Mistral partners with Cloudera for sovereign enterprise AI

Mistral published a partnership with Cloudera to run Mistral models on Cloudera’s hybrid data platform — including private cloud, on-prem, and air-gapped — and to train custom models on proprietary data customers keep.

Mistral put open-weight models inside Cloudera’s hybrid and air-gapped data estate with an explicit ownership pitch — how a European lab sells to companies that will not ship data to a public API. The “30 exabytes” line stays the companies’ figure.

Mistral published “Cloudera and Mistral Partner to Bring Specialized, Sovereign Intelligence to Enterprise Data.” The company announced a strategic partnership with Cloudera. That news post is the primary.

The page lists two components. First, inference in the customer environment — running a trained model to answer questions. Mistral says its models will be integrated with Cloudera’s hybrid data platform so enterprises can deploy across private and public cloud, on-prem (on the customer’s own premises), and fully air-gapped environments — networks with no outside connection — while keeping control.

Second, custom model training: Mistral says enterprises can train models on large amounts of proprietary data in controlled environments, and keep ownership of both the data and the resulting intelligence.

Kamal Brar, Mistral’s SVP of Partnerships & Alliances, quoted on the post: “It’s a privilege to have the opportunity to bring Mistral’s sovereign AI to Cloudera’s 30 exabytes of customer-managed data running on its platform.” An exabyte is a billion gigabytes. The 30-exabyte figure is the companies’ claim — not an independent audit.

Abhas Ricky, Cloudera’s Chief Business Officer and GM of Applied AI, quoted on the same page: every enterprise is heading toward “specialized intelligence,” and the shift is “from renting generic AI to owning intelligence that’s uniquely theirs.” That is Cloudera’s framing.

Sovereign AI, as defined on Mistral’s post: customer control of data, intelligence, compute, and operations; models adapted and owned on open weights; training and inference on infrastructure and in jurisdictions the customer chooses. That is the companies’ positioning — not a regulatory finding.