Raju
Datla

Most founders get one shot at building something Cisco wants to buy. Raju Datla got two - eight years apart, from different angles, on different problems. That's either extraordinary luck or extraordinary pattern recognition. Everything about his career suggests the latter.

In 2003, Datla built Jahi Networks, a company that handled network management appliances - the unglamorous infrastructure plumbing that makes enterprise devices actually work. Cisco acquired it in late 2004. In 2009, he co-founded Cloupia with Raju Penmetsa, targeting the next-generation problem: unified cloud management for data centers. Cisco bought Cloupia in 2012 for $125 million. The technology didn't disappear into a drawer - it became the foundation of Cisco UCS Director, UCS Manager, and UCS Central. If you've managed a Cisco data center in the last decade, you've used Datla's work without knowing it.

After the Cloupia acquisition, Datla stayed at Cisco. He ran the UCS software division - the $3B business he had, in part, built - as Vice President of Cisco's Datacenter Group. That's a rare story: the founder who didn't just cash out but ran the enterprise, learned how big companies operate at scale, and then chose to leave anyway.

In 2015, he founded CloudFabrix Software in Pleasanton, California. The premise was specific: IT operations was drowning in fragmented data, siloed tools, and alert noise that no human team could actually process. His answer was not another monitoring dashboard. It was a platform that automated the flow of data itself - ingesting, enriching, routing, and acting on operational telemetry at machine speed. He called it a Robotic Data Automation Fabric.

When AI agents became the industry's most overused noun in 2024, most vendors rushed to show autonomous action. Datla's focus was different. His argument: an AI agent is only as good as the data it can see, the context it carries, and the controls placed on what it can do. Without those three things, you don't have enterprise AI - you have an expensive demo.

Fabrix.ai's platform centers on an Enterprise Knowledge Graph - a continuously updated map of IT entities, dependencies, and topology. When an agent fires a root cause analysis, it's not guessing from raw logs. It's navigating a live map of how everything connects. That's the difference between an agent that sounds confident and an agent that's actually right.

Datla has been consistent on the governance piece: every automated action should be auditable, every cost visible, every failure explainable. "Built-in observability that ensures that you can monitor what the agents are doing, where they're succeeding, where they're failing, how much are they costing." In an industry where 'AI' is marketing shorthand, that specificity reads as engineering credibility.

The rebranding from CloudFabrix to Fabrix.ai in February 2025 wasn't a pivot - it was the company catching up to where Datla had been pointing it for a decade. The same robotic data automation logic that made CloudFabrix useful for AIOps turned out to be exactly the infrastructure substrate that Agentic AI needs: real-time data ingestion, policy-aware orchestration, multi-LLM integration, and agent lifecycle management. The product didn't change. The world changed around it.

There's a thread through all three of Datla's companies - Jahi Networks, Cloupia, Fabrix.ai - that is easy to see in hindsight. Each one built infrastructure that made the next layer of enterprise technology manageable. Jahi made network devices manageable. Cloupia made cloud infrastructure manageable. Fabrix.ai makes the data generated by that infrastructure - and increasingly, the AI agents acting on it - manageable.

He doesn't chase trends so much as find the place where complexity is about to become unmanageable and build the layer that absorbs it. In 2003, that was network configuration at scale. In 2009, it was cloud orchestration. In 2015, it was data fabric for IT operations. In 2025, it's agent governance. The timing on each has been ahead of the market by enough years to be useful, not so far ahead as to be ignored.

The approach has attracted institutional attention. Fabrix.ai holds partnerships with Cisco, IBM, NVIDIA, and HPE - which is an unusual collection of heavyweights for a 120-person company. GigaOm ranked Fabrix.ai as a Top 3 Leader and Outperformer in their 2025 AIOps Radar. Gartner named the company in multiple Agentic AI reports for 2025-2026. Enterprise Management Associates recognized it as a Leader across AIOps, Observability, and automation solution categories.

At Cisco Live 2025, Fabrix.ai's multi-agent platform generated what the company described as "strong attendee engagement." Given the audience - Cisco's own practitioner community - that reception is a meaningful signal. These are people who spent years using the UCS infrastructure Datla helped build. They know what enterprise-grade actually means.