🚨EMERGENCY POD! The Department of War (DoW) has officially released a new AI strategy (note, this is my first 2026 prediction to come true, just 2 weeks into the new year), and along with it a whole slew of new reforms to the Defense Innovation Ecosystem. In the newest episode of the Techquisition Edition of the Mission Matters podcast, David and I break down the DoW’s new AI strategy, the major changes to the defense innovation ecosystem, and what all these new initiatives mean for startups.
We discuss:
What it means for DIU and SCO to be designated as “field activities”
What new major AI projects the DoW is starting
How DoW is investing in data and compute infrastructure to enable rapid AI adoption
New initiatives in the services and PAEs designed to bring innovation into the hands of warfighters
And more!
You can listen to the podcast on Spotify, Apple, the Shield Capital website, or right here on Substack.
As always, please let us know your thoughts, and please reach out if you or anyone you know is building at the intersection of commercial technology and national security.
And, we encourage anyone interested to read the AI strategy and innovation ecosystem memos for yourself to get all the details on these policy changes:
Transforming the Defense Innovation Ecosystem to Accelerate Warfighter Advantage Memo
Transforming Advana to Accelerate Artificial Intelligence and Enhance Auditability Memo
Transcript
Maggie 00:39
Welcome to the second ever episode of the Techquisition Edition of the Mission Matters podcast. In this episode, we’re going to be discussing some pretty big news that came out just yesterday about the Department of War’s new AI strategy and their new plans for the DoW innovation organizations. Just yesterday, Secretary of War Pete Hegseth announced a series of reforms for the DoW innovation ecosystem during a speech he gave at SpaceX headquarters in Starbase, Texas. He also announced the DoW’s new AI strategy. In conjunction with the speech, the department also released three new memos codifying these initiatives. So David, just to kick it off with the first question, can you cover what were some of the major themes highlighted in this speech and in these memos?
David 02:07
Well, Maggie, it’s a pleasure to be back on with you for this Techquisition Edition emergency podcast. Yes, we’re recording the day after Secretary Hegseth’s speech at SpaceX, which is a pretty cool venue to announce things at. But what did he talk about? He was re-emphasizing the desire for the Pentagon to go fast, find a way to be able to say yes, not be overly reliant on process, grow the industrial base by being creative and leveraging private capital markets. There was also a significant emphasis on leveraging artificial intelligence and on behavioral changes around data sharing, experimentation, and making the department AI-first. Secretary Hegseth also announced new leadership at the Chief Digital and Artificial Intelligence Office, as well as at my alma mater, the Defense Innovation Unit. He created lanes of understanding for how organizations like the aforementioned CDAO and DIU, as well as DARPA, the Mission Engineering and Integration Activity, and the Office of Strategic Capital, among others, are to work in tandem together to support the mission, which, quite candidly, can be confusing at times if you are unwitting to the variety of department organizations.
Maggie 03:32
I know I’ve had to have you explain to me multiple times how all these organizations actually differ from each other.
David 03:38
Yeah, well, you know, it certainly doesn’t help that the acronym soup is always expanding. But to that end, he emphasized some consolidation. We had these innovation organizations that were not tied directly to executing missions or acquiring capability. To that, he called out specifically the Defense Innovation Working Group and the Defense Innovation Steering Group. Gosh, I didn’t know that those were actually different, unique entities, and I’m an interested outsider. I couldn’t tell you what they did. He abolished those in favor of consolidating and then sort of putting the onus back on the services to consolidate as well and figure out how they’re going to engage with this series of memos. Overall, I’d say it’s a similar pattern to what we’ve seen. He is highlighting things that seemingly are working and trying to inculcate them into the system, versus allowing these things to continually operate on the periphery.
Maggie 04:42
So David, speaking of DIU and SCO, or SCO, Strategic Capabilities Office, can you tell us a little bit about what you think of those organizations being named field activities? I mean, what is a field activity? I know that was really one of the major reforms that people have kind of attached onto. Is this new? Is this substantial? Why does it matter?
David 05:04
Yeah, well, and I’m glad you mentioned the Strategic Capabilities Office, because I forgot to mention them as one of the organizations he referenced earlier. Okay, so field activity. Does it matter? Is it new? Field activities, there are lots of them. They are sort of an echelon lower than an agency. You have the departments, the Department of War, the Department of the Air Force, which is where the Space Force and Air Force as branches reside. Under the Department of the Navy, you have the Navy and the Marine Corps as branches, and then the Department of the Army. These are secretary-level organizations, and then you have these things called agencies. DARPA is an agency, DITRA is an agency, and then you’ve got the National Security Agency, which is also an agency and an intelligence community organization, so they are dual-hatted to the Office of the Director of National Intelligence and the Department of War. But what you might call less substantial, though still chartered to do a specific service on behalf of the entire department, are these things called field activities. I’ll actually give a commendation to my good friend Kevin McGinnis, who I worked with at Defense Innovation Unit back in the day. He wrote a paper stating that Defense Innovation Unit should be a field activity all the way back in 2017, and he was quick to remind me of that this morning over a message. We talked about it and asked, what does that mean, and why is it important? I would say for the most part things will stay the same. It does put the onus on Defense Innovation Unit and Strategic Capabilities Office to fulfill the mandate as articulated by Secretary Hegseth, and he’s going to submit that through the appropriate Title authority to Congress to ratify it. But now they are, lock, stock, and barrel, supposed to do exactly what these memos highlighted, and the department, or other organizations within the department, cannot or should not do that work. A good example of a field activity is Washington Headquarters Services. Most people don’t really know who they are or what they do. They support from a back-office perspective, contracting personnel among other things, and they provide that service for all of the Office of the Secretary of War’s needs. Whether you’re in Research and Engineering or Acquisition and Sustainment, and you’re one of those principals tied to doing something, if you need a contracting entity or contracting activity, you must use Washington Headquarters Services. That’s what they’re mandated to do, and that’s what a field activity is getting at. I would also say it will provide those of us interested in the financial machinations much better transparency into what Defense Innovation Unit and SCO are submitting through the President’s Budget. They’ll also now have to defend it. It won’t be locked in with all of the other Office of the Secretary of War activities where you have to figure out after the fact what DIU spent money on. That’s interesting. They’ll have to fight for it through the OSD Comptroller and the Office of Management and Budget, but it will help us understand what they’re spending to execute their mission. Hegseth mentioned that DIU will still have the ability to report to the Secretary, as codified in the Fiscal Year 24 National Defense Authorization Act, which was really important at the time. But their direction is going to come from the Under Secretary for Research and Engineering, Emil Michael, who now also appears to be dual-hatted as the department’s single Chief Technology Officer. For what it’s worth, a lot seems to be getting put on the plate of the freshly minted CTO, and I think we’ll get into that a bit more with the artificial intelligence transformation memo.
Maggie 09:19
Yeah. Yeah. That seems like a major point that Secretary Hegseth made, which was that all of these innovation organizations are now going to be reporting into this single CTO, Emil Michaels, who is really bringing all these organizations and their missions together. So according to Secretary Hegseth’s speech, one of the themes that I really took away is that innovation really can’t just be for innovation organizations like DIU. It needs to be part of warfighters and operators. It needs to become a core part of the way that the services departments and acquisition organizations, like the new program acquisition executives, do business. And there were maybe two major initiatives that he announced as part of that. The first one was this change to PAEs, this innovation insertion increment, or triple I requirement, that he mentioned. So could you tell us a little bit about what that actually means?
David 10:19
Yeah, well, I mean, I’ll say maybe to the first part about where is innovation supposed to happen. You certainly have these organizations where it’s expected to sort of be a front door for commercial entities. But I thought it was really cool that Secretary Hegseth sort of called out that innovation can happen at any echelon. He regaled an anecdote with a captain in the Army who was working on some AI transformation and saying, like, we need to be celebrating these activities and encouraging people who see problem sets in different ways and solutions to go out and do it and be encouraged, and then the command structure ought to support it. But specific to this innovation insertion and increment move, I’ll say off the bat, I think this is going to be really tricky to do, and a lot of that comes down to the budgeting process, known under the whole planning, programming, budgeting, and execution, PPBE. Maybe for the purposes of this conversation we’ll truncate it to how it’s colloquially known in the department, and that’s the POM process, Program Objective Memorandum. The POM process forces program managers to outline with extreme specificity exactly how the funding requested from Congress will be spent two years in advance, to include who the performers are going to be and what contract vehicles they’re going to exercise and leverage well in advance. And so how can the PAEs, across their portfolio, which we assume to be a grouping of related programs, be provided enough flexibility to incorporate innovation? This is going to be a challenge. Today, program offices can use programs like the Small Business Innovation Research, SBIR, or the AFWERX program. They also sometimes leverage the Defense Innovation Unit and some of their funding lines that are a bit more flexible, and then SCO, to do some of these innovative, non-programmed activities. So I think the department will need to negotiate with the appropriators off the bat what sort of a percentage, maybe five or ten percent of a program that is being POMed for, could be available for non-specified experimentation. And so I think it’ll be interesting to see how this plays out and how the services respond. But I do think that there’s an element of funding flexibility that is required in order to get after this mandate.
Maggie 12:55
So just to make sure I’m understanding this right, the way we should think about this triple I initiative is that it’s really trying to force the PAEs to spend a specific percentage of their budget on trying to integrate some innovative new technology into an existing program. So I need to spend some money to put AI on the F-35. But what you’re really getting at is that the way money is spent on a major program today has to be very detailed and submitted to Congress far in advance, and so we’re going to have to see how they’re actually able to get this kind of flexibility. Am I understanding what this initiative is the right way?
David 13:37
I mean, it’s almost as though you could ask me, you know, “David, here’s the NFL schedule. Predict every single team that’s going to win based on paper,” and not be able to take into account injuries or crazy random acts. And if you’re not right, there’s nothing you can do during the season in order to change the plan for how you thought it was going to play out. And I think what Secretary Hegseth is getting after is that the velocity and the pace of technological change are so fast that if we’re not taking advantage of it while the program is still in development, by the time it’s fielded, it’s already going to be outdated. And so if we don’t have these opportunities for innovation to occur, to then insert it and integrate and scale, because we’re not all-knowing, you know, we’re setting ourselves up for failure.
Maggie 14:40
So next, I want to turn to another one of his initiatives that he outlined, both in his speech and in these memos relating to the services. That is, he has mandated that the services need to brief plans for innovation within the next 30 days. So what should we take away from this mandate?
David 14:58
So, just like the Office of the Secretary of War started to consolidate a lot of these organizations or councils or steering groups, he is sort of asking the same of the services. And so I would say that, in some respects, it feels like the services are out in front of this. We’ve seen announcements around the Army FUSE program, and they’ve consolidated a lot of these seemingly related but not connected organizations that are doing innovation, experimentation, and scaling, and that’s now all under one house. The Navy is doing a road show with their new Navy Rapid Capabilities Office, and that seems to be a strong insertion point for innovation to get out to the fleet. They’re doing it in concert with the Office of Naval Research, which has historically been responsible for a lot of early-stage technology readiness level experimentation and working across a variety of national and government labs. And then Space Force, for a couple years now, has had this organization called the Commercial Space Office, COMSO, and they’ve been doing a really good job working with commercial organizations and figuring out how they can apply that to their mission mandates.
And so I guess the one service I didn’t really mention is the Air Force. They have a pretty famous innovation organization called AFWERX, but of late it’s been a little bit unclear exactly how that’s being tied into larger programs like Collaborative Combat Aircraft, which is pretty forward-leaning and leverages innovative companies. How is the Air Force going to be adopting this mandate to be more innovative? I think it’ll be really interesting to see how the services submit their plans. I really hope that, at least in some respect, we can see that at the unclassified level and provide some sort of comment, because one of the things that Secretary Hegseth reiterated is just how confusing and challenging it is for industry to understand all of this. So really what we ought to do is accomplish this transparently and make sure that we’re getting feedback from industry, so that we’re not just recreating the same thing as it was before.
So Maggie, maybe switching gears here, in addition to all these changes across innovation organizations and the need to go faster, Secretary Hegseth announced the Department of War’s new AI strategy. I think you’ve had some time to absorb it, but what were some of the major initiatives and themes from that strategy that stuck out to you and that our founders really need to know about?
Maggie 17:58
Yeah, definitely. I was excited to see a new AI strategy from the department. The last strategy the Department released was back in 2023, and if you can believe it, it did not mention generative AI a single time. Times have definitely changed since then. And I think, as you know, I’ve written about and we’ve talked about in the past that there are a lot of changes the department needed to make to really stay competitive in the AI space, as the pace of innovation has moved so quickly in the commercial sector. Overall, the strategy is focused on how the department can quickly adopt AI at speed to become an AI-first warfighting force. They’re trying to ensure that the entire department has access to leading-edge models on all of its networks, both classified and unclassified. The strategy starts out by detailing what they’re calling seven pace-setting projects, which are essentially, as far as I can tell, pilot initiatives to get the ball rolling on AI adoption within the department. Each of these is going to have a single, accountable leader who has to regularly report progress on aggressive timelines.
David 19:14
What would be a pacing project? Because that does sound pretty exciting.
Maggie 19:19
Yeah, yeah. So they’ve already announced one of these, GenAI.mil. I know, David, you’ve actually had the chance to play around with it, so I don’t know if you want to share anything about what that project is and how it is.
David 19:31
Well, sure. So GenAI.mil, a couple weeks ago when I was doing my reserve duty, all of a sudden one day we had an application on our desktop and we had access to Gemini. When you opened up the application, you could launch Gemini, but it also highlighted that coming soon would be a couple other of the more prominent large language models. And at the event yesterday, AHEG Seth announced that Grok is now available. The other two that I believe I recall seeing were going to be OpenAI ChatGPT and Anthropic. So it seems like the department is getting access to some of these cutting-edge large language models, and it’s really great to see. I will say, during my two weeks of reserve duty, I was using it quite a bit. Because I’m used to using it in my day job, I use OpenAI among others, and I was explaining to my coworkers how they might engage with it and leverage it. So it’s really great to see things that are in the commercial world coming to the department. I mean, I guess three years late, but who’s counting?
Maggie 20:43
I mean, I thought that this was a great initial project for the department to roll out, and they got basically just these basic AI tools in the hands of 3 million folks within the department. They did not overthink it, right? They are just giving people access to the exact same commercial tools that you and I have access to in our day jobs, really allowing people to see for themselves the power of these tools and find for themselves the major applications that are actually going to make a difference in their day-to-day workflows. So that was the first pace-setting project. They have a few more that they have not really released any details on, but they are looking at things like using AI for battle management, decision support, and improving simulation. The next major initiative that they announced is that they are actually going to be investing in more AI compute resources for the department.
David 21:33
I will just interrupt. I mean, that is huge to me, and as somebody who has watched the budgets be developed, it always feels like information technology and infrastructure are always a bill payer. So maybe just a little bit more on how exactly we are going to get to this AI utopia with the resources necessary to make it happen?
Maggie 21:59
Yeah, so I think that really investing in these AI compute resources is going to be one of the major enablers of actually getting this technology into the hands of warfighters, as well as folks in the back office. I have written about this before. One of the major roadblocks to getting AI where it is needed is literally just access to compute, access to GPUs, access to hardware that can run on classified systems or that can run at the edge. You know, some people I have talked to actually talk about bringing GPUs, like an NVIDIA Jetson, just in a suitcase to make sure that they always have it when they need it, because they know that they are not guaranteed to have it. Yep, I have talked to people at our portfolio companies that do this, and I have talked to people in the government that do this. So, you know, what the Secretary has said is that the U.S. military is going to put real capital behind acquiring compute resources. He even went so far as to say that they are going to build data centers on military bases or on federally owned land. We will definitely see what has to come of that. These are going to be secure data centers where maybe we could actually do some AI model training on classified data.
David 23:15
Sounds awesome and definitely needed. I hope that the government is able to strike a deal, because what we are seeing is certain communities resisting data centers, you know, not in my backyard. Well, you know, federal land on military bases, and there are certainly plenty of old, ugly buildings that ought to be knocked down, and we could put some server farms in there. Maybe if the warfighter can get access to it first, that would be wonderful, yeah.
Maggie 23:45
Well, you know, we know there’s a lot of real estate professionals in this administration, so I’m sure they’ll be thinking carefully about where the best spots are to put these centers. The next major piece that I saw in this AI strategy actually harkens back to the 2023 strategy, which was, once again, really focused on data availability for use with AI models. This was the major focus of the 2023 strategy, and the new strategy takes it another step further. One of the major challenges to actually deploying and training AI models in the department is that a lot of data is stored in silos. It’s hard for the Air Force to access Navy data, or even within the Navy it can be difficult for different groups to access data, particularly classified data. And as we all know, our AI systems are only as good as the data they actually have access to. So what the Secretary’s memo does is direct all departments and components within the DoD to release a federated data catalog to the CDAO, the Chief Digital and AI Office, that exposes all their system interfaces, data assets, and other access mechanisms across all classification levels, and it says they need to do that within 30 days. I think it will be interesting to see if this can actually technically be done, just literally whether the technology is there to do this in 30 days. But this is certainly an aggressive mandate to force organizations to get their data into a ready state so it could potentially be used by AI. We also see that the memo directs CDAO to release this data to any user who needs it and who has all the security requirements necessary to access that data. It says that if CDAO wants to deny a request, they have to report it within seven days, and that denial can actually be escalated to the Deputy Secretary of Defense for Research and Engineering, Emil Michael. So it will be difficult for CDAO to deny data requests. This is really about trying to open up these data silos. Related to these data decrees, it also directs the services to accelerate rapid ATO reciprocity, authority to operate reciprocity. I don’t want to get too deep into the details of ATO reciprocity. David, I don’t know if you want to chime in here, but basically this is the idea that if I’ve gone through the full cybersecurity process to deploy my software with the Navy, then I should be able to deploy it with the Army or the Air Force without having to go through that same very rigorous cybersecurity process as a commercial vendor to deploy my software on service networks.
David 26:35
You know, it’s a long time coming. We’ll see if they’re able to exercise this. It seems to have been a thorn in the side, and Secretary Hegseth did mention that there will be no sacred cows, and this would certainly be one I’d be willing to sacrifice. Maggie, maybe just, like, what should this mean for our startups? How should they interpret these announcements around data reciprocity, data access, network reciprocity, network access? I think they maybe mentioned stuff around one of the buzzword bingos of modular open systems architecture, trust and responsible AI. They even, I think the memos talked about Advana. The Secretary didn’t mention Advana in his speech. What does this mean for our startups, or emerging companies that want to do business with the department?
Maggie 27:31
Yeah, well, I think it means that this is the time to really start accelerating your engagements with the department. This is a time when they are actively kicking off new projects to explore how these technologies can be used. I also think this is a time when you can actually help shape the infrastructure that the department is going to need to deploy the applications that startups are building. If you have opinions on what compute resources the department needs, and what data catalogs and data infrastructure they need, or even better yet, if you have a solution to support building out that data infrastructure and that compute infrastructure, this is the time to make your thoughts known. Start engaging with people at CDAO and within R&E. Start engaging with people at each of the services. The service chiefs and combatant commanders are going to start designating AI integration leads; they’ve been directed to do so within 30 days. Find a way to get in front of those people when they’re announced, and then really spend a lot of time with warfighters, with operators, deeply understanding their needs, understanding how your solution can solve a real problem for them, and work your way up from there. Yeah. I mean, David, what do you think? What did I miss?
David 28:47
I think that’s right. I think now is a compelling time. I’d like to maybe go back to talking about the Army FUSE program. They’re doing a series of experimentations with the intent that the things that are working will have on-ramps into programs like Next Generation Command and Control. So I think if you’re a startup, you need to start participating in these activities. I’m not saying that there is a well-worn path to scaling, to becoming like the next Anduril per se, but one thing that I thought was interesting that the Secretary mentioned is that he highlighted two companies that are now a pretty vaunted part of the fabric, Palantir and SpaceX, and said that for them to get a hearing and an opportunity, they had to sue the government, and we need to change that mindset. He also mentioned that the Gen AI application is now available to three million people. While our loyal listeners will definitely be listening to the things that the Secretary said, do not assume that the three million people in the department who now have access to Gen AI will have necessarily heard the Secretary. So I think it is important to take the time to read the memos, and we’re probably going to link them in the show notes, and then make sure you have that with you as you’re engaging with your customers, because they’ll need education about how much this transformation is trying to change the narrative. Again, I would say get active and engage. If you feel like you’re being told no a lot, I would say that people within the Office of the Secretary of War will definitely be hearing their boss loud and clear. If you become the squeaky wheel, it might get that higher headquarters organization to reach down and smack somebody to do the right thing, because these are directives. If there’s one thing that the military is supposed to be good at, it’s taking orders in a very direct, top-down fashion. So I think it’s important to get engaged and to give these new emerging organizations the opportunity to prove you right, that working with the department is the right use of your time and resources, and that you can start solving some of these very vexing national security needs.
Maggie 31:18
So David, what were some of the topics that were not discussed in this memo, in this speech, that you might have expected to be in there?
Top of Form
Bottom of Form
David 31:26
So you certainly covered a lot in a short amount of time. In fact, I was joking with one of my friends that I feel as though Secretary Hegseth sort of recreated all of the greatest hits that a lot of us have been saying for a long time about leveraging private capital, growing the industry base, and being innovative. He talked a little bit about contracting. He talked about authority to operate, okay? But what didn’t he get to? I guess I sort of highlighted a lack of what we’re going to do with funding flexibility, right? That’ll continue to be a thorn in the side of all of these initiatives. So I think we’re going to need to get more creative in the coming year with whoever will listen over at Congress. I was also surprised to hear nothing about personnel clearance, right? How do we get people access so that they can know about things that are happening, so that they can build products that will meet the needs of the warfighter.
Maggie 32:26
It was especially interesting that they didn’t include anything about personnel clearance, because they did make talent such a central part of this strategy. In his speech, the department just announced their new initiative, the Tech Force Initiative, which is really working to get early career technical and AI talent into the federal government in general, but the Department of War in particular, to come help build technology to solve some of their crucial challenges. And the AI strategy also directed organizations to be flexible when it comes to pay and hiring to get talented people in there. And, you know, I think, as many people know, getting through the clearance process is actually one of the biggest barriers to getting technical talent in the door. Totally.
David 33:13
I mean, I’m definitely bullish on Tech Force. I remember when the Defense Digital Service first stood up, and some of the initiatives and the education, honestly, that they provided others in the department around how technology was transforming, or should be transforming, mission sets. I’m very bullish on that. But yes, the personnel clearance and getting people through the pipeline, attracting that talent, is going to be a challenge. Also, the International Traffic in Arms Regulations, ITAR. I know that part of it is with the State Department, but, you know, the Secretary has talked a bit about foreign military sales, and I kind of put those two in a similar bucket of how do we allow our industry base to sell to our partners and allies and make sure that we’re not cutting our nose to spite our face from a market standpoint for the proliferation of some of this technology. So I would have liked to have seen a little bit more there, and then maybe just access to classified spaces. Again, if we’re trying to build trust and communicate, and this touches a little bit on personnel clearance, how are we lowering the barrier to entry to even having these types of conversations in the first place? Again, we need to be growing, and growing the amount of people that are in the tent so that we can solve these vexing problems with people’s imaginative and creative ideas. So that’s what wasn’t stated. But gosh, if there’s one thing we can count on, it’s maybe another emergency pod within the next few months as this administration continues to check through a lot of the things that will help modernize the department.
Maggie 34:51
Yeah, absolutely. There are definitely more changes to come, I have no doubt, over the next few months. I’m curious, David, what do you think our next Tech Position podcast episode is going to be about?
David 35:03
Man, this is kind of like a bit of a sign-off for us, right? Like presuming what’s next. Okay, so gosh, I really hope that we can either talk about an appropriations bill getting funded, and I am going to go out on a limb and say that with that appropriations bill will be a reauthorization for the CIBER program. So we might get two emergency pods within one activity. I’m really hopeful for that. So maybe early February we’ll be back with Tech Position episode three, and maybe four, depending on how we decide to split that baby.
Maggie 35:45
I definitely hope we get both CIBER authorization and a budget passed this year. Well, on that note, David, thank you so much for coming on to record this emergency pod. I know it’s 10:00 p.m. now as we’re finishing this recording, and I really appreciate, as always, learning all of your wisdom on all subjects innovation and acquisition in the Department of War.
David 36:09
I really appreciate it, Maggie. It’s good to do this one in person, so that’s also fun. All right, until next time, until next time.
