The data is in: production AI is moving to private cloud
The clearest signal yet that the AI infrastructure conversation has flipped came from Broadcom’s Private Cloud Outlook 2026, a survey of 1,800 senior IT leaders across eight countries. Net intent to grow private-cloud investment over three years jumped from 51% to 72%, and private-cloud spend is now growing at more than twice the rate of public cloud. A majority — 56% — now run or plan to run production AI inference in a private cloud, while public cloud as the primary environment fell from 56% to 41% in a single year.
The why is the part that matters. For the first time in the study, cost overtook security as the top concern about public cloud. 86% of IT leaders said geopolitical and regulatory factors now directly shape their IT strategy, and security and compliance remained the single biggest factor in deciding where a workload runs. As AI moves from pilot to production, enterprises want it on infrastructure they control.
- This is the whole thesis, in someone else’s survey: production AI wants private infrastructure — for cost, control, and sovereignty.
- The hard part isn’t deciding to go private; it’s building and running it. That is exactly what we do.
Source: Broadcom Private Cloud Outlook 2026. See the cost calculator and how it works.