In 2035, a SOF team operating forward in a denied environment will generate more data in a single hour than an entire joint task force produced in a day during the early years of Afghanistan. The environments where operators need AI the most are precisely the environments where traditional AI architectures fail.

What happens when AI models get small enough to run on a device you can hold in your hand?
You get intelligence that works anywhere — without waiting for the cloud, without sending sensitive data across networks, and without shutting down when comms go dark.

In a world where data is growing more distributed, bringing AI to the data is not just an option, it’s a strategic imperative. It’s how you’ll comply with emerging data laws, how you’ll deliver instant experiences to customers, and how you’ll outmaneuver competitors who are still busy shuffling data back and forth.

Adaptive AI agents will become pervasive in the fabric of national security and public services, much like software and networks are today.

AI offers vast promise in areas from intelligence to operations to cyber to resource allocation. But that promise hinges entirely on whether you can explain every decision it makes. If you can't map a recommendation back to reasoning and data, it’s not ready for primetime.