Two-Thirds of Enterprises Now AI-Dependent, but Operating Models Lag

  • 66% of enterprises now run processes that would degrade or break if AI were switched off.
  • Only 35% of enterprises have a formal operating model for AI.
  • 42% of production AI use cases started in business units, not IT or central AI teams.
  • 73% of leaders prioritize IP and differentiation over cost for AI infrastructure control.

The study highlights a growing disconnect between the rapid adoption of AI and the slower evolution of operating models to support it. As enterprises increasingly depend on AI for critical processes, the ability to integrate, secure, and monitor these systems will become a key differentiator. CoreWeave and Humatica's findings suggest that companies prioritizing IP control and decentralized AI building are likely to gain a lasting competitive advantage.

Operating Model Gaps
How enterprises will bridge the gap between AI dependency and formal operating models.
Decentralized AI Building
Whether the trend of business units independently building AI will lead to operational inefficiencies.
IP Control
The pace at which enterprises will shift from buying AI as a service to owning proprietary models and data.