Replicator Initiative Looks to Swarm Through ‘Valley of Death’
Army officials operate a swarm of drones during an exercise at the National Training Center in Fort Irwin, California.
Defense Dept. photo
Deputy Secretary of Defense Kathleen Hicks laid out an ambitious goal when she announced the Replicator initiative in August: within two years, the Defense Department will field thousands of all-domain, attritable, autonomous swarming robots to counter China’s military mass.
While that objective in and of itself is quite eye-catching, Replicator’s other goal — to revolutionize how the Defense Department does business by building a bridge across the so-called “Valley of Death,” where innovative defense technologies fail to make it to the field — could be the real success story of the initiative.
There are several programs across the services focused on autonomous systems: the Army recently announced plans to deploy human-machine integrated formations, the Navy’s first and only uncrewed, high-altitude drone reached initial operational capability in 2023 and the Air Force is developing unmanned, collaborative combat aircraft to fly alongside its Next Generation Air Dominance fighter jets.
While these examples may not reflect exactly what Replicator is after — Hicks has described the initiative’s shopping list as “small, smart, cheap and many” systems that could potentially “move the needle” in the Indo-Pacific — the initiative is “not a new program of record” and will leverage “existing funding, existing programming lines and existing authorities to accelerate production and delivery at scale,” she said at the Defense News Conference in September.
Congress made clear during the 2023 budget cycle that it wants the Defense Department to “use its full authorities” for acquisition, Hicks said during a Defense Writers Group event in November. “‘We don’t think you’re pushing hard enough, do what you can with what you have’ is the gist of what we read.
“And I’m a big believer [that] we need to be problem-solvers in DoD,” she continued. “The answer to every question is not, ‘I need more money and more people.’ So, Replicator is a great example of how we are taking that feedback from Congress, looking to problem solve for ourselves and working with what we have at hand.”
The Defense Department has a number of “alternative acquisition strategies,” such as middle-tier acquisition to field systems within two to five years and the software acquisition pathway, to facilitate rapid, iterative delivery of software capabilities, but “we are very acclimated to — for perfectly logical reasons — a particular acquisition strategy: a program of record that we can track through the budget that has specific line items, it moves through different colors of money, etc.,” Hicks said. Changing the acquisition system will require “a more systematic, senior-level focus … on how do we remove barriers beyond just the resourcing piece … to make sure things make it across the Valley of Death.”
Jaret Riddick, a senior fellow at Georgetown University’s Center for Security and Emerging Technology, said the fact it was a senior leader such as Hicks who announced the Replicator initiative sent a “clear demand signal” to industry.
The initiative is a “clear declaration” for the United States to “create an offset against one of China’s greatest strategic advantages, and that’s mass — having more ships, more planes, more missiles, more people,” Riddick said during a recent discussion hosted by the Center for Security and Emerging Technology.
Additionally, the announcement coming from the deputy secretary of defense gives Replicator “sort of the gravitas, the serious focus on this, in a condensed time period to really learn how we can accelerate and how we can get these sort of processes in place that help us to be able to do things quickly in the environment that we’re in.”
In response to Replicator, “we’ve seen mobilization” both inside and outside of the government, he added. To lead the initiative, the Defense Department has established the Deputy’s Innovation Steering Group — co-chaired by Hicks and the vice chairman of the Joint Chiefs of Staff — along with the supporting Defense Innovation Working Group, led by the Defense Innovation Unit.
Igor Mikolic-Torreira, the center’s director of analysis, said the Deputy’s Innovation Steering Group is the “critical piece” to make Replicator’s ambitious goal a reality.
“Having a group like this with all the relevant people in the room … really focuses the department and the decision-making processes on this issue and what needs to happen,” Mikolic-Torreira said. “The key players that have influence on the process who have authority, who have powers to make this happen, are all in the room with the vice chairman and the deputy secretary, [who are] in some polite fashion breathing fire on them, and the system responds very effectively to this.”
That doesn’t mean the group won’t run into challenges, such as how to acquire systems on as great a scale as Hicks has stated, he said. That is where the Defense Innovation Unit will play a key role, Riddick said.
DIU is “oriented to really engage” across the “network of industry partners,” particularly nontraditional defense companies — “the folks who have been sort of maturing these technologies in the commercial space,” Riddick said.
DIU Director Doug Beck said expanding the aperture “of players that are providing capability into this space” — and thus expanding the industrial base “on which you can rely … makes an enormous difference.
“A lot of those players — including some of those more emerging commercial tech players who are playing in this space — are working very, very hard on solving” potential issues for Replicator, such as scalability and supply chain challenges, he said during a media roundtable in November.
Aditi Kumar, DIU’s deputy director of strategy, policy and national security partnerships, said for Replicator the organization is not only “looking at the programs already in the department’s pipeline that fit the mission need,” but also identifying capability gaps, such as domain area or enabling infrastructure, “and this is where we’re looking at leveraging our solicitations process to bring in new entrants,” exploring if industry can fill those gaps.
DIU is planning to host a Replicator technology summit in early 2024, “where we’re bringing in our partners here to think through what capabilities are critical to this, and how we build some of the common architecture pieces and other enabling pieces to deliver a capability that, over the long term, the department can benefit from,” Kumar added.
DIU is not only oriented to form unique connections with the private sector but can also build relationships with the “operational community — so, the combatant commands to the actual warfighters,” Riddick said. “Those types of connections will certainly be necessary to understand” the needs of the warfighter, and what systems will fit those needs.
Defense Department officials have shown reluctance to publicly discuss details about what systems will be bought as part of Replicator. In November, Hicks said the department will select the initial batch of candidates “in the next about three weeks,” but as of press time, no announcement had been made.
“We’re being very careful … about the way in which we talk about Replicator” so as not to reveal too much information to China and other strategic competitors, Hicks said — similar to how the Pentagon has remained cagey with details about the B-21 Raider strategic bomber and the Next Generation Air Dominance fighter programs.
The department will be “clear and transparent” with Congress about Replicator, “but how we choose to speak about it” publicly, “in terms of the particular programs or projects that we’ll be accelerating through Replicator, is to be determined.”
What is known is that the systems will be bought in “tranches,” or as Kumar put it, “successive batches.”
“The reason we’re approaching it in successive batches is because there are some … systems that are ready to go from a scaling perspective,” that have been “tested and validated” and “we have high confidence that they will meet the mission need,” Kumar said. “There might be other systems where some, or all, pieces of that work still need to be done,” such as analysis, testing or evaluating scalability. “As we answer those questions, those systems will start getting wrapped into the successive batches.” New entrants can deliver systems in later tranches that fill gaps in the current portfolio.
Buying systems in tranches also gives the Defense Department flexibility if a robot isn’t performing correctly, “buying down risk” that would otherwise be passed to the warfighter, she added.
“How do we buy down risk so that we can move forward [and] give ourselves the optionality to iterate on those decisions?” she said. “Because at the end of the day, taking that risk now is taking that risk away from the warfighter later. … We want to proceed in steps so that we are learning at each point and then making the next decision based on those learnings, as opposed to making big decisions up front and then locking ourselves into something.”
Michael O’Connor, a major in the Space Force and a Department of the Air Force fellow at the Center for Security and Emerging Technology, said buying systems in tranches solves a common problem, “where you have huge waves of demand … out of the military for capabilities, for either commodities or new systems, every time the world gets extra interesting. And then you have these peace dividends where industry has been scaled back because the DoD is scaling back.
“If you can have this constant demand signal, it’s much easier to plan for, it’s much more efficient, and I think it also helps keep a more healthy industrial base. And ultimately, that will also be useful” as a process that can be repeated going forward, O’Connor said.
The goal is to replicate this process for other defense technologies, Beck said. “It’s about delivering concrete results in 18 to 24 months while driving the process to accelerate technology at scale in a way that is replicable. We’re focusing first on … unmanned, autonomous, attritable systems, with multiple thousands across multiple domains, and there will be other Replicators” in the future.
The end product of Replicator will be “simultaneously the impact of fielded systems that [help] to deter major conflict — or win if forced to fight — in the Indo-Pacific, and the learning that we have from having done that, that allows us to do it again and again and again,” he said. “And in order to do that, we’re going to have to hit snags, learn from them and break down systemic barriers along the way. If, in fact, those barriers were already all broken down and it was simple, we wouldn’t have to do this. So, we’re doing that on purpose, and that’s part of what this is all about.”
Lauren Kahn, a senior research analyst at the Center for Security and Emerging Technology, said while fielding thousands of autonomous drones within the next two years would technically be a success, what would actually make Replicator a notable achievement is if the Pentagon can indeed “replicate this system again elsewhere” and prove it can “regularly overcome this kind of Valley of Death … to make sure that the DoD can actually access all the resources that are available to it.”
Mikolic-Torreira concurred. “Are we going to be successful if after two years we have a few thousand drones of various kinds in the force? Yes, in some sense, but I think that’s the disappointing success,” he said. “The really useful success is if we get this institutionalized where you’re continuously buying fairly large tranches of drones” and sustaining this “clear demand signal for products.
“Replicator is a real success if it spawns its continuation,” he said. If consistently buying and upgrading systems “gets instantiated as part of that process … now you’ve got the demand signal that the industrial base will respond to. … So to me, that’s the biggest long-term measure of success is if it changes the department in that way.”
While driving change in the Defense Department’s acquisition processes is important, for Hicks, the warfighter is Replicator’s top priority.
“What to look for on Replicator is, does the United States Department of Defense field within 18 to 24 months thousands of attritable, autonomous systems? If the answer to that question is ‘yes,’ and we’re able to talk about the way we burned down risk to make that happen and deliver capability, that builds trust on the commercial sector side that we can move within that two-year cycle and increase investment opportunities for companies that are trying to be in that space,” she said.
“And most importantly for us, it delivers capability to the warfighter faster. That’s what Replicator is about.” ND
Topics: Defense Department