Cloudera’s Quiet Comeback in the Age of Hybrid AI

When a company born in the open-source trenches of Apache Hadoop evolves into an AI platform, the change maps a decade of technological shifts. 

Cloudera’s evolution—from Hadoop to a lakehouse architecture and now to AI—reflects a deliberate pivot toward hybrid deployments that keep enterprise data close to compute.

In a conversation with AIM, Abhas Ricky, chief strategy officer at Cloudera, condensed the timeline and said, “We’ve come a long way from being a company that was powering open source analytics all the way to AI.” He framed the transformation as product plus capital. After going private in 2021, Cloudera re-engineered its stack and leaned into R&D around AI.

Ricky also cited platform adoption as proof of change: “Over 95% of our ARR is now tied to the new platforms, which is Cloudera Data Platform (CDP).”

Building the Hybrid Edge

Cloudera’s competitive edge lies in its hybrid approach. As Ricky put it, “The whole story of Cloudera’s differentiation has been, we’re the only game in town for hybrid.” 

For Cloudera, hybrid means a single practical outcome: application portability without costly refactoring. “Why? Because application refactoring, as you might know, is an expensive line item, and we help customers do that.”

The company describes ‘private AI’ as the capability to run AI where it’s needed. 

“Private AI means you can do AI wherever you want at the edge on your desktop, on public cloud or private cloud,” he added. 

That flexibility is the backbone of Cloudera’s pitch, which is to let customers host models or bring models to data without rewriting apps or moving sensitive datasets wholesale. 

When discussing relationships with hyperscaler, Mayank Baid, regional VP for India and South Asia, described them as “frenemies”. 

Furthermore, Ricky highlighted, “In the private cloud domain, there is no competition that we have.” 

Cloudera has deliberately built an ecosystem of partnerships and integrations to reinforce its platform. Recent moves include an inferencing service with NVIDIA, connectors for ServiceNow, as well as tooling and partners for retrieval-augmented generation (RAG), observability and tabular models.

Direct alternatives to Cloudera are few, but companies like Databricks, IBM and Snowflake, among others, can be considered to see if they meet their requirements.

For instance, Databricks’ Lakebase, an AI-native operational database built on Postgres and Neon technology, sits within its Lakehouse architecture. It unifies operational and analytical data, offering real-time access, autoscaling and governance via Unity Catalog. Designed to minimise vendor lock-in, it provides scalability, transparency, and speed for AI-driven enterprise applications.

Turning Data Into AI Workloads

Cloudera’s commercial model has shifted to monetise actual workload consumption. 

“Cloudera monetises through a subscription and consumption in public cloud,” Ricky explained. “Now we have a pricing structure called a CCU—Core Compute Hours.” 

Ricky explained why that linkage matters: “If you are going to drive enterprise AI workloads on the platform, there will be more usage of the platform and that will drive revenue growth to get through with that.”

The firm is approaching AI from two angles. The first is adding intelligence to the lakehouse. “AI into CDP is the AI-powered Lakehouse,” said Ricky, referring to features such as intelligent catalogues, co-pilots and NLP helpers. 

The second focuses on enabling an ecosystem of model and agent partners. Through “AI with CDP”, Cloudera is expanding with offerings like low-code studios for synthetic data and agents, RAG and fine-tuning studios, applied machine learning prototypes (AMP), a CML workbench and NVIDIA-backed inferencing.

Ricky was clear about where he believes the struggle will be won. “Now, we believe the world is moving to a point where inference is where the battle will be. Everyone wants to get to inference,” he added. “If we can help customers do inferencing for a large percentage of the use cases in a form factor of their choice, at a price point of their choice, that’s a significant advantage.”

He also referenced the scale of Cloudera’s data footprint as a moat: “We probably have more enterprise data and context than almost anyone else out there, with 27 exabytes of data under management.”

India: Growth and Selective Startup Engagement

Baid called India “one of the fastest growing regions globally”, highlighting government, banking and telecommunications as core verticals driving adoption.

“Most of the large enterprises with a huge volume of data and sensitive data, which comes under regulation, are using our platform,” he noted.

When it comes to startups, Baid said the company engages where scale and use case fit.

Ola and PhonePe are some notable customers using Cloudera’s lakehouse solution for various use cases.

Moreover, confirming the company’s investment in India, Baid said, “We are investing heavily in India, both in terms of engineering and sales team.”

Cloudera’s trajectory is less about refinement and more about strategic positioning—keeping enterprise data where it lives, integrating AI where it adds value and monetising the compute that follows. 

The post Cloudera’s Quiet Comeback in the Age of Hybrid AI appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source