By Paul Montgomery, Executive Director, EMEA, Ocient National Security Solutions
Walking the floor at DODIIS Worldwide 2026, I heard plenty about AI, cyber, OSINT and modernization. That was fitting, given this year’s theme: “DIA Next: Intelligence Technologies for Battlespace Lethality.”
But as the conversations went deeper, a more fundamental question kept surfacing: Can the data underneath all of this actually keep up?
It’s a question our team at Ocient National Security Solutions (ONSS) spends a lot of time thinking about. At our booth, my colleagues Galen Huss, Bill Minarchi, Tony Ibanez and Ronnie Geronimo and I held back-to-back conversations with practitioners and leaders from across the intelligence community. Again and again, those discussions came back to the same challenge: organizations have more data than ever, but keeping it, analyzing it and getting answers from it quickly enough remains difficult.
Tom Hofmann, VP Solutions for ONSS, captured one part of the problem during his Coffee Talk: You can’t find what you don’t keep.
The data an organization discards today because it’s too expensive to store or too difficult to query may contain the connection that matters tomorrow. An indicator that looks like noise today can become significant months later when a new name, location or device enters the picture. If the underlying data is gone, so is the opportunity to make that connection. But keeping full-fidelity data is only useful if analysts can actually work with it.
That was the other side of the challenge we heard throughout DODIIS. Network engineering teams described managing petabytes of telemetry that they are forced to age out earlier than they would like. Several people posed a simple question: What could we find if we could keep all of it and query it quickly?
For many, that’s something they’ve never been able to test.
This is where the conversation moved from storage to mission impact. The goal isn’t simply to retain more data. It’s to give analysts the ability to ask questions across that full history and get answers quickly, without moving data through a long series of systems or reducing it first. That can shorten the distance between an analyst’s question and the intelligence delivered to warfighters and decision-makers.
AI adds another dimension to the problem.
We heard from teams that have begun pointing AI agents at their own data, only to see costs rise as those agents repeatedly query and verify results. If the underlying data platform can’t reliably return consistent answers, the AI has to do more work: asking again, checking results and consuming more tokens in the process.
That makes the quality and consistency of the data layer increasingly important. AI can help analysts move faster, but it still depends on the systems underneath it to return reliable answers.
The sheer scale of that data was also a central topic during the “Hyperscale Intelligence Tech Stack” panel, which featured current senior leaders from ODNI and DIA alongside others with experience at CIA and NCTC.
Andrew Borene, ONSS Vice President and a former senior intelligence official, described a challenge we heard throughout the event:
“As usable datasets grow past the petabyte scale needed for analysis in minutes, the answer isn’t more AI workload and tools alone. The highest modernization priority for any warfighting intelligence effort has to be new foundational architecture that brings the compute to the data where it lives. Given the stakes in missions involving lethality, AI queries have to run inside each agency’s data ownership, policy, and compliance controls, with full auditability. We can get there, but it will take new thinking about how data is processed.”
That point gets to what I took away from DODIIS. The intelligence community isn’t short on new tools. The harder challenge is making sure the underlying data infrastructure can support them at the scale, speed and reliability the mission requires.
Back to my colleague, Tom Hofmann, who summed up that shift well.
“The DODIIS conversation has matured,” he said. “Less about AI as a headline, more about whether the data beneath it moves at mission speed. Billions to trillions of rows, real-time decisions, that’s not a future problem. It’s a today problem, and it’s exactly where OcientAIQ™ lives.”
For our team, that’s the opportunity: to help intelligence organizations retain more of the data they already generate and analyze it quickly, while keeping that data within the controls required by the mission.
Ultimately, the value isn’t in keeping more data for its own sake; it’s in being able to find the connection that matters when it matters.
Learn more about OcientAIQ and Ocient National Security Solutions at ocient.com/national-security.