Jeremy B. Terk

⚙️ “The side that iterates fastest will win.” That’s how Domino Data Lab COO Thomas Robinson describes the future of defense in a new Axios Future of Defense Q&A — where he shares why software-defined warfare and rapid AI redeployment will define mission success. From the U.S. Navy to the frontlines of innovation, Domino is helping defense leaders operationalize AI faster, safer, and with trust. Read the full interview with Colin Demarest on Axios: https://gag.gl/6MztHJ #DefenseInnovation #AI #SoftwareDefinedWarfare #OperationalAI

16

I've been watching defense organizations pour billions into AI while their acquisition timelines stretch beyond a decade. The contradiction is stark. Here's what I'm seeing: Defense contractors raised over $3 billion in 2024 alone. AI-related M&A in aerospace shot up over 2,000% quarter-over-quarter. The Pentagon oversees 685 AI projects. Yet the average major defense acquisition still takes 11 years to deliver initial capabilities. Meanwhile, adversaries are fielding equivalent systems in months. The typical analysis blames bureaucracy. I think the problem runs deeper. Modern AI systems are fundamentally compositional. They're built from multiple interacting components - models, retrievers, vector databases, evaluation frameworks, monitoring layers. These components evolve independently. They connect through APIs. They get swapped out as better alternatives emerge. This is how software works now. But defense procurement was designed for monolithic platforms. Aircraft carriers. Fighter jets. Systems with 30-year lifecycles where every component is specified upfront. The entire acquisition framework assumes you can define requirements completely, lock them down, then spend years building to spec. You cannot procure a compositional system through a monolithic process. It's not a speed problem. It's a category error. I believe the solution requires procurement itself to become modular. Acquisition processes need to mirror the architecture of the systems they're buying - with rapid evaluation cycles for components, interoperability standards that let modules from different vendors work together, and the ability to swap pieces without re-validating the entire stack. Defense organizations that figure this out first won't just move faster. They'll be able to actually deploy the AI systems the modern threat environment demands. This architectural alignment between how we build and how we buy is what we explore in our work at www.kriraai.com. #KriraAI #DigitalTransformation #AIProcurementInnovation #ModularAcquisitionProcesses #CompositionalAISystems #DefenseTechModernization #InteroperabilityStandards #RapidEvaluationCycles #DefenseAIStrategy #CloudNativeDevelopment #OperationalExcellence #CTOChallenges #FounderGrowthHacks #TechLeadership #BusinessGrowth #StrategicITConsulting #AIEngineering #InnovationPipeline #CustomSoftwareDevelopment #EnterpriseAIIntegration

1

US Congress, delaying SBIR reauthorization threatens to starve out the most innovative and adaptable part of our industrial base. Large businesses have immense resources and momentum, but that makes them slow (and expensive) to adapt. Small businesses rapidly evolve and innovate to fill the missing links and niches of our current and future supply chain... when resourced to survive. Foreign competition is evolving and accelerating. If you want lasting American tech superiority, don't starve out our industrial edge. #ReauthorizeSBIR

21

6 Comments