If you doubt that A.I. is creating opportunities for new players to disrupt whole industries, look no further than Shield AI, which might just become the 21st century’s first new, big-time defense contractor.
The San Diego-based company was co-founded by ex-Navy SEAL Brandon Tseng. He served two tours in Afghanistan and other deployments in Asia and the Pacific, and says he saw firsthand soldiers risking—and sometimes losing—life and limb when entering compounds and buildings with little information about what lay around the next corner or behind the next the door.
He knew the right kind of reconnaissance technology could save lives. After an MBA at Harvard, he co-founded Shield AI along with his brother Ryan, a tech executive who had previously founded a wireless charging company acquired by Qualcomm, and Andrew Reiter, an engineer who had worked on computer vision applications for Draper Labs, the non-profit advanced engineering company in Cambridge, Massachusetts.
Shield’s first product is a small quadcopter drone, Nova2, that can be easily carried by a small infantry unit and deployed as a digital scout, even inside crowded apartment complexes or underground tunnels. Because Tseng knew that more sophisticated opponents were increasingly turning to GPS and radio jamming to protect themselves, Shield designed the Nova2 to operate autonomously, without the need for a human to control it remotely or for a constant connection to some distant datacenter. It also doesn’t need a GPS signal to navigate and map its environment.
Fleets of Shield’s drones can also operate in a swarm, coordinating with one another. That may be critical for keeping watch over a large base or mapping out a large set of buildings. And while the drones Shield currently makes aren’t armed, Tseng tells me the company has no objection to having weapons on board in the future. A single drone can always be shot down. A drone swarm, on the other hand, might be able to quickly overwhelm an enemy’s air defenses.
Shield’s Nova2 drones have already been used by the U.S. Special Operations Command since 2018, including in combat. That’s not bad for a startup defense contractor. But what really impressed me about Shield, and why I think it might just become the next big defense contractor, is the company’s vision. Incumbent defense contractors tend to think of their products as silos, each centered around a big piece of hardware designed to meet a particular customer spec, often with bespoke software designed from scratch just for that product. That’s not how Shield thinks, Tseng tells me. (Shield has also done some interesting things to keep its workforce connected and motivated during the pandemic, as my colleague S. Mitra Kalita chronicled here.)
Whether it is a small team of Army Rangers or a sortie of sophisticated F-18 fighter-bombers, or a Navy destroyer trying to get close to a dispute island chain, these units face a common overarching dilemma: the increasing need to operate in what are called “denied environments.” That means an area where, through some combination of terrain, digital and electronic warfare, and sophisticated weapons systems, the enemy can prevent U.S. forces from operating without incurring prohibitive losses. “The premise of the U.S. military that it can project power and maneuver in the environment freely, for the first time since World War II, that has been challenged by Russia and China,” Tseng says.
The solution to denied environments, Tseng says, is autonomy. Unmanned systems that can operate on their own, even when communication and navigation is disrupted, is a common piece of A.I. know-how that needs to be implemented across a wide range of military hardware and scenarios. This A.I.-first, A.I.-centric approach is not what one finds at Lockheed Martin or Raytheon.
Shield intends to bring this kind of autonomy to increasingly larger “platforms.” “We want to climb the unmanned systems food chain,” Tseng says. And Shield plans to get there fast, both through organic growth and acquisitions. The key is figuring out what elements can be shared across these platforms – such as algorithms for autonomously mapping and navigating a world—and what’s different.
To that end, last week the company announced it is buying Heron Systems, a small Virginia- and Maryland-based defense contractor that burst onto the scene last year when its software won a series of simulated dogfights against top human fighter pilots. Heron is known for its use of reinforcement learning, where A.I. software learns from experience, usually in a simulation, rather than from labelled data. While Shield already relies on sophisticated simulators and reinforcement learning to train its quadcopter software, Heron’s know-how about simulating the physics of fighter planes and other larger, fixed-wing aircraft, and the kinds of combat environments they operate in and threats they face, was what attracted Shield to Heron, according to Tseng.
Of course, there’s a big debate about exactly how disruptive A.I. will be to warfare. A provocative recent blog post by Jack McDonald, a professor of war studies at Kings College London, argued that rather than upsetting the balance of power, the technology might just be wash. Major powers will all invest in A.I. technology, meaning no one may gain a decisive upper hand. Combat, especially against non-state actors, is likely to be driven increasingly into urban areas, where military forces are more difficult to spot—even with technology like Shield’s Nova drones—and even harder to take out, without incurring unacceptable levels of civilian casualties.
Tseng’s retort to this: even if its true, the U.S. military can't afford not to invest in A.I. capabilities. “In this contest, you don’t want to be second place,” he says.
That’s a lesson about A.I. that business probably needs to take to heart too.
***
Before we get to the rest of this week's A.I. news, do you want to learn more about how A.I. is disrupting your industry and its potential to transform your company? You'll find plenty of insights at the inaugural Fortune Brainstorm A.I. conference, coming up November 8–9 at The Ritz-Carlton Boston. You'll hear cutting-edge case studies from senior executives who are using A.I. and meet the C-suite leaders, big tech companies, startups, and Fortune 500 businesses leading the charge in developing and deploying A.I. You can see the program and learn more about the event here. We'd love to see you there! Apply here to attend.
Jeremy Kahn
@jeremyakahn
jeremy.kahn@fortune.com
