Databricks co-founder outlines what stalls enterprise AI — and how to fix it
At TechCrunch Disrupt 2026, the Databricks co-founder delivered a pragmatic message: enterprise AI is no longer judged by how exciting it is, but by how safe and operationally ready it is for broad deployment. That shift marks a maturation of the market—companies are increasingly asking for repeatable, auditable processes that reduce risk and demonstrate measurable value.
The talk highlighted the top deal-killers: unclear governance, lack of security and privacy assurances, poorly defined deployment paths from pilot to production, and insufficient alignment across legal, security and executive teams. Rather than presenting these as blockers, the session framed them as solvable checkpoints. Vendors and internal teams that proactively address these issues can turn due diligence into a competitive advantage.
Practical fixes and playbooks
- Build standardized MLOps and observability pipelines to ensure reliable, auditable models in production.
- Document governance, data lineage, and privacy controls so procurement and legal teams can sign off faster.
- Prioritize business outcomes and ROI in sales conversations to move discussions from theoretical risks to concrete returns.
- Invest in cross-functional onboarding so security, compliance and executives are aligned early.
Overall, the message from Disrupt was optimistic: enterprise AI isn’t failing, it’s maturing. As vendors and customers adopt clearer safety and deployment standards, more pilots will convert into production systems that deliver real business value—accelerating adoption across industries.