Cloud and data services provider Cloudera has partnered with France-based AI lab Mistral to bring AI models to enterprise data in private and sovereign environments. The companies say the move will help organisations ensure their data is AI-ready, which is critical to successful deployment.
Cloudera’s hybrid data and AI platform will be combined with Mistral’s AI models, enabling organisations to build, customise and run AI applications using their own data.
The technology will support deployments across cloud, on-premises, edge, sovereign and air-gapped environments, with the emphasis on organisations retaining control over sensitive information.
The partnership will allow organisations to run inference privately, meaning the process through which a trained AI model uses data to perform a task takes place within the confines of a business.
Cloudera says a lack of AI-ready data remains a stumbling block for organisations looking to deploy and leverage AI.
According to the company’s Data Readiness Index 2026, 89% of EMEA IT leaders said they had complete visibility into where their data resides, while only 26% said that data was fully governed.
Warren Olivier, regional VP for Africa at Cloudera, says Gartner predicted in 2025 that this gap would cause 60% of corporate AI projects to fail by the end of 2026 because they lack AI-ready data.
Independent AI expert Johan Steyn, founder of AIforBusiness.net, agrees that failed AI initiatives are rarely model failures. "Most failed AI initiatives are not model failures but data failures – fragmented, poorly governed or trapped in silos," he says. "The real value in a partnership like this is not simply access to a frontier model, but the ability to bring that model to your own well-governed data without surrendering control of it."
Olivier says stalled pilots can prompt organisations to blame AI hype, switch models, spend more or add an agent without addressing underlying data problems.
“This is where tokenisation stops being a little billing footnote and becomes the whole story. Generative AI is metered by the token. You pay for the volume of text and content going in and coming out, not for whether the answer was worth anything. On a clean, governed foundation, that is affordable. On a messy one, it is a slow leak that turns into a flood scarily quickly,” Olivier adds.
According to Cloudera, fragmented data can force organisations to add more context to prompts, while untrusted data can push teams towards larger and more expensive models. Unreliable answers can lead to retries and additional metered calls. AI agents can increase costs further by reasoning, re-reading, calling tools and looping.
Olivier adds: “In the EMEA findings of our own index, 42% of leaders said complicated access requirements were their primary barrier to using the data they can see, while only about a third have fully integrated data sources across environments."
Abhas Ricky, chief business officer and GM for applied AI at Cloudera, says: “Enterprise AI is entering a new phase where organisations need more than access to powerful models. They need the freedom to unlock specialised intelligence using their data, on their terms.”
Cloudera and Mistral say enterprises in highly regulated and data-intensive industries face regulatory, security and operational challenges when moving sensitive or proprietary information to external AI services.
The partnership covers Mistral’s portfolio of AI models and tools for reasoning, chat, coding, document intelligence and voice, and is intended to address these challenges by enabling AI to run within the enterprise data perimeter.
Kamal Brar, SVP of partnerships and alliances at Mistral, says: “Our mission is to make frontier AI available to enterprises without requiring them to give up control over their data, infrastructure or intellectual property.”
Share
