We’ve talked a lot around here about how AI is changing the way wars are fought, but we haven’t spent all that much time talking about how it’s going to change the way intelligence is gathered and analyzed. That changes today.
This morning, AI-powered intelligence company Whitespace announced in an exclusive release to Tectonic that it has raised a $3.2M seed round to scale its “warfighter-ready intelligence analysis agent,” called Iris.
What that actually means: The money will go towards making sure Iris can work in edge deployments and onboard hardware, the company said.
“Operations don’t wait for perfect data or perfect connectivity,” Jackie Barbieri, co-founder and CEO of Whitespace, said in a statement. “Iris is the first AI agent built to do intelligence analysis for the warfighter.”
The round was co-led by MaC Venture Capital and Caffeinated Capital with participation from “family offices and strategic angels.”
In the shadows: Now, you might not have heard of Whitespace in VC-defense tech land, and that’s for good reason: This is the first time the company has raised money since it was founded 11 years ago.
- Barbieri told Tectonic that the company was “bootstrapped” for a decade. She only decided to raise because they realized how fast demand was growing. “In 2024, we saw our ARR grow modestly from like $250K-300K to $900K,” she said. “From there…we finished 2025 at $3.99M in ARR.”
- Barbieri knew the company had a backlog and that there were a ton of interested customers out there—they needed funding to supercharge development.
Eye Spy: At its core, Whitespace is an intelligence analysis company. Barbieri herself has spent much of her career as an analyst specializing in—this is important—Activity Based Intelligence (ABI).
- ABI, put simply, is an intelligence methodology that focuses on understanding what people, objects (like vessels), and systems do over time—their patterns, routines, and changes in behavior—rather than just identifying who or what they are.
- The idea is to use lots of different sources (ISR, SIGINT, commercial imagery, you name it) to try and identify patterns—as well as aberrations from them.
The minute Barbieri started doing ABI, she was hooked. “Once I saw the world that way, I couldn’t unsee it,” she said. “I developed this really strong conviction that this was how we had to do intelligence, period. Not just because it made people more effective, but because the way the data was being captured and stored was also machine-readable.”
Race to automate: You won’t be surprised to learn that Iris—Whitespace’s flagship AI-powered intelligence analysis tool—is built on ABI.
Here’s how the whole thing works:
- When you open up the tool, it looks a lot like the chatbots we’re all familiar with (admit it). At its core is an off-the-shelf LLM trained by intelligence experts like Barbieri.
- Analysts can plug in a query (say, “Where has this Chinese vessel traveled in the past week?”) and Iris uses a suite of commercial satellite imagery, third-party device data, and her ABI training (yes, she is referred to as “her”) to not only figure that out, but also intuit what those movements mean.
- When using Iris, users can choose what LLM they want to use (in the example we saw, ChatGPT and Claude were both available) and which agent.
- Specific agents have been trained to have specific skills—for example, Barbieri said, certain agents are trained specifically to work with the DoD.
- The user can then chat with Iris to build out an intelligence analysis. This can either be done as a back-and-forth (where you instruct her on each step of the analysis), or she can do it automatically.
Barbieri showed us two use cases.
- The first was sea-based. Iris used satellite imagery to track a Chinese aircraft carrier that participated in a military exercise—she was able to figure out what days the vessel likely participated in the exercise based on its absence from port.
- The second was land-based. The tool took locations where drugs had been seized in Ecuador, then used third-party aggregated cellphone data (like, where devices ping) to figure out the likely trafficking route between these spots.
Each of these queries took just a few minutes. Iris is also able to churn out reports and mission planning documents. In the future, Barbieri said, they’re hoping to take these mission plans and automatically task autonomous assets, like UAVs.
Crowd favorite: And people are actually using Iris—those ARR numbers speak for themselves, after all. Barbieri couldn’t tell us who, specifically, they were contracted with, but that they were working with “three different DoD organizations.”
Plus, according to the company’s statement, “Over the past two years, Iris has delivered more than 1,000 operational planning products across seven combatant commands, with outputs shared among more than 50 partner nations.”
On the edge: So, what’s the plan from here? Bringing Iris to the edge.
Specifically, “The new funding will support additional deployments, deeper integration with crewed and autonomous systems, team growth, and continued development of Iris for contested operations at the edge,” the company said.
They’re already “validating deployment of Iris onboard an aircraft under constrained compute” with the US military, and are hoping to make her work with even less connectivity with this new pile of cash.