Podcast: Why Amazon Could Be The Next Big Defense Prime

The differentiator in military programs is shifting from platforms to software. Defense industrial specialist Steve Grundman joins our editors to explain the generational change—and how it will transform industry. 

Don't miss a single episode of the award-winning Check 6. Subscribe in Apple PodcastsGoogle PodcastsAmazonAudible and Spotify.

Discover all of Aviation Week Network's podcasts on our Apple Podcasts channel or aviationweek.com/podcasts.


Rush Transcript

Joe Anselmo:

Welcome to this week's edition of the Check 6 podcast. I'm Joe Anselmo, Aviation Week's editorial director, and editor-in-chief of Aviation Week and Space Technology Magazine.

The competitive dynamics in the massive US defense industry are shifting from hardware and bending metal, to software encoding, and that evolution could open the door for some dramatic changes in the defense industrial base. That is the view of our special guest today.

Steve Grundman is a senior fellow at the Atlantic Council Executive in Residence at Renaissance Strategic Advisors, and runs the executive education program in National Security studies at Syracuse University. He's also a Pentagon veteran having served as Deputy Under Secretary of Defense for Industrial Affairs & Installations during the Clinton administration.

Joining Steve for what I'm sure will prove to be a stimulating discussion are two Aviation Week senior editors. Graham Warwick is our Chief Technology Editor, and Jen DiMascio leads our team of defense and space writers. Steve, I wanted to jump right into this recounting when you and I were sitting down outside at the Paris Air Show and the F-35 fry was roaring overhead and you pointed up and said all these hardware platforms here are great and really impressive, but they won't be the differentiators in the next big war. What did you mean by that?

Steve Grundman:

Well, I guess the particular impetus for that comment was a reflection on a project and a resulting report that we had done at the Atlantic Council over the winter. This was a report written by TX Hammes entitled "Game Changers." I strictly forget the subtitle, but something along the lines of, "What should we be learning from the war in Ukraine about, in this case, ground warfare?" And Tacair does not figure in the game changers of that war. What figures in the game changers of that war, according to TX, and with whom I agree, in this regard were, well, let's start with, for example, the sort of proto-JADC2, joint all domain command and control system, that the Ukrainians seem to have conceived and implemented, that create battle command and fire control solutions at about the pace that I think the great ambitions of the Americans, JADC2 would be, just as an example.

Joe Anselmo:

And Steve, one other thing you were talking about is that you're predicting that in the next 15 years, the list of top Pentagon contractors is going to get some new entrants. You named Amazon as one possible example. Tell us about why you think that.

Steve Grundman:

Well, yes, as you and Jen at least know, I beginning in 2021 was writing in your magazine about what in various ways I called the waves of change, quoting from the eminent Richard Rumelt. There are waves of change that I could sense coming to the defense industry. Indeed, although it was 2019 when the merger of United Technologies and Raytheon was announced, that for me was a pretty sharp leading indicator of change, an inflection point, if you will, or at least one of the big contributors to an inflection point in what it was going to take to be a successful defense contractor going forward. Among the observations that I started making then was the value of big balance sheets, the value of this back-to-the-future reintegration. For a guy as old as me, we can remember, I think Graham too, can remember when the normal structure of the defense industry had lots of multi-industrials in it, back in the Cold War.

That was really one of the hallmarks of the strength of the defense industrial base back then. Anyway, I saw the Raytheon Technologies merger as a back-to-the-future indicator that, at least from a balance sheet perspective and a capability perspective, was going to be one of the comparative, at least competitive advantages that the next successful structure of the defense industry would exhibit. And so, yeah, among the other of the indicators of change coming is the place of entrance in the marketplace. Let's go back to Ukraine again. What maybe arguably is the single most important technology that's been deployed to great effect in the Ukraine war? Well, it's Starlink, it's SpaceX Starlink. In other words, who's running the backbone of battle management for the Ukrainians? Is it one of the familiar names in US or European defense and C2? Nope. Turns out it's an entrant, what I would regard as an entrant, SpaceX Starlink.

The other example that at least comes quick to my mind is the reentry of GM defense into the market. One of those great multi-industrials, General Motors, that is to say, which was a really important part of the US defense industrial base. Indeed, all of the automakers were back in the Cold War. Well, there's another ironic, perhaps re-entrant, let's call them to the defense market, and then there's the whole class cohort of venture financed entrants in the market. Prominent among them certainly would be, let's say, Anduril, just to pick one, but there are many others, particularly in the space sector.

Joe Anselmo:

And Steve, when Jen DiMascio and I were catching up with you before this to chew on some ideas, you laid it out in pretty simple terms. You had said the last great inflection point in this industry was in 1993, the Last Supper, when all the major defense CEOs were called to the Pentagon and said, you guys need to consolidate, the Cold War's over. This industry has to get a lot smaller in terms of the number of companies, which obviously led to the Northrop Grummans and the Lockheed Martins and such. You view Raytheon Technologies as the last shoe to drop in that, and in your view, we're entering a new inflection point now, correct?

Steve Grundman:

I do. Let's call the Raytheon Technologies merger the pivot point. Maybe it's the last of the previous generation of the defense industry structure and/or the leading indicator of the next.

And the color I would put on that is that in 1993, again marked by the well-regarded Last Supper, the Pentagon wanted industry to work on a particular problem, and that was cost. It wanted to ring out from, it wanted to create incentives and room, from a regulatory point of view, give dispensation to mergers that would allow the industry to ring out from the aggregate of the Cold War industrial structure, excess capacity. That was the dominant problem that the Pentagon was asking in many different ways, not even just at the Last Supper, but just more generally in its contracting practice and otherwise. That was the problem it was asking them to work on.

The new problem, and let's mark the Secretary Mattis 2018 National Defense Strategy, at least as the indicator of the new problem, the new problem that the Pentagon was exhorting industry to work on was innovation, and that's quite a different problem in terms of the behaviors that it was trying to induce from industry, but that, and obviously overlaid by the Indo-Pacom challenge, let's just call it maybe even euphemistically, is the new problem, if you will, or at least the priority problem that the Pentagon is asking industry to work on.

And so we see that. And again, that has a different effect. I think it will. These inflection points don't play out in a year. They play out in a decade. The full restructuring of the Cold War industrial structure actually, by my clock, ran all the way to the Northrop Grumman combination with TRW in 2002. So it takes about a decade, at least it has historically for these inflections to play their way all the way out in the industrial structure. And so too, what I expected, again, this inflection to play out, not over a year, but really over a decade, but by the end of this decade, I think we're going to see, again, not a mystery, but we're going to see a notably different structure of the defense industry, by which I mean both different players in the top 10 or 20 of the companies, but also different structure to the existing players.

Joe Anselmo:

Graham Warwick, when Jen and I told Steve you'd be joining us, he was quite enthusiastic and he planted a question with us. He said to ask you, how is Steve wrong?

Graham Warwick:

I'm not sure I would say wrong. Okay, I have several thoughts as I'm listening to this. One is, if you're in the defense industry and you're not already a software company, you are in trouble because it is all about software. But I would also make a distinction between a software company like Amazon or Google, and a software company that's at the pointy edge of defense. So if you go back to, gosh, I don't know how long it was ago when there was all this thing about "it's not about the platform stupid, it's about the network," and everything like that. Here we are like 20 years later and it's still about the platform. We still need a vector with which to apply force. We need to, even if it's electrons we're getting into the battle space, we've got to get those electrons into the battle space somehow.

So we talked about Ukraine and about the challenges with tacair in Ukraine, and we're also talking about a ground war, where it's hard to actually envisage a ground war in the Pacific, because there's not a lot of ground where the battlefield is, but what is being used hugely is drones, from cheap throwaway drones for artillery spotting to longer range drones. It's still a platform. It's still a pointy bit of the spear that's applying something to the battlespace.

I think when I now look at the "it's not the platform, stupid," type thing of 20 years ago, we had become so focused on the platform that everything was tied to that procurement of that platform. You had to get the program office to buy into anything and everything before you could get anything into the defense structure. If you were the F-15 program manager or the F-16 program manager, you ruled all in terms of what the defense instruct base could be.

You had the money, you could buy things. You had a platform and a program office, and I think what they're beginning to successfully do, still have the platforms, but they're actually making them less the focus of where all the money goes and they're trying to get it onto the pieces that those platforms will need or will connect to, but they're disconnecting it from having to have a major platform acquisition program as the reason that they're doing something. So as we go forward, that opens up all sorts of things. We are seeing this open system thing beginning to get some traction. That opens up to new players who could be completely software with no hardware at all. It just does in effect the capabilities that it's purely a software function that becomes a business, but that's all they do is just apps that go onto things like that.

They don't have to be a Raytheon with huge depth of industrial capability. They just write apps that go run on something. But at the same time, I think that's the way that the big guys are going to go. We'll never get rid of the physical bit of the cyber-physical, but the physical becomes almost impossible without the cyber bit. There is almost nothing on an airplane now that doesn't have some bit of smarts attached to it. That means that code gets written for every rivet that goes into an airplane. Almost. So go into it. Yes, I think we'll see new players at the pointy edge writing apps that go onto things. We will see the traditional players become way more like software companies, but we still need a physical vector to get the effect into the battle space, and therefore I think we're going to spend billions on NGAD. NGAD probably won't operate the way that we think a fighter operates, but it's still a platform to project effects into something.

So I think we're going to need platforms, but it's going to become increasingly cyber-physical and the cyber part is going to be the biggest piece of any program going forward. I would point out the automotive industry. If you go buy a Mercedes, it's got something like a hundred million lines of code. The money that goes into Mercedes developing a new car or BMW or anything like that, the bulk of the money they spend on developing that car is writing the software for that cart. It's not bending the metal to make the nice shiny shapes. It's the software that goes into the car, and that's the way this industry's going.

Joe Anselmo:

I wanted to bring Jen into the conversation and then Steve we'll circle back with you and let you respond to what Graham and Jen say. Jen, your team is covering the Pentagon on a daily basis. Is what you're hearing from Steve and Graham aligning with what they're telling you?

Jen DiMascio:

Yes, absolutely. The Pentagon is still spending billions upon billions of dollars to create large platforms, and they are at the same time increasingly turning towards software development. I think the challenge, and what I'm interested to hear from Steve is, how they bridge the gap of having so much metal and so much hardware and embedding that new software. How do they bring the old stuff into this new era of warfare? If you look at what's happening in Ukraine and you think, well, Palantir and Anduril have really brought great software to bear a way to get into the electronic warfare loop of the Russians and show some air launched effects there. That's happening on a really small scale.

And so at the same time you have a high-tech enabled trench warfare. It's on the one hand really far-reaching, and on the other hand, it's the same thing that we saw in World War I. So it's changing the game, but it's also more of the same. How do you change it further and can a high-tech enabled joint strike fighter help you do that? Then the question is how do you embed those technologies into an F-35 and when we've seen all of the problems just getting from block three to block four to upgrading the software programs on our legacy tacair equipment. So I guess those are the questions I'd like to know from Steve, and also how do you do that within the industrial base that exists?

Steve Grundman:

Yeah, let me make a response both to Jen and to Graham, and that is to say essentially, I agree. Indeed, not only do I agree, but what each of you is saying strikes me as, to borrow, again, a phrase from the title of my friend Richard Ramel's book last year, the crux. This is the crux of the challenge which this evolution or inflection presents both to the Pentagon and to corporate strategists in the defense and aerospace industry, and that is to say, maybe putting it a little bit more bluntly than it will be felt in those places, how do we reduce the cost of the metal, so that we can invest relatively more into the code? And also, I would add, into the CONOPS.

How can we make that trade-off in this industry, by which I collectively mean the buyers and the sellers that don't have those habits, that have all of these big muscles for building exquisite hardware, for building exquisite metal. To take that metaphor, that's the crux of the challenge it feels to me like for, again, both for the buyers and the sellers in the defense industry, to make this or at least accelerate this transition to putting the really high value coding, which I mean both literally, but also to some degree figuratively, at the top of the hierarchy of value rather than somewhere in the middle or at least in second seat at best.

Graham Warwick:

I'm trying to think how this would play out in Next Generation Air Dominance (NGAD), which is really is the biggest platform out there. I'm just trying to visualize the balance between hardware and software, old hardware skills and new software skills in developing something like NGAD, and how do you get the cost balance. If you're a company, where's your investment balance? If you're looking at something like NGAD. The fact is NGAD, CCA, collaborative combat aircraft, any platform that we envisage, there's new airlift platforms being looked at. There's new refueling platforms being looked at. There's new air defense platforms being looked at. To some extent, they're just airplanes. So there are some fundamental skills that we need to have if we continue to build aircraft. We need to understand how landing gear work. We need to understand how flight controls work. We need to understand aerodynamics. We need to understand propulsion and things like that.

Those are fundamentally hardware skills. Now, they're being massively digitalized through computational design, so you still end up with a piece of hardware, but it didn't come around by the same way as it used to do. It comes through a digital process that refines, optimizes, and then eventually you just push a button and it gets produced. But I honestly think if you look at NGAD, that's the real hardware pieces of NGAD are going to be the least important bits of it. It's going to be a composite airframe. It's going to have stealth characteristics. It's going to have some sort of advanced type of flight control for both performance and safety and various things like that. The biggest part of NGAD is going to be the software environment that it lives within, and that's the software environment inside the platform itself that enables it to be a platform, that enables it to do its job, and that then how it connects to the larger environment around it.

And again, I think it gets back to the analogy with the car. I forget what the F-22 ended up like. It ended up with I think about 12 million lines of code. It might be up from that now, but cars are a hundred plus million lines of code. NGAD is going to be at 1, 2, 3 orders of magnitude more lines of code running on the platform and its connected ecosystem. That's what the companies are going to be doing. That's what they're doing now. Yes, they have to keep their skills, to keep the wheels on the bus, sort of thing, as it goes down the runway, but they also have to realize that its ability to do its job is not in anything physical on that airplane. The ability to do its job, as long as it can avoid being hit, which is a design issue, but once it's in the air, its entire ability to do its job is going to be defined by the software on board.

Steve Grundman:

I think that that broaches a question in the industrial structure about. At least a question, I really don't know the answer about. Whether we keep calling it software, why not, and hardware are integrated in individual companies. Is that the efficient structure of industry or are they disintegrated? Do the guys who know landing gears and the physical features of the vectors, does that become a specialized part of the industry?

And those who know processing, sensing, integration, network become a different part of the industry? Back in the early aughts, in maybe the second wave of the post-Cold War restructuring, you all remember TSPR, total systems performance requirements, or integrator. There was this idea, and I think Boeing, it might be said, was the most prominent company that took a flyer on this idea of the pure system integrator and all of the manufacturing and creation was done by subcontractors. I think in the end, I think Boeing wouldn't deny that experiment failed, but maybe it just was before its time. And in terms of the structure of the industry, say 10 and 15 years from now, maybe that the software and the hardware actually will be disintegrated. Maybe. These are the, if I may, provocative ideas I hope corporate strategists are playing with.

Graham Warwick:

That does actually gel with, because if you look at what we've done with software, just, again, I'm a platform guy. I'm coming at this with a very specific viewpoint. But if we look at how we've gone so far, we always go too far. We always go too far. Every time, we always go too far. With the F-22, we went highly integrated. Literally, you couldn't pull the boxes apart. You couldn't touch anything without touching everything on an F-22. And it's a major problem with that airplane trying to keep it updated. But what we did with the next generation is, and we've done increasingly is, we've separated the vehicle into two pieces, the vehicle management system and the mission management system. And I could argue, and if I was to be given a chance to argue, I would argue that the vehicle management system should be the total responsibility of the prime contractor and should not be an open system.

It should be the very best optimized proprietary system that that OEM can come up with. The mission management system should be as open as we can possibly be because that's where for the next 50 years, that's where we'll have to make changes to that platform. We are not going to go in and change the tires. We're not going to go in and put new coffee cup holder in the cockpit for the pilot. We are going to go in and put directed energy weapons. We are going to go in and put new EW effects. We are going to go in and do other things like that. So I would say if you are a traditional aircraft prime, I would say your role is to come up with the very best closed platform vehicle management system and vehicle and the most open mission management architecture so that you can then, once the customer's got it, he can't touch the vehicle without coming back to you, because you know the vehicle better than he does.

But the customer can come in and say, today, I want to put the ability to drop blue painted bombs on people rather than green painted bombs on people. So I want to change that software, without having to go back to the prime. And I think that's part of the reason that we go into this whole "it's not the platform stupid," was because the pen was billing to realize they were always having to go back to the prime every time they wanted to do something and they wanted to kill the platform to kill the prime. I mean, that's an exaggeration. They wanted to break the lock by breaking the idea. They needed just platforms. We still need platforms. We just need to break the lock.

Steve Grundman:

And if we wanted to put a new sensor on the platform, we didn't also have to want to change the tires. I particularly like this juxtaposition. I think it's cleaner and more evocative between mission management and vehicle management. That's better. Frankly, that's more helpful to our thinking about this than software/hardware, which is I think freighted with all sorts of ideas about what one might mean by those things. Mission management, vehicle management, that's a cleaner distinction between these needs, that I think are in play in terms of how we organize them industrially, whether entrants are going to claim some of those roles, whether the legacy companies will step into and up into those roles. Well, one hardly should doubt, not least with the cashflow that they have to invest that these companies could change, that these legacy companies could change. And indeed, there's plenty of indications already that they plan to do just that. So it's interesting time to be a strategist of these things. Acquisition strategist in the Pentagon, corporate strategist out in the industry.

Jen DiMascio:

Well, it is interesting. Steve Trimble, our defense editor and I were discussing this yesterday. What is the idea of bringing more, essentially, apps into an aircraft in that mission management system? And a challenge that he pointed out in doing that is, really, how do you contract for it? Do you take a page out of the book of commercial satellite operators and charge some kind of service fee? Because when you're dealing with the government, they tend to buy one product in bulk, and it's one development and that's it. It's not a recurring payment or a recurring continuously upgradable system. And so I think the Pentagon will have to rethink how it handles all of that, or maybe it already has when Amazon Web Services gets in on the deal.

Joe Anselmo:

Unfortunately, we are out of time, so we're going to have to end on that note. But Steve, thank you for coming on to share your insights. We're definitely going to have you back, because I think there's still a lot more to talk about. That is a wrap for this week's Check 6 podcast. A special thanks to our editor in London, Guy Ferneyhough. Thank you for your time, and be sure to join us next week, when Graham will be back to talk about how the advanced air mobility sector has reached a critical juncture. Until then, have a wonderful week and stay safe.

Joe Anselmo

Joe Anselmo has been Editorial Director of the Aviation Week Network and Editor-in-Chief of Aviation Week & Space Technology since 2013. Based in Washington, D.C., he directs a team of more than two dozen aerospace journalists across the U.S., Europe and Asia-Pacific.

Graham Warwick

Graham leads Aviation Week's coverage of technology, focusing on engineering and technology across the aerospace industry, with a special focus on identifying technologies of strategic importance to aviation, aerospace and defense.

Jen DiMascio

Based in Washington, Jen manages Aviation Week’s worldwide defense, space and security coverage.

Steven Grundman

Contributing columnist Steve Grundman is the principal of Grundman Advisory and a former deputy undersecretary of defense for industrial affairs.