The Third Offset Strategy: the impending mutations of American military power

Report: Written by Rian Whitton.

Since 2014, the Pentagon has been developing a new doctrine to combat America’s real and perceived vulnerabilities. It is this policy development, glossed over in major media outlets that will become the shape and character of America’s armed services for the foreseeable future.  

Named the third offset strategy, this doctrine is similar to the first and second offset strategies that defined US military procurement throughout the Cold War. First, Eisenhower hoped to offset Russian forces with increased nuclear stockpiles, after which the Reagan administration more successfully countered the threat of Soviet numerical superiority with advanced weaponry like precision-guided missiles, stealth aircraft and missile-defence systems.

Now, the Pentagon top brass argue the US is once again vulnerable; with the rise of China as a Pacific rival posing the most significant threat. While the previous strategies were made possible by huge increases in defence spending, the relative thrift of modern times has led to more novel pathways for America to maintain its military edge.

Above all, artificial intelligence, swarm doctrines and the disaggregation of expensive platforms into smaller, specialised and affordable systems are key themes in the early discussions about future force configuration. Such developments could significantly increase lethality, cheapen the cost of military production, and alter the ability of organisations, large and small, to make war.


We are used to hearing of an expert system such as IBM’s Watson or Deep Blue beating a human at Chess or Jeopardy. By weight and speed of processing power, computers can beat humans in narrowly set tasks, while in the chaos and uncertainty of situations like aerial dog fighting, human adaptability has proved more vital than computer software.

However, a little-known company called Psibernetix has recently caused a stir by building an artificial intelligence (ALPHA) that, through fuzzy logic, can beat experienced USAF pilots in combat simulations.  In fact, even when significantly handicapped, ALPHA has been able to outsmart and outmanoeuvre human pilots with no losses of its own.  Former Pilot Gene Lee referred to ALPHA as “the most aggressive, responsive, dynamic and credible AI I’ve seen to date.”

The potential for  AI in aerial combat is self-evident. Modern fighters are designed to operate at 40,000 ft. at speeds of over 1,500 mph (Mach 2 +). In such a scenario, speed of action and situational awareness is critical. ALPHA, which runs on novel language-based algorithms, can respond to changes in the environment 250 times faster than a human can blink. Despite this formidable processing power, the cutting edge technology can be run on a low-end PC.

Partially designed to operate unmanned aerial vehicles (UAV’s), it’s conceivable to imagine fleets of high-performance aircraft being controlled by algorithms, independent from human control.

However, this scenario is unlikely. ALPHA and its successor AI’s will instead be incorporated into existing and future aircraft, and will make the job of the pilot significantly easier, complimenting humans instead of supplanting them.  According to Nicholas Earnest (the creator of ALPHA), AI will begin to act as a digital assistant to pilots, easily processing incoming sensor data at speeds impossible to human cognition. Beyond the improved processing of big data, ALPHA has also demonstrated capability of remotely controlling up to four aircraft. This allows for one piloted plane to have control of multiple vehicles; acting as something of a pseudo-satellite.

While automation has up till now supplanted human tasks, the future of war will increasingly see human-machine teaming, where a hybridisation of machine and man create a strong and adaptable system.

While ALPHA provides an example of software designed to engage enemy fighters and supplement pilot performance, other systems are being developed. Mayhem is an AI that can quickly scour vast quantities of code to determine vulnerabilities thus highlighting and mitigating the risks of cyberwarfare.

ALPHA and Mayhem serve as milestones in the evolution of military-grade software. As targeting, electronic warfare, data-processing and situational awareness are all improved and further delegated to on-board systems, the pilot’s role could change drastically. Airmen may stop resembling Tom Cruise’s Maverick and consist increasingly of ad-hoc data scientists.

Swarms- small and many

While the revolution in AI and machine learning is a central pillar in the offset strategy, innovation in hardware will be just as vital.

One of the legacies of America’s preference for high-tech solutions has been increasing costs and development time for new military platforms, exemplified by air-superiority aircraft. The F-22 fighter took 25 years to develop, with each platform costing around $700 million. Of the 750 planned, only 138 were built before Congress shut down production in 2013. The newly introduced F-35 programme has cost close to $1 trillion, with myriad setbacks.

The expense of these aircraft systems has had serious repercussions. In 1984, Norman Augustine prophesised that the exponential increase in platform cost with the linear increase in budget would mean by 2054, America’s entire defence budget could pay for one plane. Though this is not quite the case, between 1996 and 2010, the number of fighter jets at America’s disposal dropped from 3,002 to 2,159. The reduction in platforms, coupled with higher cost, is causing much worry in America’s defence hierarchy.

As a result, think tanks have been pushing for the development of ‘small and many’ systems, or swarms. Secretary of Defence Ash Carter stated that swarm technology was a key part of a broader effort to adapt emerging technologies to existing and future war fighting needs.

Already, scientists are working on smaller drones whose algorithms allow them to coordinate in large swarms. T. X Hammes, a former US marine, argues that 3D-printing already allows for 10 small drones to be produced in a single day, and suggests that as printing technology improves, up to 1000 micro-drones could be produced every 24 hours (with one printer).

These drones would naturally be small, lacking in range and deficient in payload, making them no match for current air-superiority fighters. However, their sheer numbers have the potential to overwhelm adversaries in a number of ways. Airborne swarms of flying explosives could devestate enemy air-defences, while the US Navy has worked on underwater gliding drones that can operate for up to 5 years without refuelling. These systems could act as mines or self-propelled torpedoes, posing a threat to enemy surface and submarine vessels.

The increased potential of swarm technology is being enabled by aforementioned developments in AI, as well as additive manufacturing, compressed gel fuels and nanoexplosives. The production and deployment of many small, inexpensive systems in a short space of time would cut through the bureaucratic waste of the American defence industry, while significantly altering the nature and doctrines of the US military.

Barges- Big and Many

A swarm of semi-autonomous drones overwhelming Chinese air defences sounds exotic, but there are some difficulties. Particularly, the feasibility and small size of such systems limits their range and payload.

Another idea being touted is to leverage improvements in automation by building unmanned ‘arsenal platforms’ that carry huge quantities of missiles and smaller systems.

By the lowest estimate, a Tomahawk missile carrying 1000 pounds worth of explosives costs as little as $500,000. One B-2 bomber therefore costs up to 4000 Tomahawks. Strategists like Hammes and Paul Scharre have thus revitalised a fringe concept of large retrofitted aircraft or boats; specifically designed to carry as many missiles as possible. Given advances in automation, these enormous platforms would not even have to be manned. Space previously dedicated to crew would instead by focused on the number of missiles. By Scharre’s understanding, a retrofitted B-52 bomber could carry up to 66 air-to-air missiles, compared to the maximum 6 of current US fighters.

These platforms would likely not be manned. Representing huge targets and designed specifically to carry ordinance, they would represent highly visible and vulnerable targets. The ability of pilots to increasingly control multiple vehicles in the air thus allows ‘arsenal platforms’ to act in support roles, extended the persistence of the US military in potential conflict zones.

Automation is thus allowing the DOD to disaggregate expensive platforms that do anything for a huge price tag, into a variety of more simple systems that do one thing well. Foreseeing the future, there will be far greater variety in the type of military hardware being used; from miniature swarm vehicles, to submersible drone mines, to gigantic pilotless aircraft filled to the brim with missiles.

What does this mean?

The developments discussed are recent, and are not happening exclusively from other more traditional developments in the US military (i.e. more Aircraft Carriers and expensive bombers). Rather, the quantitative arguments for larger numbers of specialised weaponry is beginning to gain traction in the decision-making process.

At a time when devastating armaments and sophisticated platforms are becoming ever easier to access and manufacture, the highly expensive weapons and platforms of yesteryear are becoming increasingly unattractive to defence planners.

One final consideration should be the possibilities for smaller non-state actors to compete with states in the production of weapons systems. Difficulties that terrorists have had in mass producing explosive formed devices (EFP’s) have been overcome by the sophistication of 3D-printers. Meanwhile, commercial drones are becoming cheaper and more numerous, even as they become more sophisticated. Aerovel’s Flexrotor has an effective range of over 2000 miles (comparable to large military aircraft) yet only costs $200,000. The gulf in platform quality and volume between states and non-state actors, well defined for so long, is becoming increasingly tenuous.

The third offset strategy is America’s attempt to leverage new technology to maintain its military edge over rival states. However, technological innovation is increasingly democratising the ability to produce and use sophisticated and lethal weaponry in large quantities.

The production of nuclear weapons, fighter jets and aircraft carriers all necessitated enormous state resources, across military, technical and scientific sectors. States have thus enjoyed an enormous asymmetric advantage on the modern battlefield.

As sophisticated weaponry becomes cheaper, America’s ability to control the battle-space could become increasingly contested.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s