Actian, the data and AI division of HCLSoftware, has announced the release of Where Enterprise AI Runs: Deliberate Workload Placement Across Cloud, On-Premises, and Edge, a research study conducted by BARC. The report examines how organizations are making decisions about where to deploy AI workloads as operational, security, and governance requirements evolve.

According to the research, 77% of organizations use the cloud for some AI workloads. At the same time, 76% report that deploying AI outside the cloud has become more important over the past year.

"Production AI brings more stringent requirements for control, protection of critical data, and auditability that experimentation does not demand," said Dr. Carsten Bange, founder and CEO of BARC.

Bange continues, "While cloud is still the starting point, trust, dependency, and sovereignty are increasingly deciding factors in AI workload placements."

Factors Influencing AI Workload Placement

The BARC research identifies several factors influencing organizations’ decisions to deploy AI outside the cloud, along with the benefits they associate with these deployments.

• Internal drivers: Data sensitivity (55%) and mission-critical AI processes (49%).
• External drivers: Cybersecurity concerns (45%); customer, partner, or contractual requirements (45%); and dependency on specific cloud or technology providers (42%).
• Reported benefits: Improved data privacy and confidentiality (62%), a stronger regulatory compliance posture (40%), and cost optimization (37%).

The study also identifies an association between hybrid architectures and AI maturity. Among organizations using hybrid architectures, 65% report having AI in production or delivering measurable business value from AI, compared with 60% of organizations using cloud-only architectures.

Understanding Data and AI Sovereignty

Actian describes data and AI sovereignty as the ability to control, enforce, and demonstrate how data, AI models, and agents are used across deployment environments. This concept encompasses three dimensions:

• Control: Defining who can access data and models, under what conditions, and from which locations.
• Enforcement: Applying established policies during operations, rather than relying solely on documented requirements.
• Proof: Maintaining auditable evidence that policies have been applied.

As organizations evaluate AI deployment strategies across cloud, on-premises, and edge environments, governance requirements may extend across multiple systems and locations. Consistent governance requires policies to remain applicable as data moves between environments. Machine-readable data contracts can help organizations define and apply data policies consistently wherever data is processed.

"The research shows that organizations are becoming much more deliberate about where they run AI, particularly as concerns about sensitive data, security, and technology dependencies grow," said Emma McGrattan, Chief Technology Officer of Actian.

McGrattan continues, "But moving workloads out of the cloud doesn't automatically give you greater control. Sovereignty comes from being able to enforce your policies wherever AI operates and prove they're working. That's especially important as AI agents become more autonomous and harder to oversee."

Research Methodology

The study surveyed 329 respondents in June and July 2026. Participants held roles across executive leadership, architecture and engineering, data and AI leadership, and other functions. Respondents represented industries including manufacturing, IT, financial services, and healthcare, with participation spanning Europe, North America, and Asia-Pacific.

Source: Actian

Stay Connected with ISE Magazine 

Subscribe to our newsletters and magazine for the latest telecom insights, explore the current issue for in-depth features and strategies, and register for upcoming webinars to learn directly from industry leaders.