The Future of Warfare: Small, Many, Smart vs. Few & Exquisite?


In the 1970s, faced with the USSR’s overwhelming superiority in numbers, the Department of Defense decided to compensate by focusing on high technology platforms.  This led to the highly successful F-15, F-16, F-18, Abrams tanks, and Bradley fighting vehicles.   Since then, the United States has continued to pursue cutting edge technology that has resulted in the highly capable F-22 and, when the testing and software development is complete, perhaps a highly capable F-35.  Unfortunately, cost has accelerated faster than capabilities.  And thus numbers have declined precipitously.  The U.S. Air Force initially planned to buy 750 F-22s, but the high cost led Secretary of Defense Robert Gates to cap the program at 187.  Nor has the Air Force been alone in pursuing top end systems.  The Navy attempted an entirely new concept with “Streetfighter.”  Meant to be a low-cost, highly capable ship to replace the Navy’s frigates and minesweepers for operations in brown water, it evolved into the Littoral Combat Ship. High cost and poor performance led the Navy to cut their planned purchase from 55 to 32 with potentially more cuts in the future.  Similarly, the Zumwalt class destroyer program was initially planned for 32 ships, but rising costs mean only three will be built. The result has been a pattern of fielding exquisite platforms in diminishing numbers at great cost.

While it was the right decision to pursue high end systems in the 1970s, dramatic improvements in the fields of robotics, artificial intelligence, additive manufacturing, biology, and nano-materials are changing the cost/effectiveness calculation in favor of the “many and simple” against the “few and complex.”   The convergence of these technologies and the steady decrease in costs even as capabilities increase is rapidly expanding the destructive power, range, and precision of weapons that soon will be both widely available and relatively cheap.  As Frank Hoffman has noted, we are putting ourselves on the wrong side of a self-imposed cost curve.

To illustrate how small, many, and smart are emerging as major shifts in warfare, this article will start by examining why it is now possible to create small, smart, and cheap platforms that have sufficient range and combat capability to fulfill the very challenging role of power projection. It will then examine the implications for U.S. defense.


The last decade has made the global public familiar with expensive high end drones.  Yet, perhaps the most interesting developments have taken place at the low cost end of the spectrum.  In 1998, an industry/university consortium flew a composite drone from Newfoundland to Scotland on two gallons of fuel.By 2003, a hobbyist launched a GPS-guided model airplane/drone that flew autonomously from Newfoundland to precisely the right landing point in Ireland.  Built of balsa and plywood with a tiny gasoline engine that burned less than one gallon of fuel in the 26 hour flight, it was cheap enough that the hobbyist built 23 to ensure he could be the first hobbyist to fly across the Atlantic.  He made it with the third launch.  In the intervening 12 years, governments, hobbyists, and businesses have steadily increased the range and capability of these platforms.  Hobbyists and businesses have made use of the rapid technological convergence to decrease the cost of long-range, autonomous systems at least an order of magnitude.  Today they are routinely flying smart systems with intercontinental range — they lack only a payload to be a precision weapons system. Their composite construction and very low energy usage mean they will be very difficult to detect.

Of even greater concern, these small, inexpensive drones are designed specifically to be used by people with no particular skills and no in-house maintenance system. Most still require a remote human operator.   But flying them has become so easy that realtors and wedding photographers are using quad-copters with stabilized camera mounts to film properties and events.  Industry has already taken the next step and provided farmers with inexpensive autonomous drones to monitor their crops. As one article explained:

What “drones” means to Kunde and the growing number of farmers like him is simply a low-cost aerial camera platform: either miniature fixed-wing airplanes or, more commonly, quadcopters and other multibladed small helicopters. These aircraft are equipped with an autopilot using GPS and a standard point-and-shoot camera controlled by the autopilot; … Whereas a traditional radio-­controlled aircraft needs to be flown by a pilot on the ground, in Kunde’s drone the autopilot (made by my company, 3D Robotics) does all the flying, from auto takeoff to landing. Its software plans the flight path, aiming for maximum coverage of the vineyards, and controls the camera to optimize the images for later analysis. …  At the heart of a drone, the autopilot runs specialized software—often open-source programs created by communities such as DIY Drones, which I founded, rather than costly code from the aerospace industry.

Since air is the simplest environment, it is not surprising that fully autonomous, cheap, and long-range drones emerged there first. They will be followed quickly by maritime and ground systems. In 2010, Rutgers University launched an underwater “glider” drone that crossed the Atlantic Ocean unrefueled.  This year, the U.S. Navy has launched an underwater glider that harvests energy from the ocean thermocline and plans for it to operate for five years without refueling.

In short, small air and sea platforms have demonstrated the capability of achieving intercontinental range while producing very little in the way of radar or heat signatures.


While most of the public focus has been on high-end drones conducting anti-terror strikes, the Chinese have fielded the Harpy Unmanned Combat Air Vehicle (UCAV).  Initially developed in the 1990s by Israel as an anti-radar system, the Chinese version has a range of 500 km and a 32kg warhead with multiple types of seeker heads.  One Chinese configuration has 18 Harpies in box launchers mounted on a single truck bed (other configurations use 6 launchers per truck). Essentially, these are expendable drones capable of saturating defensive systems.  Both China and Israel have displayed these weapons at trade shows in an effort to sell them to other nations.  The system is currently operational with the Turkish, Korean, Chinese, and Indian Armies.  Today, the Israeli version has an electro-optical sensor to attack non-emitting targets and an extended range of 1000 km. One can assume China has made similar improvements to its systems. This system represents a first step toward inexpensive swarms.   Yet Harpy class UAVs, while expendable, are still relatively expensive and thus major deployments of weapons of this class require a well-funded state military.

The primary driver of how many systems are purchased is cost.   But additive manufacturing is driving down the cost of many manufactured products.   Today researchers in England have prototyped a printed drone that will cost roughly $9 a copy.  And additive manufacturing is not only low end products.

Mark Valerio, vice president and general manager of military space for Lockheed, told Reuters, “In the next decade, we will completely change the way a satellite is designed and built. We will print a satellite,”

Valerio suggests such a satellite will cost 40% less than current models.  These trends indicate that dramatic cost decreases will be the norm for these widely used and increasingly capable commercial drones.

We don’t have to wait for additive manufacturing, either.  The U.S. Navy has announced it will repurpose the commercially produced Slocum Glider – a five foot long, autonomous underwater research vehicle.   The glider can patrol for weeks following initial instructions, surfacing periodically to report and receive new instructions.  Such drones are being used globally and cost about $100,000.  Clearly such drones can be modified into long-range autonomous torpedoes or mine delivery vehicles.  For the cost of one Virginia class submarine, a nation could purchase 17,500 such drones. Additive manufacturing can and likely will reduce the cost of these systems even more. And the skills needed to build and employ a glider are orders of magnitude less than those needed for a nuclear sub.

“Smart” sea mines should be a particular concern for the United States.  Simple contact and influence mines have the distinction of being the only weapon that has defeated a U.S. amphibious assault – the landing at Wonson, Korea in 1950.  While lanes were eventually cleared through the primitive minefields, forces attacking up the west coast of Korea had already seized the amphibious objectives before the first amphibious forces got ashore.  Frustrated at cruising up and down the coast as the Navy tried to clear the mines, the Marines nicknamed the landing “Operation Yo-Yo.”  Not much has changed. In February 1991, the U.S. Navy lost command of the northern Arabian Gulf to more than 1,300 mines that had been sown by Iraqi forces …”  These were simple moored sea mines.

Since 1950, mines have become progressively smarter, more discriminating, and more difficult to find. They have sensors which can use acoustic, magnetic, and other signals to identify and attack a specific kind of ship, allowing – for example – commercial vehicles to pass unmolested.   As early as 1979, the United States fielded CAPTOR mines.  These are encapsulated torpedoes that are anchored to the ocean floor.  When they detect the designated target, they launch the captured torpedo to destroy it out to a range of 8 KM. Today China possesses “self-navigating mines” and even rocket propelled mines. We are seeing early efforts to use unmanned underwater vehicles to deliver mines.  Since commercially available drones are already crossing the ocean autonomously, pairing drones with mines will almost certainly make it possible to mine sea ports of debarkation and perhaps even sea ports of embarkation.

Ashore, mobile land mines/autonomous anti-vehicle weapons are also under development. The natural marriage of IEDs to inexpensive, autonomous drones is virtually inevitable.  The obvious targets are parked aircraft, fuel dumps, ammo dumps, communication sites, and command centers.  Non-state and state actors alike will rapidly transition to drones that can hunt even mobile targets.

Today’s inexpensive drone systems mean states and even non-state actors can afford large numbers of lethal air, sea, and ground drones.  Within the decade, U.S. forces should expect to be attacked by these weapons on every combat deployment.


We can also expect the inexpensive autonomy seen in today’s agricultural drones.  The autonomy has been made possible by impressive technological advances combining tiny sensors, GPS modules, microprocessors, and digital radios, all of which are dropping in price and are commercially available.

These same technologies can be applied cheaply to military systems.  While the Pentagon faces the “Innovator’s Dilemma” and will be severely challenged to keep costs low, other nations, start-up companies,  and non-state actors will not face the same bureaucratic hurdles and thus are likely to produce cheap, smart, and deadly drones using commercially available parts.  They won’t be highly reliable or reusable. They won’t need to be.  If only half of a swarm works correctly, it may be more than sufficient to overwhelm advanced defenses.  John Arquilla and David Ronfeldt’s concept of swarming drones is on the verge of realization.

Don’t Look Back, They Are Not Behind Us

Unfortunately for the West, autonomous drones will initially favor the less technologically advanced actor because their targeting problem is simpler. For instance, a non-state actor may not own armored vehicles or aircraft, so its autonomous drones only have to find and attack any armored vehicle or parked aircraft.  They do not have to discriminate but simply fly a pre-programmed route to a suspected target area and identify the target.   Target areas for many locations in the world–to include most airfield flight lines–can be determined using Google Maps.  Cheap optical recognition hardware and software that provides rough target discrimination is also becoming widely available. If the software of a farmer’s autonomous drone can point and shoot a camera, it can point and shoot an explosive device.

Clearly, these commercial products have demonstrated the ability of autonomous drones to reach a target area, but what weapon could it use? Commercially available quadcopters carry the 3 ounce GoPro camera and are achieving flight times of over 30 minutes.  Against the thin skin of an aircraft, a simple point detonating 3 ounce warhead is sufficient. Against armor, the drone designer may choose the heavier and more complex explosively formed penetrator. This will obviously require larger drones but will also provide standoff distance.  In 2009, the U.S. Army told CNN that such weapons can penetrate armor from 100 meters.  This potential marriage of proven, cheap technology represents a direct threat to a wide range of potential targets.

The addition of cheap, persistent air and space based surveillance will provide the information necessary to use these cheap drones.  Sky Box Imaging, which was recently purchased by Google, is deploying CubeSats.  Their goal is to sell half-meter resolution imagery with a revisit rate of several times a day – to include interpretation of what the buyer is seeing.  A buyer could track port, airfield, road, and rail system activity in near real time.

While the cheapest of these systems can carry only small payloads, the rapidly developing field of nano-energetics or nano-explosives will dramatically increase their destructive power.  As early as 2002, nano-explosives demonstrated an explosive power twice that of conventional explosives.  Since research in this field is classified, proprietary, or both, it is difficult to say what, if any progress has been made since that point.  But even if double the power is as good as it gets, a 100% increase in destructive power for the same size weapon is a massive increase.

Western forces should not assume they will have the technological edge when deploying to a conflict zone.  The higher standards for target discrimination will inhibit their fielding of autonomous lethal but cheap drones.  In this field, they should expect to the non-state or less ethical state to be the first to field such systems. Implications

The convergence of technologies and techniques is already producing small, smart, cheap, and long-range drones capable of carrying significant payloads.  Fuel gels and nano-explosives will increase the range and lethality of these commercially available systems.  Additive manufacturing will dramatically reduce the costs.  The Pentagon needs to rethink the exquisitely capable but extremely expensive weapons procurement programs it is pursuing.  While these systems were a major factor in the tactical successes of the last 24 years, the United States needs to think hard about the shift from exquisite and very few to cheap and very many.

For instance, rather than insisting on building the next generation bomber, we need to examine how best to execute the mission of effective long range strike. Even if one accepts the mantra that “we must be able to hold at risk what they value,” this does not mean the United States needs a new bomber.  We need to consider other approaches.   For instance, what other strike platforms could we purchase for the same investment as the proposed Long Range Strike Bomber?

In 2012, Air Force Chief of Staff General Norman Schwartz projected a cost of $550 million per Long Range Strike Bomber.  Tom Christie, the Pentagon’s Chief Weapons tester from 2001 until his retirement in 2005, is skeptical.  He thinks $2 billion per aircraft is a more accurate estimate.  Christie’s estimate aligns closely with the experience of building the B-2 bomber.  A 1997 report from the Government Accountability Office showed that while initial estimates for the B-2 were $456 million in 1997 dollars, the actual cost was $2.1 billion per aircraft.  While not usually included in the estimates the extremely high operating costs of stealth aircraft must also be included.  According to U.S. Air Force data, the B-2 costs $164,000 per flight hour to operate.  We should plan for similar or higher operating costs for the Long Range Strike Bomber.

The 2012 Pentagon budget shows Tactical Tomahawks costs $1.1 million per Tactical Tomahawk for a buy of only 196 missiles. This missile boasts a 1,000 pound warhead and “features a two-way satellite data link that allows the controller to switch target during flight to pre-programmed alternate targets or redirect it to a new target.”

According to the Naval Air Systems Command, the older Tomahawk Land Attack Missile cost $607,000 in FY-1999 dollars.  Using the U.S. Navy Deflator figure for Procurement, the Tomahawk currently costs $785,000.  Due to advances in additive manufacturing production techniques, the cost should drop to $470,000.  This year, the University of Southern California’s School of Engineering revealed a method for 3D-printing multi-material objects in minutes instead of hours.

Even if you think we will get the next generation bomber for less than the current B-2, a single Long Range Strike Bomber would pay for 4000 Tomahawks.  Given the historical record of bomber costs, it is more reasonable to assume that we will pay at least half again as much per aircraft for the new generation. Thus we could buy 6000 Tomahawks for the price of a single next generation Long Range Strike Bomber.  And of course, every month the bomber fleet will consume tens of millions in operating costs – which otherwise could purchase even more Tomahawks. For the price of a single bomber, we could provide a full load out for all Tomahawk capable ships in the fleet.

Advances in additive manufacturing, composite materials, energy densities in gel fuels, and nano-explosives indicate we will be able to build longer range, more powerful, and stealthier cruise missiles for much less money than the loitering Tomahawk.

Historical precedents

Remembering the history of another era of rapid, broad technological change might help DoD decision makers put the problem in perspective. In the early 1900s, navies were making very rapid technological gains in metallurgy, ammunition, explosives, engines, and communications. The 1906 launch of Dreadnought ushered in thebattleship era, and within a decade, capital ships were powered by turbines, had main batteries of 14-inch guns, rudimentary wireless, and vastly improved armor. By the beginning of WWI, battleships were considered the decisive weapon for fleet engagements, and the size of the battleship fleet was seen as a reasonable proxy for a navy’s strength. The war’s single major fleet action, the Battle of Jutland, seemed to prove these ideas correct.

Accordingly, during the interwar period, battleships received the lion’s share of naval investments. Displacement more than doubled, from the 27,000 tons of the pre-WWI New York class to the 48,500 tons of the Iowa class. The main batteries shifted from 14-inch to 16-inch guns, secondary batteries were improved, radar was installed, speed increased from 21 to 33 knots, cruising range more than doubled, and armor improved. Yet none of these advances changed the fundamental capabilities of the battleship. This is typical of mature technology, for which it costs much more to improve performance than it does for immature technology.

In contrast, naval aviation was in its infancy in 1914. Aircraft were slow, short-legged, lightly armed, and primarily used for reconnaissance. Air combat was primitive; attempts to bring down opposing aircraft included pistols, rifles, and even a grappling hook. After the war, aviation remained an auxiliary and was funded accordingly. While the Navy built 18 new battleships during the interwar period, it built only eight carriers, a total that includes the conversion of a collier (USS Langley) and two cruisers. And yet by 1941, carrier aviation had developed to the point that it dominated the naval battles of WWII.

The Navy’s failure to understand where genuine technological advantage lay carried real opportunity costs. In the late 1930s, the Navy spent heavily on fast battleships, eventually commissioning 10 and starting construction on two more. These ships turned out to serve primarily as very expensive anti-aircraft escorts for the fleet carriers. (Most naval shore bombardment was done by older battleships that dated to the 1920s.) One has to wonder: if the Navy had dedicated even half of the battleship funds spent in the 1930s to naval aviation, how different would the opening year of the war been?


Investment in highly capable and expensive new weapons systems is predicated on specific assumptions about the future. Unfortunately, it is a truism that one can never predict the future with certainty.  Thus a hedging approach is more functional than a predictive approach.  With the widespread commercial shift to small, many, and smart systems as a substitute for a few, exquisite systems, it is time for the United States to rethink its equipment procurement approach.

The critical military functions will remain – but how we accomplish them will change.  Rather than investing everything in a single type of fighter or a long range bomber, it makes more sense to limit our buys of these systems and augment them with systems that conform to small, smart, and many.  For missions like reconnaissance, strike, jamming, communications relay, and others, the United States needs to explore relatively cheap and even disposable systems.

Obviously, this will not be a rapid shift.  For instance, the United States is already heavily committed to the F-35.  But rather than buying over 2400 F-35s and continuing to build Ford carriers, we should examine limiting the buy of F-35 to six or seven hundred.  These aircraft, along with the current inventory of approximately 180 F-22s, this will provide sufficient numbers of aircraft for high end penetrating missions.  For other missions, existing and upgraded F-15, F-16, and F-18s can carry the load – particularly when augmented with large numbers of inexpensive penetrating platforms.  Rather than expensive manned Wild Weasel SAM suppression platforms, we could employ cheaper but more capable versions of the Harpy 2.  Cheap platforms even reduce the need for air-to-air capability.  Rather than destroying the aircraft in the air, swarms of cheap penetrators would strike at an opponent’s base.

In a similar ways, our current inventory of carriers will ensure that we have these ships available until the 2040s or later.  We can continue to build Fords to ensure that carriers operate until the Navy’s goal of 2100 – or we can seriously investigate how the convergence of robotics, artificial intelligence, additive manufacturing, biology, and nano-materials is going to change the character of future conflict.  Rather than decades-long, monolithic procurement programs, we can return to the process we used early in the days of aviation.  This was a time of wide commercial innovation in a variety of fields – internal combustion engines, metallurgy, design, radios, ordnance.  It was impossible to predict which designs would work best.  The industry used a model of build, test, improve, test, improve; only afterward did the Navy or War Department actually field the systems. The cost was low enough that they could abandon an aircraft if it was not working.  Despite investing a fraction of the money it spent on battleships during the interwar period, the Navy developed the carrier aviation team that dominated WWII naval warfare.

It is critical that we examine the few exquisite systems we are planning to buy – aircraft, ships, armor — and see if their missions could be accomplished by many, smart, cheap platforms.  Given the inherent political advantages of large, complex systems, this will be a difficult step.  The F-35 is a poster child for the difficultly of reconsidering a program of record.  Built in 45 states at a cost of $399 B for 2,443 aircraft and with expected lifetime operating cost of $1 trillion, the F-35 has powerful Congressional support.  Further, U.S. doctrine and powerful service constituencies heavily favor these exquisite systems.  This is natural since doctrine and preferences are usually based on experience and current U.S. experience is based on exquisite systems.  These two powerful factors will make it difficult to dispassionately examine other options.  However, we must do so soon.  Our experience with the F-35 shows that the decision to pursue a different path needs to be taken before the new system gains a powerful constituency that insists it be built regardless of its capability.


T. X. Hammes is a Distinguished Research Fellow at the U.S. National Defense University.  The views expressed here are solely his own and do not reflect the views of the U.S. government, Department of Defense, or the National Defense University. 


Image: Adapted from D. Dibenski (public domain) and Reilly Butler (CC)