AI security and compliance concerns are driving a private cloud boom
Cost is a key driver behind the shift back to private cloud, but growing concerns over AI security and compliance appear to be having a bigger impact.
That’s according to a survey by GTT Communications that found growth rates for private cloud far outstripping investment in public cloud, following a wider trend of cloud repatriation.
The research follows a survey earlier this week from Node4 that showed a whopping 97% of respondents planned to move at least some workloads back inhouse. Analyst firm IDC found similar results earlier this year, saying cloud repatriation was not “a wholesale migration” but reflected missed expectations around costs.
SaaS company 37 Signals made headlines for saving $10 million a year by ditching public cloud and buying its own infrastructure — suggesting at least some repatriation is cost focused.
GTT Communications’ research said its own results “supports that trend” and found that primary drivers include unpredictable public cloud costs and concerns about vendor lock-in. However, its research also suggests some private cloud adoption is driven by privacy, security, and compliance requirements — with control over AI a major motivator.
The survey of US and European enterprise leaders found that private cloud spending above $10 million annually will increase from 43% last year to 54% this year — a 24% growth rate that’s far ahead of the 13% growth rate in public cloud spend for the same group of companies.
Indeed, public cloud remains favored for use cases that require high reliability and scalability.
Is AI driving private cloud growth?
The shift back in-house or to private cloud is partially because of AI, the company suggested. Notably, GTT’s survey found more than half of AI workloads are already in private cloud or on-premise environments, largely for better security (56%), to meet compliance requirements (51%) or because of the specific needs of AI workloads (50%), in particular secure, partitioned AI testing or deployment.
GTT said that cost remained a factor but was much lower at 35%, but noted that AI applications are resource-intensive and can further expose enterprises to higher costs that are less predictable.
“We know many companies are now shifting their sensitive workloads to private clouds as part of broader multi-cloud and hybrid strategies designed to support agentic AI and other complex AI initiatives at scale,” said Bastien Aerni, VP of Strategy and Technology Adoption, GTT.
“This distributed approach allows enterprises to balance performance, security and compliance requirements while optimizing costs to advance their AI ambitions.”
The report added that enterprises can either move their private data to where the AI model is hosted or move the model to where the data lives.
“It is widely recognized that enterprises are appreciating more and more the value and sensitivity of proprietary data for AI models and are showing a preference for private cloud for AI,” the report added.
GTT said that complex AI workloads face challenges with repatriation. The survey found 43% of respondents faced issues with migrating apps and data from public cloud deployments.
“Enterprise infrastructure strategies are evolving fast as AI workloads surpass the limits of traditional architectures,” Aerni said.
“Organizations are refining their cloud environments, whether public, private or hybrid, to support their AI initiatives at scale. But many still underestimate the complexity involved. In our experience, without reengineering connectivity and security architectures, even the most ambitious private cloud strategies can fall short.”
MORE FROM ITPRO
Source link
