI capability for our customers. We are building the defining platform of AI-Native Resilient Observability, and every hire, every deployment, and every engineering decision we make is in service of that. Everything we build is in service of one thing: giving the organizations that put their trust in us more value, more control, and more confidence in the infrastructure they depend on.

Gabriel-James Safar, Co-Founder and CEO, Tsuga

Paris, 23 June 2026