“The single biggest obstacle to AI adoption is distributed data. When you’re adopting agentic AI, where machines are going to take action based on the information they have available, they need access to that information in real time,” said Kevin Bohan, Director of Product Marketing at Denodo.
Golden Dome and Combined Joint All Domain Command and Control (JADC2)
Modernization programs such as Golden Dome and Combined Joint All Domain Command and Control (JADC2) aim to combine data from sensor networks, intelligence platforms, logistics databases, and operational commands so decision makers can see across previously separate domains. Those programs assume a level of data sharing and integration that, the reporting shows, most defense agencies have not yet achieved because the necessary data infrastructure layer is missing.
The Missing Data Layer
What defense organizations lack is a connective data infrastructure layer that sits between distributed systems and the AI applications that require their data. This access layer is meant to provide unified, governed access regardless of where data lives. Hyperscalers, cloud service providers, and enterprise-grade technology companies already offer these capabilities as managed services or embedded core infrastructure, but many defense agencies have not built an equivalent layer across their networks.
When that layer is absent, predictable failures follow: AI applications struggle to access information reliably; teams duplicate work as they individually address access problems; security policies are applied inconsistently across systems; and the same piece of data can be interpreted differently by different commands. In short, systems designed to interoperate can end up working against one another.
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scrambleSemantic consistency and unified access
The data infrastructure layer performs four key functions, the article explains. The first two are unified access and semantic consistency. Unified access gives applications a single interface instead of forcing them to navigate several incompatible systems. Semantic consistency translates outdated field names into common meanings so that applications — including AI models — can treat equivalent concepts as equivalent.
“Most systems weren’t designed with AI in mind, so without context, the models are forced to guess,” Bohan said. The report argues that when different systems use different terms for the same concept, AI can make incorrect assumptions; semantic consistency reduces that ambiguity and thereby enhances the reliability of AI outputs.
Centralized governance and live context
The third and fourth functions of the layer are centralized governance and live context. Centralized governance means access controls and security policies are applied where queries happen rather than being set separately on each underlying system. Live context means retrieving data directly from operational systems in real time instead of relying on snapshots or delayed extracts.
According to the reporting, agentic AI — systems that take action based on available information — requires real-time access to current information. Without live context, AI systems may act on data that is hours or days old, degrading reliability and increasing operational risk.
What this means for military services, technologists, and hyperscalers
- Military services and the Department of War: The piece notes the U.S. military has made large investments in AI capabilities but warns the underlying data infrastructure has received less attention. For the Department of War deploying AI at scale, the article concludes, the data infrastructure layer is essential.
- Technologists and security teams: Engineers and security practitioners must prioritize unified access, semantic models, and centralized policy enforcement so AI can draw from current, consistent data rather than fragmented, inconsistent sources.
- Hyperscalers and cloud service providers: The reporting highlights that these vendors have already built managed services and embedded infrastructure that address the problem, framing them as existing models for the capability gap defense agencies face.
The takeaway is concrete: AI capability is necessary but not sufficient. Without a data infrastructure layer that provides unified access, semantic consistency, centralized governance, and live context, AI systems will continue to struggle in distributed military environments. That infrastructure, the reporting says, is the foundation that determines whether AI can scale effectively across distributed environments and operational domains. For the Department of War deploying AI at scale, the data infrastructure layer is essential.
