Cloudera’s Abhas Ricky on Private AI, Sovereignty and the Real Economics of Enterprise GenAI
As enterprises rush to embrace generative AI, many are learning the hard way that enthusiasm alone doesn’t translate into measurable value. According to Abhas Ricky, Chief Strategy Officer at Cloudera, success begins with a clear business case.
“GenAI belongs to a different genre of use cases, but the thesis remains the same,” he says. Companies must first understand why a use case matters, whether they have the skills to execute it and how it fits into broader priorities. Only then should they begin plotting the economic realities behind it. That means factoring in compute, scalability, GPU requirements, cost per query and the hours required to operate models at scale.
This is where Cloudera makes its strongest argument: private AI. By allowing customers to train and run enterprise AI applications in public cloud, private cloud, edge environments or even desktops, without letting data leave their estate, organisations gain tighter control over costs. Abhas shares an example of a customer able to deploy 400 private AI models for the same price as 100 models in the public cloud. It’s a stark demonstration of how economics shift when infrastructure and data stay close together.
Bringing AI to the Data, Not the Other Way Around
Cloudera’s architectural philosophy flips a long-standing industry norm. Instead of transporting massive datasets into external AI environments, the company focuses on moving models to where the data already lives.
This is powered by containerised data services and its Anywhere Cloud strategy, which supports any query engine or preferred database. The aim is simple, flexibility without reengineering applications. The ability to move data and assets bi-directionally across environments without refactoring is especially vital for hybrid and regulated industries.
Abhas points to financial services, healthcare and life sciences as examples. With policies like IFRS, HIPAA and SOC-II dictating how data is handled, enterprises need AI systems that accommodate regulatory expectations, instead of challenging them. Cloudera’s Open Data Lakehouse and security controls are designed to meet that obligation, and the company says it will expand further into compliance-heavy sectors including public and defense.
The Rise of Shared AI Infrastructure
With GPU shortages shaping global AI strategy, organisations are rethinking how compute is sourced and shared. Abhas believes the trend is already underway.
Telcos, for example, are experimenting with offering GPUs as a service. He references discussions in Helsinki, where operators are exploring whether unused Radio Access Network capacity could support customer inferencing. The emergence of NeoCloud reinforces the same direction, shared AI infrastructure that reduces the burden on individual enterprises.
National AI Sovereignty Becomes a Priority
As more countries define data residency and AI sovereignty policies, Cloudera is positioning itself as an enabler rather than an observer. The company recently partnered with AWS in Europe to support sovereign cloud initiatives, and Abhas notes ongoing collaborations in India and Singapore, particularly within GovCloud environments.
He credits Cloudera’s hybrid and containerised architecture for this momentum. The ability to keep workloads within national borders while maintaining scalability, governance and security gives governments and public entities more confidence to pursue AI transformation at their own pace.
The Hard Part of Private AI: Building the Foundation
For organisations planning to deploy private AI behind firewalls or on-premises, Abhas says three challenges consistently appear.
First, enterprises need a knowledge hub capable of mapping relationships between data assets. Without shared ontology, AI adoption remains fragmented. Second, they must rethink workflows for AI rather than simply digitise old processes. Cloudera’s AMPs, inferencing service with NVIDIA and low-code tools aim to support that shift. Third, strong governance and guardrails are non-negotiable. Whether through partners like Galileo or other platforms, metadata governance remains essential for building an enterprise intelligence centre, which Abhas describes as the foundation of private and sovereign AI environments.
“That’s the hard piece,” he says, “but it’s exactly what we help large customers do.”
The Workforce Will Change Rapidly
For Abhas, AI transformation isn’t only about infrastructure. It’s about talent. He expects entirely new or evolved roles to emerge across data engineering, model operations, streaming, security and AI workflow design.
He compares it to the rise of Web2 or digital transformation roles a decade ago. As tools like Cursor and Cloud Code simplify previously complex tasks, skill sets will expand, shift and multiply. Robotics and automation will accelerate the trend even further, widening the market for data practitioners and AI-led engineering careers.
The Next Horizon
Cloudera’s enterprise AI outlook is clear, keep data where it belongs, reduce model deployment costs, respect regulatory boundaries and enable hybrid flexibility. By focusing on private AI economics, sovereign infrastructure and architecture built for mobility rather than migration, the company believes it is positioned for the next wave of enterprise adoption.
For organisations trying to balance innovation with responsibility, Abhas’s message is pragmatic. AI is not a race to deploy faster. It is a discipline built on business value, infrastructure choice, data governance and long-term sustainability.
