Anything's a Swarm with Sufficient Quantity?

Good evening and welcome to those of you that joined us these past few weeks! The Arsenal Team went away for an extended vacation, but we’re back and cannot wait to provide our community with weekly updates on the defense tech industry! We want to also take a moment and acknowledge the sacrifices and heroism that came 22 years ago on Sept 11th, 2001. We want to acknowledge the first responders who sacrificed all, as well as the men and women who fought the global war on terror for the next two decades in the aftermath of that horrific event. We will never forget.

This week’s post:

• 🤖 Anything’s a Swarm with Sufficient Quantity?

• 💰 Term Sheet

• 🚩 Red Team Update

If this email was forwarded to you, sign up here

Anything’s a Swarm with Sufficient Quantity?

The Deputy Secretary’s remarks at the NDIA Emerging Technologies for Defense Conference two weeks ago outlined some rough edges for DoD’s “Replicator” project. So far, it’s more of a concept than anything concrete, but the vision she laid out integrates some of the trends DoD has been slowly building towards: quantity achieved by sacrificing quality (“exquisiteness”), as well as autonomy and AI as key enablers for coordinating quantity. The bid is to counter PRC mass with mass - quantity has a quality all its own: “We’ll counter the [People’s Liberation Army]’s mass with mass of our own, but ours will be harder to plan for, harder to hit, and harder to beat.” Which begs the question: supposing the DoD is able to scale production of unmanned systems (UxS) as envisioned by Replicator, how will these systems work together? Enter swarming.

There’s no consistent definition of swarming in the DoD, which is not always a bad thing for an emerging technology, but it also betrays how far we are from operationalizing it. In 2019, the Defense Systems Information Analysis Center (DSIAC) attempted to unify various published definitions with the result that the only common thread was “multiple UAS [sic] that use individual behaviors to work as a unit.” Most commonly, swarming means something more than predefined actions for many autonomous units.

It does entail many units, but those units share information locally to give rise to emergent (the sum is more than its individual parts) collective behavior based on individual actions that are reactive and often surprisingly simple. This is the sense in which it is found in the natural world - think insect swarms, flocks of birds, schools of fish, etc. - from which the military usage draws inspiration. A nice working definition is from a recent Joint Counter-small Unmanned Aircraft Systems Office (JCO) Request for White Papers: “a sUAS swarm system is defined as a group of unmanned aircraft that demonstrates coordinated behavior to achieve a common objective.”

Lately, swarming has been envisioned as a way to pierce anti-access/area denial (A2/AD) bubbles. It’s often associated with aerial systems given the widespread proliferation of those systems, and features prominently in concepts to overwhelm defensive systems with saturation, or as sacrificial targets enabling more capable munitions to hit their targets. But don’t take our word for it. From DARPA’s Autonomous Multi-Domain Adaptive Swarms-of-Swarms (AMASS) Broad Agency Announcement (BAA) to dynamically command and control unmanned, autonomous swarms of various types: “Under the umbrella of counter-A2/AD are the mission sets of…

• counter-Integrated Air Defense Systems (IADS);

• counter-Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance (C4ISR)

• counter-radar

• counter-maneuver

• Beyond Line-of-Sight (BLOS) / Over-the-Horizon (OTH) targeting

• ISR

All of the Services are swarming to get in on the action (sorry, couldn’t resist). Under the Navy’s Super Swarm mission, there’s at least a few different lines of effort:

• LOCUST (low-cost UAV swarming technology) - up to 30 drones flew together in 2017

• DEALRS (deployment and employment of autonomous long-range systems) - mothership drones for lots of smaller drones

• MASS (manufacturing of autonomous systems at scale) - 3D printing for lower-cost drones (see our interview with a similar company, Firestorm Labs)

• MATes (manned and autonomous teams) - human/swarm interactions and C2

The Air Force’s Science and Technology Strategy from April 2019 mentions swarming to “attack adversary weaknesses in unexpected ways by exploiting numbers and complexity” to “provide adaptability, rapid upgradability, and the capacity to absorb losses that manned systems cannot.” For the Marine Corps, updates to its Force Design 2030 guidance mention prioritizing the “Long Range Unmanned Surface Vessel with associated swarming drones,” “swarming unmanned aircraft system (UAS) capability” for infantry battalions, and “a low-cost, air launched family of loitering, swarming munitions.” And the Army’s Rapid Capabilities and Critical Technologies Office (RCCTO) awarded BlueHalo a $14M contract to develop “a transformational sUAS autonomous swarming capability not currently available to the warfighter.”

There are a handful of swarm-specific start-ups (say that three times fast) out there, but most firms seem to be established UxS entities using swarming as an enabler, or systems integrators and large primes (Northrup and Raytheon were performers on DARPA’s past Offensive Swarm-Enabled Tactics - OFFSET - program). A past Strategic Capabilities Office program called Perdix used MIT’s Lincoln Laboratory and was featured in 60 Minutes. And lots of action in counter-swarm, probably more so right now than in offensive swarming.

So there could be a play on the hardware side with the systems themselves, or the software side to operationalize the systems to actually swarm. That’s arguably the harder of the two right now, especially given the need to communicate information amongst all the systems at or close to real-time.

But operationally, even pre-planned actions or rudimentary rules-based autonomy significantly complicates a defensive response. Maybe starting with simply building, procuring, and delivering UxS systems in sufficient quantity to matter is a great place to start. And leads us back to the Replicator idea.

Red Team Update

China Defeats UAS Jamming with SWARM and Autonomous Tech

• In a recent trial by China’s National University of Defense Technology, a swarm of dozens of drones overcame test jamming signals by aiding each other, according to state media. Then, without the help from a human operator, found and destroyed a target as a loitering munition

• A recent study of hundreds of AI-related military procurement records conducted by the Center for Security and Emerging Technology at Georgetown University in Washington, DC found that a third of all known contracts in the U.S and China from an eight month period in 2020 were for intelligent and autonomous vehicles, the largest share in both countries

• Both the US and China are spending billions on military AI research and development which only aides autonomous vehicles and swarming technologies. In 2021, Chinese military publications showed officials discussing an approach to warfare that incorporates advances in big data and AI to help identify US vulnerabilities and utilize precision strikes against those vulnerabilities

About Us

Our team has 30+ years of combined experience as military officers using the end products. We’ve worked in both government and industry. From MIT to Wharton, Wall Street to biotech, and DARPA to the flightline, we offer you a unique perspective on how to navigate America’s defense tech industry.

The opinions expressed in this newsletter are solely those of the authors and do not necessarily reflect the views of DoD, our employers or any affiliated organization. This newsletter is for informational purposes only and is not intended to provide legal, financial or professional advice.