In case you didn’t know, the U.S. Army is trying to change the way it does things.
Since 2023, under Chief of Staff Gen. Randy George, the service has launched a program called Transformation in Contact, which pushes new technology out to soldiers for real-time testing in the field. And in May of this year, it kicked off the Army Transformation Initiative, aimed at turning the Army into “a leaner, more lethal force.”
At the center of these efforts is Alex Miller, the Army’s Chief Technology Officer. He’s been in the seat since 2023 and, as he puts it, “advise[s] the Secretary of the Army on—using big, big hand-wavy language here—all technology for what the Army’s doing.” In practice, that’s included everything from leading the Army’s AI and data architecture strategies to prototyping next generation command and control (NGC2). In other words, if there’s one person pushing innovation inside the Army, it’s Miller.
Tectonic sat down with him a few weeks back to talk about the work he’s doing, the changes he’s advocating for, the tech he’s excited about—and what, exactly, keeps him up at night.
Note: This interview has been edited for length and clarity.
Tectonic: Tell us a little bit more about how you got into this role.
Alex Miller: People ask me this pretty often because it sounds like a very prestigious job—and it is. I love my job, my boss, and my team. I couldn’t draw you a process map of how to get here, so I’ll just tell you what happened to me. I was the Science and Technology Advisor for Army Intelligence, so I had all of the spooky intelligence technologies in my portfolio. The Army was doing a big exercise called Project Convergence, which was intended to identify a lot of the architecture and technology gaps in joint warfighting.
Because I was the intel guy—and all warfighting starts and ends with intel—I ended up owning a lot of the architecture, working for Lt. Gen. Scott McKean at Futures Command. I was dual-hatted in the G-2 there and had an opportunity to be in the field for a couple of weeks. I came back to the Pentagon for one day to brief our Under Secretary on a couple of topics. I happened to meet the Vice Chief of Staff of the Army, Gen. George, who’s now the Chief. I briefed him for about five minutes. He said, “Hey, I want you to send me your observations.” I sent those observations that night, including what we were doing to fix them or change processes. The next day, I flew back out to the Mojave Desert, to the National Training Center. I got a text from his XO before the Vice came out: “Hey, you gave an honest view to the Vice. He wants you to brief him when you’re out there.”
A couple of days later, when we were in the field at NTC, he asked me to come work for him. It was a flashbang moment that I am thankful for—having a lot of expertise, being willing to go to the field and get dirty, being really willing to do the hard stuff, and having a senior leader say, “Okay, I’ll give you more responsibility.”
What were the observations you made? When you looked at what the Army needed in terms of tech, what was it?
This was two and a half, maybe three years ago. I’ll be very candid: Tech wasn’t our problem. I’ll tell you some of the tech we were looking for, but our biggest challenges were process-related—falling into a post-Cold War mindset for how we implemented technology.
Everyone talks about standards. Any founder, investor, or defense company talking to a government—whether it’s the US Army or the British Army—starts with interoperability and standards. But most of the standards we have are older than I am, and we still haven’t solved the problem. That tells me standards aren’t really the problem.
In the real world, where people have to make money and things actually need to work, plenty of standards are implemented—TCP/IP, 3GPP, 4G LTE, and enterprise protocols. These are standards implemented in a common way. Our biggest challenge in the joint warfighting enterprise was that everyone knew the standards, but implemented them slightly differently. It was like putting a LEGO next to a Duplo block—there are pegs and holes, but they’re not the same size. You haven’t done it the same way.
That kept us from having more sophisticated conversations about what we actually needed. Instead of buzzwords like AI/ML, we needed to talk about putting a foundational data layer in place, accessible to everyone. We needed to talk about design patterns and shareable data formats. Those were our real challenges.
Now, we’ve solved a lot of that by falling back on what commercial industry does really well: put things out fast, test them, kill what doesn’t work, and scale what does.
Tell me a little more about how your assessment has changed over the last three years.
Right now, in 2025, we’re at an inflection point. We’re working to scale the things that work, because we’ve spent the last 18 months going into the field and infusing our units with technology.
We started this process called Transforming in Contact. It means exactly what it sounds like: Instead of going through a bespoke process where people who aren’t in the fight or in the field write requirements and make purchasing decisions, we take a bunch of commercial and military tech and give it directly to units, and let our acquisition professionals and requirements writers engage directly with our soldiers. We let them self-organize—their formations and their kit—around some common principles.
For example, they must be able to talk, using voice and data. Their network has to function both with access to cloud-based technologies and in disconnected environments, because this isn’t a standard business use case. You don’t always have infinite bandwidth or constant connectivity. They need primary and alternate means of voice communication with commanders and subordinates.
We flooded them with technology—almost to the point of breaking their formations—to see what worked. Then we asked for their feedback. They came back and said: here’s what worked (lightweight tablets and phones, drones), here’s what was directionally correct but needs improvement (mobile EW capabilities), and most importantly, here’s what does not work and won’t meet our future needs (like heavier vehicles, legacy radios, and larger server stacks).
That let us skip the usual 18-month process of killing bad ideas—we just didn’t start them.
How widespread was this tech experimentation, and how hard has it been to scale the tech that worked into major programs?
We started with three Brigade Combat Teams: 2nd Brigade, 101st Air Assault; 2nd Brigade, 25th Infantry Division; and 3rd Brigade, 10th Mountain Division. That sounds like a lot, but we have over 50 brigades across the Army—so this was version 1.0.
Now we’re in version 2.0: how do we scale to all three divisions, their headquarters, and their sister brigades—three to four per division. Now we are sorting through how we push it to logistics brigades or the DIVARTY, who own the artillery. It’s not just about shooters and door-kickers. We’re scaling to enablers, too.
It’s been challenging, not because the technology isn’t good, but because of how we fund things. Traditionally, we pick something and plan to use it for 20 to 30 years. Then we spend five years fielding it—about 20 percent of the Army per year. That makes sense for vehicles or aircraft. It makes zero sense for algorithms or software patches.
Changing the culture of how we deploy and scale technology has been the hardest part.
And how successful have you been in changing that?
I actually think we’ve been wildly successful, and that’s been focused in a few areas up front.
Gen. George—who I think is one of the coolest officers in the Army—commanded at every echelon in combat: battalion, brigade, division, and corps. When he came in, he said the first thing we needed to fix was the network—our command and control had to work.
That became my big rock for about a year and a half. We stripped out a lot of complexity. Think about your day: You just switched from centralized Wi-Fi to a puck to get better bandwidth. Two years ago, our soldiers didn’t have that kind of flexibility. They were stuck with their legacy programs. If it didn’t work, it just didn’t work.
We said: Let’s add diversity to comms. Let’s use what industry does well—routing, redundancy, flexibility. Use indigenous cellular if it works. Use Wi-Fi, use Starshield. Only fall back to MILSATCOM if absolutely necessary. Then layer on security so our soldiers aren’t creating a giant signature in the middle of nowhere.
Also, let’s get rid of the 400-pound server stacks—that’s a real number. Let’s give them powerful tablets or laptops instead. In the past, if you walked into a tactical operations center, it looked like Christmas in London—brightly colored cords strung everywhere.
What if we removed all that complexity? If we start by setting that foundation correctly, almost everything else rides on top. We can plug in new radios. We can roll out new software.
How do you keep all of that secure during the transition?
A couple of ways. For digital security, we’ve got great relationships with the NSA, GCHQ, and other allied security agencies. But we also borrowed from the commercial world. AES-256 and other commercial cryptography solutions are used by the finance sector and email providers. Starting from that baseline—instead of building bespoke crypto—made more sense.
Here’s the example I like: AES-256 is still the standard on the internet. It has billions of users and decades of hacker attempts to break it behind it. So when someone comes to me with a new security algorithm, I ask: “Do you have billions of users? Have hackers tried to break it for 20 years?”
We partner with NSA and commercial providers to get this right.
On the physical security side, we assess what actually needs to be secure. Not every piece of military data needs Fort Knox—level lockdown. That mindset is wasteful—it costs processing and comms bandwidth. Of course, real secrets get locked down. But if you’re in a firefight and someone’s shooting at you, they already know you’re there. So we’ve tried to be realistic about what needs to be protected and when.
What does the Army of 2030 look like to you?
I’m going to judo-flip the question a bit—and here’s why. I could list a bunch of tech and get everyone salivating, but 2030 is five years away. When someone answers a “five years from now” question, it gives them a lot of time to do nothing and scramble at the end.
What we’re trying to do is set conditions right now. So I’ll be a little ambitious: By 2027, our goal is that every division in the Army has unmanned systems. Everyone’s enamored with FPV quadcopters—and sure, that’s part of it—but we’re also talking about heavy-lift unmanned platforms for ammo resupply and medevac, or one-way attack munitions to lower cost per kill if we ever go to war.
And I’m not a war hawk. I’m not a warmonger. But if we go to war at scale, we need the best stuff to create the most havoc for the enemy, as fast as possible.
That also means revitalizing our organic industrial base. Most people don’t realize the Army makes all the small arms ammunition for the U.S. military. We make 155mm rounds like you’ve seen used in Europe, that we’ll use for operations globally, plus a lot of vehicle and radar repairs—all organic to the Army.
But much of that infrastructure was built for WWII and never updated. Now we’re thinking about how to leverage robotic automation—the same way Detroit reinvented itself, we need to do that too.
And for our close combat forces—our door-kickers and shooters—we’re making sure they have the best weapons, the lightest armor, and the most capable night vision. I grew up playing first-person shooters. I had a HUD, I could see my squad, and get alerts. How do we do that for real soldiers, without loading extra weight onto their necks?
Even body armor: When I was in Afghanistan, our level IV plates were steel and heavy. You’d be carrying 30–40 pounds before water, ammo, or anything else. How do we reduce that weight without compromising protection, especially as soldiers move longer distances in Europe or the Pacific?
So when someone asks me what 2030 looks like, I answer: We’re doing the work now so that in two years we’re different, and in four years, we’re two full revs beyond where we are today.
I notice a lot of the innovation you point to isn’t futuristic—it’s better versions of things soldiers desperately need right now. Do you think there’s too much hype around AI, unmanned systems, or other pie-in-the-sky tech versus things like night vision and lighter plates?
No—hell no.
As a sci-fi enthusiast, my gut says no. But here’s the balance I’m trying to strike in my role, and with the Chief. There’s a linear future for some of the tech we have today. A lot of the drones we’re going to buy are a generational leap from what we had just a year ago—but they’re still consumer-grade. We need the consumer market to help us scale that.
Once we start using them as dual-use tech, we’ll see wholesale, radical changes in how they function.
On the other end of the spectrum is where we make revolutionary changes. I’ll give you two examples: AI and autonomy.
I use AI and ML as buzzwords because they are buzzwords. I’m a computer scientist by training—I understand the power of everything from simple automation to agentic and generative AI. I’m not a non-believer. I’ve used it.
What frustrates me is when we use those terms without asking what decisions we’re trying to augment. Otherwise, people just say, “I’ll use AI to make a bad process slightly faster,” instead of asking how to make a better process.
What decisions need to happen at superhuman scale—meaning you could throw infinite humans at it and not solve it? What needs to happen at superhuman speed—faster than any human can think? That’s where we need to focus.
This is where I get excited. If you imagine a graph: On the X-axis is the leap in capability—evolutionary (what’s already emerging) to revolutionary (what’s possible). Then on the Y-axis is how much we can borrow from commercial markets versus what’s uniquely military.
On the commercial side, you’ve got things like the financial sector’s advances in data minimization and delivery speed. But on the military side, you’ve got jamming, duress, fog of war, sleep-deprived analysts, and shooters who are crushed, tired, dirty, hopped up on Rip-Its and nicotine. That’s the real operating environment—and the tech has to work under those conditions. Those are the different tradeoffs I get to think about.
Let me give you one concrete example.
When we think about drones and quadcopters, commercial markets are already doing that. What no one thought about five years ago was: How do I get a case of ammo to a soldier using a drone?
What we want to do next is pair that with predictive logistics. If a unit’s in a firefight and going to run out of ammo in seven minutes, I need to launch a drone in two minutes to get it there on time. I also need to launch a recovery asset. And because we’ve got a good network, I need to alert their peers in other sectors that something’s happening and they need to be ready.
Those are the automation, logic, and reasoning tasks I know we can do—but we have to lay the foundation first.
It doesn’t sound like you think warfare in five or even ten years will be human-free.
I don’t—and maybe I’m not thinking hard enough. But warfare is a fundamentally human condition.
Yes, there are zones in Europe where you can’t go without being killed by a drone—they’re no-man’s-lands. But if you want to win, someone has to be there. Otherwise, it’s a stalemate.
That said, the way we’re thinking about formations is absolutely machine-augmented. And not in five years—in one year. There should be no engagement the U.S. Army enters where we’re trading blood for blood.
We should be trading steel for blood—our steel, their blood.
The first thing the enemy sees should be a robot. Airborne, ground, doesn’t matter. We should never be sending our boys and girls through the fatal funnel first. It should always be a robot.
What’s the tech—or software—you’re most excited about?
We just awarded the Next Generation Command and Control contract last Friday. Three weeks ago, that would’ve been my top answer. Now that it’s on a glide path, what I’m really excited about is autonomy—because of the unique use cases we’re working on.
Commercial industry is focused on optionally manned, on-road vehicles. What I deal with is: how do I take that vehicle, trained on hardball roads, and make it work off-road—through woods, snow, rocky terrain, marshes?
How do I help it distinguish friend from foe? How do I let it reason in a way that protects our soldiers—or goes offensive, if needed?
How do I take 250 years of U.S. warfighting doctrine—because the U.S. started as a guerrilla force—and imbue machines with that ethos, so robots form the front line before our troops do?
What are the things that keep you up at night?
Everyone else is doing the same thing.
What really keeps me up is the cost of countering democratized technology. It’s a pit. Look at Russia-Ukraine. Ukraine has taken down extremely expensive Russian bombers with cheap drones. It’s become a cliché, but it’s true—and we don’t have infinite money.
Taxpayers deserve the best value for every dollar. So I spend my time thinking about how we build lower-cost options to counter mass: directed energy for drones, lasers for counter-UAS, attritable robots to provide our own mass, and rebuilding the organic industrial base to make cheaper bullets.
It’s not about bean-counting or secretly wanting to be an accountant. It’s about scale. If we don’t generate mass, our soldiers won’t have the kit or capabilities they need to fight the way the U.S. Army fights.
We’ve talked a lot about commercial capabilities. What do you think the role of the primes is in the future fight—and will they have to change?
The primes are going to have to transform.
The good, bad, or ugly of it is that defense primes know how to scale production. If you’re a founder building hardware—something that exists in three-dimensional space—one of your biggest challenges is scaling. There’s a future where founders and innovative thinkers partner with primes.
But I don’t mean that in a predatory way. I actively discourage primes from buying small, innovative companies just because they see them as competition. That hurts everyone: taxpayers, soldiers, the joint force.
There’s a better model where small startups—VC-funded, PE-funded—scale with partners. But the primes need to understand: The era of the government blindly paying every bill and taking on every risk is ending.
They expect the government to take on all the financial, policy, and functional risk. We’re pushing back.
In areas like munitions—where costs haven’t come down in decades despite the tech staying the same—we’re investing in other competitors. Not just to compete, but to send a message: We’re not doing things the same way and expecting different outcomes. We expect on-time deliveries and full value for every dollar.
If you fail to deliver, we’ll do two things: make it public and do something different.
What do you think defense tech companies, founders, and VCs are doing right—and what are they doing wrong?
That’s a good one. I’ll give you three: what they’re doing right, what they’re doing wrong, and what we are doing right and wrong.
On the “right” side, defense tech is doing a great job using their networks to get in the door. That’s huge. I don’t know every company out there—there’s no way to—so networks help.
On the “wrong” side, venture doesn’t always work in the DoD. We don’t pay that quickly. We don’t have disposable income or discretionary funds to throw at novel things. We’ve asked Congress for more agile funding to address that, but companies need to understand: our budget isn’t really ours.
For example, in FY25, the Army’s budget was about $186 billion. That’s the smallest of the three departments. We’re about 21 percent of the total. Of that, 55 percent goes to pay—military and civilian. Another 25 percent is congressional direction—things we can’t change. That means only 20–25 percent is available for transformation.
So when you see radical tech innovation from us, it’s happening in that sliver.
What we’re doing better now is building front doors for smaller companies. We’ve revamped our SBIR program to put out useful topics. We’re connecting companies directly to units, not just giving them a Pentagon meeting.
The unit can give real feedback—fast. That kind of feedback might have taken years in the past. If your product works or is on the right track, you’ll know. If it doesn’t, you’ll know that too.
And what is the Army doing wrong?
There’s still too much fascination with process and requirements over documentation.
Anyone who knows how the DoD does capabilities knows we tend to write long, buzzword-laden requirement documents. Everyone gets a vote. Then two years later, we finalize it. Then, a year after that, we look for funding. Then five years later, we deliver a capability—and the user says, “I asked for this three jobs and four ranks ago.”
We’re trying to change that. Our new command-and-control system? The requirement document was three and a half pages, not 300. That’s a huge shift. We’re trying to institutionalize that mindset.
Legislation like the FORGED Act in the Senate or the SPEED Act in the House helps. But honestly, I think we can go faster even without new laws.
Congress and the Pentagon keep talking about reforming acquisition. Is it enough? What else has to change for new tech to actually get in the door?
The number one thing we need is flexibility in funding—to actually change how we do business.
Right now, we have budget line items and program elements. Back in the 1960s and ’70s, the Army had about 11 of them: vehicles, ammunition, soldier systems, that kind of thing. Now we have thousands. These lines are so high-fidelity that they guarantee payment for specific items—but they also prevent us from stopping things that no longer make sense.
That kills flexibility.
We’re not asking for a slush fund—even with the best intentions, that goes badly. But we do want flexibility in areas like counter-UAS, electronic warfare, software, and munitions—anything that moves too fast for five-year planning cycles.
And do you see signs of movement toward that kind of flexibility in Congress?
Yes. I’ve been on the Hill. Gen. George has been on the Hill. The Secretary has been on the Hill. We’ve briefed every relevant committee and made our case.
I think we’re moving in the right direction. We requested flexibility in our FY26 budget, and I think we’ll get it in a few areas—unmanned systems, counter-UAS, EW. Then we’ll prove to Congress that we can be a good customer and partner, and I think that will lead to broader change.