xAI Grok models land on Databricks Agent Bricks

xAI announced at the 2026 Data + AI Summit that its Grok models are now natively available on Databricks Agent Bricks, letting enterprise teams power production AI agents directly on data within Databricks' Lakehouse architecture. The flagship grok-4.3 model — featuring a one-million-token context window — is available through the integration at what xAI calls competitive pricing, and Databricks guarantees zero data retention, a key requirement for regulated customers.

The deal is xAI's clearest enterprise-distribution move to date, planting Grok inside one of the largest data and AI platforms used by Fortune 500 data teams. xAI promoted it on its official account (2,192 likes, 327 retweets): 'Grok models are now available on Databricks Agent Bricks. Bring SpaceXAI's latest models to your enterprise data to power capable AI agents.' The branding nod to 'SpaceXAI' reflects xAI's ongoing entanglement with Musk's broader empire.

Competitively, this matches similar Databricks integrations with Anthropic, OpenAI, and Meta models, turning Agent Bricks into a multi-model battleground. The 1M-token context window is the headline differentiator, pitched for agents reasoning over large enterprise document corpora without aggressive chunking. Zero data retention is the trust hook against fears of training-data leakage.

The awkward backdrop: this enterprise-credibility push lands the same week the Pentagon disclosed Grok's role in firing 2,000 munitions and Yann LeCun publicly branded xAI a 'failure.' Enterprises evaluating Grok must weigh its technical merits against the reputational and governance baggage attached to the brand. Whether Databricks customers adopt Grok at scale — or stick with Anthropic/OpenAI defaults — will be the real test of xAI's enterprise ambitions.