Podcast: Explaining ‘Replicator’—A Plan For Thousands Of Drones

Jen DiMascio: Hello, and welcome to the Check 6 Podcast. The Pentagon has announced a Replicator Initiative to build thousands of uncrewed systems. The US Army is planning a major competition to award contracts for the manufacturer of some 12,000 small uncrewed aerial vehicles, a development that is one piece of a larger strategic effort to counter a Chinese invasion of Taiwan.

I'm Jen DiMascio, the Executive Editor for Defense and Space. I'm here with Aviation Week Defense Editor Steve Trimble, Byron Callan, Managing Director at Capital Alpha Partners, and Sam Bendett, an advisor at the Center for Naval Analysis. Thanks, everyone, for joining me. Steve, you wrote this week about this contract in a previously classified initiative at the Pentagon to defend China. What is the Replicator Initiative? And how does that fit into all of these plans?

Steve Trimble: Thanks, Jen. Yeah. So big picture right now, this transition to bigger predominance of drones in warfare has been happening for a while, obviously. But really, we're at this transition stage where, for the past couple of decades, the military has gotten used to operating drones in the order of dozens, or even hundreds, sort of simultaneously in various places within a certain theater, in a war fighting theater.

But now, you're looking at a future where literally there'll be thousands, maybe even tens of thousands of small, medium and large sized drones all operating in all different domains, all at the same time. So we're seeing a lot of concepts and initiatives that are looking at this. That includes existing programs, the Army, and we've got AUSA coming up, so we'll be looking at this quite a bit with the Army. But they've got the Short-Range Reconnaissance program, which is on track to award a contract next spring for about 12,000 small quadcopters, enough to outfit all the platoons in the Army. And then the Army has air launched effects, they have unmanned ground vehicles. Those are all in the process of development and acquisition. You look at the Air Force, same story, collaborative combat aircraft and various other ideas also in space. And then the Navy, unmanned surface vehicles, unmanned undersea vehicles, UAVs of all kinds are at various stages of development.

That's the big picture. Lots of different programs going on simultaneously. So the new thing came out initially, August 28th, Kath Hicks, the Deputy Secretary of Defense, was giving a speech at the Emerging Technologies Institute here in D.C. And this is what she said on August 28th, she said, "We're making another big bet. The latest piece of our comprehensive war fighting centric approach to innovation, it's called a Replicator Initiative, and I want to tell you about it. A replicator is meant to help us overcome the PRCs biggest advantage, which is mass, more ships, more missiles, more people. Since we need to break through barriers and catalyze change with urgency, we've set a big goal for replicator to fuel the treatable autonomous systems at scale of multiple thousands and multiple domains within the next 18 to 24 months."

All right. So several days later, September 6th at the Defense News Conference, Kath Hicks was speaking there. And it clarified this, because there was all sorts of stories that came out after this that they were going to buy thousands of Valkyries, XQ-58s or things like that within this very compressed 18-to-24-month timeframe. And I think she felt the need at least to make sure people understood what it is or what it isn't. So at that conference, she said, and I quote, "Let's be crystal clear, Replicator is not a new program of record. We are not creating a new bureaucracy, and we will not be asking for new money in fiscal year 2024. Not all problems need new money. We are problem solvers, and we intend to self-solve. So Replicator will use existing funding, existing programming lines, and existing authorities to accelerate production and delivery at scale, by exerting leadership, focus and attention on a singular operational challenge and maturing solutions, because that's what ultimately delivers."

And then she said something else that I think really sort of clarifies exactly what is going to happen with the Replicator Initiative. And she said, "Our task through this initiative is bringing leadership across the department, around, it was actually the deputy's innovation steering group table, to help ensure those ideas ripe enough to scale, actually do get scaled by elevating and accelerating what we do in cutting red tape. So they're delivered to war fighters in 18 to 24 months."

So really, what we're talking about is taking a look at all the programs that are already out there, that already have funded requirements, but maybe are not going quite as fast as the services would like, or at least OSD would like. And bringing those to this innovation steering group, which is co-chaired by Kath Hicks and the Vice Chairman of the Joint Chiefs of Staff. And then using that forum, essentially, to find the ones that they think are the most promising and figuring out ways within their existing budgets and their existing authorities to accelerate them to delivery.

Now, obviously, there are some constraints that exist in the system. If you're talking about Valkyries for example, there's only so many engines that Williams can build, the engine supplier for the Valkyrie. So unless you want to spend a whole bunch of money to build out Williams' production capacity or find other companies to augment that and somehow get them the tools and the technical data they need to do this as well. All that takes time, and all that takes money though. So there are constraints that they're living with, but they're trying to find ways to make this work in spite of all that.

There's also another program that came out right around the same time, which was DARPA and the REMA program. And this program, interestingly, didn't formally align itself with replicator, but at the same time, they talked about doing things in 18-month time cycles. And it talked about, essentially, creating a mission autonomy suite that could easily be adapted to thousands of commercial drones all at the same time, and then refreshed with the common software autonomy platform, allowing them to refresh the fleet on a monthly basis within an 18-month period. I mean, that idea is basically taking thousands of commercial drones that are already out there, we're talking like quadcopters and things like that, things that are not very sophisticated at all, and then giving them some sophisticated autonomy capabilities that the military has developed, sort of overlaying that on their own control system. So you could see how a program like that would really enable this new vision.

And then part of that is something we have also reported on is the demand driver from the Combatant Commands, particularly Indo-Pacific Command, which has this big challenge of trying to deter China from taking Taiwan back by force, which is the big fear of the next decade, really, is, "How does the US military try to deter that along with its partners in the region?" And the concept that has been articulated now by Admiral Aquilino, also at that ETI conference in August, is this concept called Hellscape.

So we don't know a whole lot about that, but from Aquilino's remarks, what we do know is they want to take the concepts they've been experimenting with, particularly by Task Force 59, which is experimenting with a whole bunch of different types of UAVs, especially unmanned surface vehicles or uncrewed surface vehicles in the Persian Gulf, and using that as sort of a laboratory of how you could scale that up, perhaps even to a Taiwan Strait scenario, and basically turn the Taiwan straight into this hellscape for an invading amphibious force, if that is the situation that they're worried about with China. With China having this advantage in mass and proximity, the idea is you could counter that by basically flooding the Taiwan Strait, so to speak, not really flooding the strait, but by putting thousands or at least hundreds of these unmanned surface vehicles in between Taiwan and China, and overwhelming the defenses of these ships and the overall Chinese military defensive complex to account for each one.

That's sort of the idea that's kind of driving this, because if you're going to do that, then you need to have an acquisition system, and you need to have an industrial base, and you need to have a logistics infrastructure, and a training infrastructure and all this stuff in place in order to use those and to field those and to operate these drones and command and control them in a way that we just haven't so far in terms of the volume and the disparate nature of them. We're talking about totally different kinds of drones, thousands of them, some very commercial, some very advanced and military specific. So that's the whole concept.

Now, Admiral Aquilino talked about this Hellscape concept is something that they've learned, especially from the experience of what they've seen in the war in Ukraine with Russia, especially what's happening with the Russian Navy fleet and the Black Sea, and how the Ukrainian special forces, or military in general, is using these sorts of asymmetric capabilities, unmanned surface vehicles or sort of advanced torpedoes in some ways to wreak havoc on the Russian Baltic fleet.

So I'd like to sort of open this up, especially with Sam Bendett on the line. He's sort of the expert in the drone warfare that we've seen in Russia and Ukraine. What's the baseline there? What are we seeing Ukraine do? And how effective has that been with their objectives, especially with the Russian Navy, or even over land?

Sam Bendett: Thanks for having me on. And I think it's a very good question. I think we have to put some things in perspective before we actually start discussing actual direct comparison between what's happening in Ukraine and other conflicts. First of all, there's a huge number of short-range, extremely cheap UAVs, which dominate the skies right now. And all of these FPV quadcopter drones, which have become such a headache for both sides, are made possible with Chinese components. And everyone who builds them can assemble these basically by the thousands, because they can still buy all the basic components from China by the barrel full, number one.

Number two, when it comes to Russian and Ukrainian use of long-range drones, which is becoming more sophisticated, more effective, and more directed at the military capability of each country, there's still a trade-off between distance and actual explosive charge. So the Ukrainian drones that attack Moscow, for example, flew the distance, but caused very minimal damage. When those drones shifted to Russian industrial facilities and military basis, there was more of an effect, but a much greater cumulative effect is needed to actually impact the war.

Russians have been pounding Ukraine with Shahed drones. Most of them are shut down, many targets are civilian stationary targets. In a military confrontation, large scale military confrontation, not all targets are going to be stationary, and certainly not all targets are going to be locked away in ports. Ukrainian unmanned surface vessels have been very successful, because Russia has exhibited limited maneuver capability for their Black Sea fleet. Their ships aren't there zipping around the Black Sea. A lot of them are really just kind of backed away in their port facilities, which includes logistic ships, port ships and others.

That doesn't mean that vessels at sea cannot be attacked, and we've seen some evidence of Ukrainians attempting to attack Russian Black Sea Fleet naval vessels, which are on the move. But what this has prompted is a response from the Russian Navy and response from the Russian military in trying to develop extensive countermeasures, both sensory and kinetic and defensive countermeasures for port facilities and ships themselves. What is actually working for Ukrainians is the fact that a lot of this cumulative experience isn't shared widely and readily across the Russian military. Some units, some formations, some parts of the Russian Navy are more adept at meeting this large threat than others. And that's why there's a difference between attacks on Sevastopol, which is now relatively well defended against the USV type boats, and Novorossiysk, where ships were basically open to this type of strike.

So there are certainly a lot of lessons that can be drawn from large scale uses of disparate types of UAVs and unmanned surface vessels. I think what both sides are calling for is for better communication between all of these assets, for better analysis of data that is gathered by all of these assets for a common operating picture. And both sides are calling for greater autonomy because of extensive countermeasures that impact communication between the operators, the users, and the actual systems themselves.

So in any Indo-Pacific region, conflict distances are going to be paramount, and therefore, different considerations have to come into play with designing and fielding expandable loading munition type, kamikaze type systems, because you'll have to make a trade-off, not just for distance versus the charge, the explosive charge, but you actually have to combine the ability to go the distance with a significant charge in order to attack targets that are going to be in play in that type of war. So let me stop there and turn it back to you.

Jen DiMascio: Well, I had a question there. I mean, where is the technology right now in terms of the distance that you're talking about? But also in terms of when you're talking about fielding or putting thousands of uncrewed vehicles into operation, the Pentagon has done a lot of work with swarming drones. Has that advanced at all, and where is the level of autonomy compared to the Replicator Initiative's 18 to 24 month timescale?

Steve Trimble: The initial focus of replicator, according to Kath Hicks, is what they call the, I need to make sure I've got my acronym correct, ADA2, which should not be mistaken for A2/AD, the anti-access area denial concept. In fact, ADA2 is their way of addressing A2/AD. And ADA2 is all-domain, attritable autonomy. That's another concept that Kath Hicks is promoting through the replicator program.

So there is a lot of different efforts underway to take, essentially, these drones, commercial drones, as well as military drones, which are right now basically being controlled on an individual basis. Like for every single robot, whether it's an airplane or a ground vehicle or a surface vehicle, there's at least one human, if not multiple humans, in the process of controlling how it's working, plus the payloads on board. So what they want to do is create an autonomy platform that can give these vehicles some control over themselves and some decision-making capability, some if then and even some more advanced machine learning and AI-based thought processes, let's say, so that you don't have to have this one-to-one ratio of human to robot control.

So in the air domain, we know there's Skyborg, that's the AFRL program, but there's venture capital-backed companies, like Shield AI has Hivemind, Anduril has Lattice. Lockheed's got their own system, autonomy system. Boeing's got theirs, Raytheon's got theirs. They've all talked about experimenting with their system. So there's a lot of different options out there.

And then you've got DARPA with REMA, and that's basically taking that instead of... It's sort of assuming that the platform that hosts it isn't going to have the advanced processors that the military might acquire for something like, well, like a MQ-28, the Boeing Australia and CCA collaborative combat aircraft type platform, which is pretty big and jet powered and all that kind of stuff.

But if you're talking about a quadcopter, it's not going to have that level of processing capability on board. So how do you figure out how to distribute that autonomy software capability, that advanced processing that's required on something like that that doesn't have, and it was never designed to host that kind of thing, right? So they're building an adapter. DARPA is building an adapter, plus a software platform. Interestingly, Lockheed's doing the same thing. They told us about that a few weeks ago at Skunk Works. They're building what they call an Adapter Kit, which actually comes from the MQ-25 program, bizarrely enough, from the Navy.

But anyway, so the 18, 24-month timescale is the really difficult challenge here, because there's a lot of things that have to happen in that 18, 24 months industrially with Congress, there's no funding in the 23 budget or the 24 budget for this. So anything that has to happen has to come from other programs, and then you need to reprogram that and get congressional authority. And there are some authorities that the OSD has that Kath Hicks's office has, like the DMAG, or what is it, Deputy Management Action Group, has some authorities to do that kind of thing, although, again, with congressional approval. So the timescales are really the driving factor here and trying to come up with a way to do this at such an aggressive pace seems to be the real challenge they're taking on. It's more of a process challenge than a technological challenge, I guess that's the way to put it.

I'm curious what Byron Callan thinks about all this and how the markets have been reacting to it. What's your perspective on it, Byron?

Byron Callan: Well, thanks, Steve, and thanks for having me on the podcast today. Look, I think a couple of thoughts here. First, we're already a month after Kath Hicks's announcement, so we're probably really talking about 17 to 23 months now. But I think it's interesting, and maybe I'll bounce this back to Sam in a second, but the problem set that the US is looking at is different than what's going on in Russia and Ukraine. I mean, if you really think about the Taiwan Strait is, what, about 200 kilometers from the shores of Taiwan into China.

I think what's being talked about here, these are not going to be weapons that would be able to sink or disable a large surface combatant that be part of a Chinese invasion force, a theoretical invasion force for Taiwan. You're not going to be sinking a large amphibious ship, or a destroyer, or a Corvette. I think the real deterrence value in having, let's just say, 50 to a hundred thousand of these on Taiwan, obviously, before the start of a conflict, would be their ability to really potentially reach into China, disrupt embarkation points, but probably equally importantly, strike at the smaller ships, part of the Chinese maritime militia force, that using an LRASM size weapon would just be overkill for these type of vessels that would be important in sustaining an initial Chinese assault on Taiwan.

So when I hear Hellscape and I think about, "Well, what's the problem set?" That's probably the problem set where these size drones would really fit into. And obviously, it would give more resilience. I think you could envision a weeklong period where there's not going to be much flying or able to move in and out of Taiwan, certainly from the United States and maybe even from Japan. So having that capability on the ground that would be very hard to detect is going to be an interesting problem for China to deal with.

Now, what I think the industrial issues are, we can scale. I mean, we've kind of showed that in the [Mine-Resistant Ambush Protected vehicle] program in a fairly short period of time. I think what's interesting about this particular notion of an autonomous system is, do we have the electronic components that are secure? We can't just buy Chinese drones and throw an applique without worrying about the cybersecurity on those drones. And I suppose the other interesting question is, "So how do you integrate these, if they're going to have sensors on them, into the rest of the DOD, or Taiwanese, or even Japanese self-defense force networks?" That's another problem. So I don't think this is just kind of a build a factory and crank a lot of drones out as much as it's, what are the microelectronics, the architectures that need to go into that. That might be the hardest problem for DOD to think through and how to scale.

And I'll turn it back to Sam, because I'm just curious, Sam, is the technology there to build these kinds of drones that would have a 100-, 200-kilometer range where you might build them for a thousand, maybe $5,000 a copy, and they could be lethal and effective?

Sam Bendett: I think it's a great question. I think most of the long-range aerial and maritime assets right now in Ukraine, I think they cost a lot of money. For Russia building Iranian Shahed version, I think each of those drones costs less than $50,000. But again, there's a trade-off between distance and the actual explosive charge and munitions. For Ukraine, they can build USVs that can go the distance, but they are expensive and they cost hundreds of thousands each. So Ukraine right now is working on the development of long-range kamikaze drones, but I think the price of each is still going to be in the tens of thousands of dollars. So I don't think there's technology there to build long-range kamikaze drones very cheaply. Short-range, absolutely, mid-range, possibly, long-range, not quite yet, at least we're not seeing that yet in Ukraine.

Steve Trimble: This is the big issue is, especially at the small drone level, because another part of this is that China moved in several years ago and dominated the industry for small commercial drones with the DJI Phantom series. There was a Nascent US quadcopter industry that was trying to build itself out, but it was almost immediately crushed by these cheap DJI phantoms that actually had a lot of really good capability and were substantially subsidized by the Chinese government coming into the market and putting them out. So in a sense, we have to kind of rebuild that on US soil. It's one thing to rebuild just the little rotors and the hardware of these quad rotors, but the electronics is going to be a real challenge in the chips. And I see this everywhere now.

So we reported over a year ago that Israel has started the Standoff In program, SOI, that envisions an entire robotic brigade that would enter a violent area, sort of a battle zone, initially, and keep their human controllers away from direct fire. And after the robots, UJVs with UAVs sort of neutralize the threat zone, then the human troops would come in, with the idea being the IDF wants to make sure that none of its soldiers get killed.

And then Taiwan, I was just there actually just a week ago, and they came out with their new National Defense White Paper while I was there, which was a happy coincidence, really. And they called for buying 7,000 commercial drones. But again, they're rebuilding their commercial drone industry. And this is kind of, when we look at the future of warfare, this is going to be a new part of it, but you have to sort of work through these industrial issues, the command and control issue, the logistics issue, the cost of these things, and how do you get everything in the sweet spot to deal with this new capability that has kind of proven itself as at least a factor in the Ukraine and Russia conflict. So that's just my perspective on it.

Jen DiMascio: Well, unfortunately, that's all we have time for today. Sam, Byron, Steve, thank you so much for joining the discussion. And everybody out there listening, join us again next week for another edition of Check 6. Thank you. Bye-bye.