According to Sifted, two Datadog alumni have raised a $10 million seed round led by General Catalyst for their new AI observability startup called Tsuga. Gabriel-James Safar and Sébastien Deprez previously co-founded Madumbo, which Datadog acquired in 2018 and grew into an eight-figure revenue product before the company’s IPO. The Paris-based founders took a break in 2023—Safar for family and travel, Deprez for trades training—before deciding to tackle the observability market again. Their new startup has already onboarded scaleups and publicly listed companies in regulated industries like financial services. European VC Singular and angels including Alan cofounder Charles Gorintin also participated in the round.
Sponsored content — provided for informational and promotional purposes.
The Datadog connection
Here’s what makes this interesting: these aren’t just random entrepreneurs jumping on the AI bandwagon. Safar and Deprez actually built and scaled key products at Datadog itself. Their previous startup, Madumbo, was doing automated testing for web applications when Datadog acquired it back in 2018. At the time, Datadog was still pre-IPO. And get this—within a year of the acquisition, their product was generating eight-figure revenues. That’s serious scale. So when these guys say they understand the observability market, they’re not blowing smoke. They’ve literally been in the trenches building the very products they’re now planning to challenge.
The bring your own cloud approach
Now here’s where Tsuga gets clever. Traditional observability platforms like Datadog, Dynatrace, and Splunk require customers to send massive amounts of data to the vendor’s cloud. That creates two huge problems: cost and security. The more data you send, the more you pay—and for enterprises with massive data volumes, those costs become prohibitive. Plus, sending sensitive infrastructure data outside your own environment raises serious governance and security concerns, especially in regulated sectors like banking and healthcare.
Tsuga’s “bring your own cloud” approach flips this model entirely. Instead of pulling data out to their platform, they deploy observability clusters directly into the customer’s existing cloud environment. Basically, they’re bringing the observability to the data, not the other way around. Safar describes it as creating “a space where I’ll be able to manage your infrastructure as if it were mine, but it remains yours.” The vendor gets limited permissions to monitor and debug, but the data never leaves the customer’s control.
Why this matters now
Observability might sound like tech jargon, but it’s becoming increasingly critical as companies deploy more AI systems. When your revenue depends on software running smoothly, you need to know exactly what’s happening across your entire stack—from user interactions to server performance. And as AI workloads generate exponentially more data, the cost and complexity of traditional observability models become unsustainable.
Safar points out that “the model of traditional actors is based on the fact that data costs more.” But data volumes are exploding, and enterprises need more granular visibility, not less. So Tsuga is betting that their architecture will become increasingly attractive as AI adoption accelerates. They’re specifically targeting larger enterprises and regulated industries where both cost and data governance are major concerns.
The road ahead
With $10 million in fresh funding and a team of 25 people, Tsuga is doubling down on product development and commercialization. They’ve been working with design partners since earlier this year and have accelerated their go-to-market efforts in recent months. The big question is whether they can actually compete against the giants in this space—Datadog alone is worth over $50 billion, and they’re up against cloud behemoths like AWS too.
But here’s the thing: these founders have already proven they can build successful products in this exact market. They understand the pain points from the inside. And their “bring your own cloud” approach could be the wedge that opens up segments of the market that traditional vendors have struggled to penetrate. It’s definitely one to watch as AI observability becomes the next battleground in enterprise software.
