IBM acquires Confluent for $11B to supercharge WatsonX AI with real-time data streaming

IBM is spending $11 billion in cash to acquire data-streaming specialist Confluent, in a move aimed squarely at fixing one of enterprise AI’s biggest headaches: getting clean, real-time data into models and agents.

Under the deal, IBM will pay $31 per share for all outstanding Confluent stock, valuing the Mountain View–based company at about $11 billion and representing a premium of more than 30% over its pre-announcement price. The transaction is expected to close by mid-2026, subject to regulatory and shareholder approval.

Confluent’s platform, built around Apache Kafka, handles event streams and real-time data pipelines for more than 6,500 customers worldwide. IBM says it plans to plug that technology into its hybrid cloud and WatsonX AI stack, pitching the combination as a “smart data platform” that can connect, process and govern information across on-prem, multi-cloud and edge environments.

The deal is IBM’s largest software acquisition since Red Hat and follows its 2024 purchase of HashiCorp as CEO Arvind Krishna continues to pivot the company toward higher-margin cloud and software revenue. IBM expects the Confluent acquisition to be accretive to core earnings within a year of closing and to boost free cash flow soon after.

Confluent co-founder and CEO Jay Kreps will join IBM Software once the transaction completes, with the unit reporting into senior vice president Rob Thomas. Both companies are framing the deal as a way to turn data streaming into the “railroads” of the AI era – the underlying infrastructure that keeps models and applications fed with fresh, trustworthy data.