Research: Enterprise AI Workloads Are Tipping Towards Private Cloud
Broadcom is framing its most current private cloud research study around what it calls a business AI “tipping point,” saying production AI workloads are altering how companies assess cloud architecture, cost, security, and governance.
The company’s Private Cloud Outlook 2026 report, entitled “The AI Tipping Point,” is based upon an international survey of 1,800 senior IT decision-makers across the Americas, Europe and Asia-Pacific. In a news announcement, Broadcom asserted that the AI experimentation phase is over which private cloud is where enterprise AI work are being deployed for security and scale. The company said the shift is being formed by “costs, complexity, and control” as business move AI work into production.
Production AI Moves Toward Personal Cloud
Broadcom said 56% of business surveyed are running or planning to run production AI inferencing on private cloud. The business likewise said public cloud use for the same work fell 15 percentage points year over year, from 56% to 41%.
Click image for bigger view.] Where AI-based Applications or Workloads Currently Run(source: Broadcom). The report says public cloud stays part of business IT strategy for experimentation, flexible capability, and specialized services. However Broadcom said AI production work present sustained compute need, sensitive information circulations, governance requirements, and efficiency expectations that can expose limits in a simply public cloud approach.
The shift is also reflected in repatriation information. Broadcom stated 83% of enterprises are thinking about or have actually currently repatriated work from public cloud to private cloud, and 50% have actually currently repatriated some work. In 2025, the corresponding figures were 69% considering or currently repatriating and 35% currently having actually done so.
AI looked like a repatriation classification for the very first time in the 2026 research study. Broadcom said 43% of organizations repatriating work are moving AI training, large language models, and inference from public cloud to private cloud.
In a Broadcom blog post on the report, Prashanth Shenoy, vice president of marketing for the VMware Cloud Structure Department at Broadcom, wrote that “as business aim to scale, the instructions has altered.” The very same post said enterprise AI “has found its facilities home. And it is private cloud.”
Expense Changes Security as Top Public Cloud Concern
Broadcom said expense has surpassed security as the leading public cloud issue in the 2026 study. According to the report, 31% of respondents cited cost management as a leading public cloud obstacle, up from 26% in 2025.
The report likewise said 97% of surveyed IT leaders think some part of their public cloud invest is wasted, while 52% stated that waste surpasses 25%. Broadcom connected those findings to AI facilities pressures, consisting of compute, storage, bandwidth, GPU prices, data motion fees, and unforeseeable use patterns.