DigitalOcean has announced the acquisition of Katanemo Labs, Inc., a move that significantly expands its “Agentic Inference Cloud” into the operational layer of AI agent systems. By integrating Katanemo’s open-source project, Plano, and its specialized “small action models,” DigitalOcean provides developers with a production-ready data plane that simplifies orchestration, safety, and observability. This acquisition also introduces novel signal-based observability research to help teams transition from experimental pilots to reliable, real-world deployments. “The agentic era demands more than GPU capacity-it requires a new class of infrastructure primitives,” said Vinay Kumar, Chief Product and Technology Officer of DigitalOcean.
Also Read: IBM and Arm Forge Partnership on Dual-Architecture AI Hardware
“Katanemo Labs has spent years building exactly that: an AI-native data plane and specialized models that make multi-agent systems reliable, observable, and fast to deploy. With Katanemo Labs’s suite of products and advanced research, we are accelerating the path from prototype to production by giving developers the predictability and performance they need to scale with confidence.” “We started Katanemo Labs with a simple belief: offer developers durable infrastructure primitives that help them build agents faster,” said Salman Paracha, co-founder and CEO of Katanemo Labs. “We built Plano to offload functions like orchestration, observability, and safety in a framework-agnostic way so that teams can ship faster and operate with confidence. Joining DigitalOcean allows us to bring that vision-and our work in agentic observability-to a much broader global community.”