When Palmer Luckey cofounded the defense startup Anduril in 2017, three years after selling his virtual reality startup Oculus to Facebook, the idea of a twentysomething from the tech industry challenging the giant contractors that build fighter jets, tanks, and warships for the US military seemed somewhat far-fetched. Seven years on, Luckey is showing that Anduril can not only compete with those contractors—it can win.

Last month, Anduril was one of two companies, along with the established defense contractor General Atomics, chosen to prototype a new kind of autonomous fighter jet called the Collaborative Combat Aircraft, or CCA, for the US Air Force and Navy. Anduril was chosen ahead of a pack of what Beltway lingo dubs “defense primes”—Boeing, Lockheed Martin, and Northrop Grumman.

“Anduril is proving that with the right team and business model, a seven-year-old company can go toe-to-toe with players that have been around for 70+,” Luckey wrote on social media platform X shortly after the contract was announced. The company declined to make anyone available for this article.

That business model has seen Anduril focus on showing that it can rapidly deliver drones, submarines, and other hardware infused with advanced software at relatively low cost. It also reflects a shift in America’s war-fighting outlook toward quicker development of less expensive systems that feature more software and autonomy.

Investors seem to think it’s working. Anduril has raised a total of $2.3 billion in funding, according to Pitchbook which tracks startup investment and, according to The Information, is seeking $1.5 billion more.

Anduril’s prototype CCA aircraft, named Fury, is still at an early stage of development. Another test aircraft will be developed by General Atomics, a 68-year-old defense firm with a history of making remotely operated systems that include the MQ-9 Reaper, which played a key role in the US expansion of drone warfare in the 2000s.

The US Air Force wants the new CCA drones to be more capable and more independent than existing uncrewed craft, which still depend heavily on ground staff. They are envisioned performing a wide range of missions, including reconnaissance, air strikes, and electronic warfare—either alone or in collaboration with aircraft piloted by a human or autonomously. A core part of the program is developing new artificial intelligence software to control the aircraft that can operate autonomously in a wider range of situations than existing military systems, which are typically autonomous only in narrow circumstances.

“This is a big shift,” says Stacie Pettyjohn, a senior fellow at the Center for a New American Security , a Washington, DC, think tank. She says that the US military has so far mostly used AI for target recognition and planning rather than for controlling systems. The CCA project is “a huge step forward for uncrewed systems and for the Air Force and Navy,” she says.

The CCA project is the culmination of years of work inside the Pentagon developing a vision of more automated aerial warfare. In 2014, the Defense Advanced Research Projects Agency conducted a study called the Air Dominance Initiative and concluded that a combination of next-generation fighter jets and uncrewed systems or “loyal wingman” capable of working in teams would be the surest way to gain an advantage in future conflict. The ultimate goal is for several drones, similar to those in development by Anduril and General Atomic, to accompany a conventional, crewed aircraft on a mission and collaborate in flexible ways.

The underlying philosophy is that on the battlefield there is safety, and overwhelming power, in numbers. Giving US pilots a clutch of robot wingmen is supposed to make them deadlier and more likely to return from missions unharmed. And the project is intended to be just the start of a bigger shift toward deploying autonomous aircraft in much larger numbers.

“The CCA represents a move toward swarms or at least larger numbers of uncrewed systems,” Pettyjohn says. “As a tactic, swarming could potentially allow smaller cheaper drones to overcome more expensive systems. It could be a game-changing asymmetric capability.”

The US pioneered use of military drones, which have become a priority for armed forces around the world. But Daniel Gerstein, a senior policy researcher at the think tank Rand Corporation and coauthor of a report published in February that warns swarms of drones could threaten US power, says the war in Ukraine has sparked new interest in nimbler forms of autonomy worldwide. Military planners are “getting really serious about unmanned,” Gerstein says, because the availability of cheaper autonomous systems is upsetting the balance of power in warfare. “Drones are here, and we’ve got to do something about them.”

Ukraine’s fight-back against Russia’s attempted invasion has functioned as a laboratory and proving ground for smaller, cheaper uncrewed systems, including aerial drones, surface vessels, and ground vehicles.

When the invasion began, remotely operated Bayraktar TB2 drones made by the Turkish company Baykar helped Ukraine mount an unexpectedly robust resistance against invading Russian tanks. Small so-called suicide drones, which attack targets by crashing into them, are used by troops on both sides. Ukrainian forces have access to US-made Switchblade models small enough to fit into backpacks. And they have also repurposed and adapted commercial technology, modifying agricultural and consumer drones to perform reconnaissance and bombing missions. These systems are all still remotely operated by humans.

Ukraine has also been a test bed for attacks using multiple drones working together in so-called swarms, which by weight of numbers can be difficult for air defense systems to track or attack. Both sides have deployed dozens of relatively small drones at a time, most likely still operated by human controllers. But the US and others are counting on more sophisticated AI to make it possible for even larger numbers of drones to collaborate in sophisticated ways.

Anduril, which has made software a central part of its products, appears to be trying to lay the groundwork for swarm warfare. The company’s Lattice platform can be used to connect and coordinate different sensors and weapons systems, providing an integrated visualization of a battlefield. Anduril now markets the platform’s ability to control a swarm of drones and has collaborated with another defense startup, Epirus, to offer a counter-drone system that uses powerful microwaves to neutralize swarms of drones. “The counter-drone swarm is a rapidly emerging emerging threat,” says Andy Lowery, CEO of Epirus. “We are geared particularly and especially toward combating drone swarms.”

The Pentagon’s vision for US drone swarms would eventually see them making decisions during a mission both individually and collectively, with little or no human input. “You will have communications and artificial intelligence mechanisms operating on all the parts of the swarm,” Gerstein of Rand says. “They could go to a target, select the target, and if they saw that they were under attack they could change their flight path.”

Such autonomy might test public attitudes toward the use of autonomy in weapons systems but it would be permitted under official policy. In January 2023, the Pentagon updated a directive on the acceptable limits of autonomy in weapons, which now states: “Autonomous and semi-autonomous weapon systems will be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force.”

The US Department of Defense is also investigating how autonomous aircraft could collaborate with sea-going drones. One US Navy project is testing whether autonomous surface vessels, submarines, and drones spot targets in the Red Sea and other parts of the Middle East.

Gerstein says there is some evidence that Israel has developed and deployed more intelligent drone swarms. The Israel Defense Forces used a swarm of AI-controlled drones to hunt Hamas militants in 2021, according to a report in New Scientist.

Although drone warfare has been pitched by militaries as more clinical than conventional weaponry, there’s also evidence US drone missions have caused significant civilian casualties and crossed legal boundaries. Some experts believe that deploying huge numbers of swarming autonomous systems will bring new risks.

“Swarming means more drones operating in more complex ways across more domains,” says Zachary Kallenborn, a drone expert affiliated with the Center for Strategic and International Studies, another think tank. “Swarming also means many more weapons operating without human control, creating more risk for error—and when the systems communicate, one drone’s mistake may cascade to a thousand more.”

The US appears committed to building and deploying more drones and giving them more complex autonomy. Air Force officials have said it will spend around $6 billion on the CCA program over the next five years, including future contracts. Another major US aerial drone project is the Replicator initiative, announced by Deputy Secretary of Defense Kathleen Hicks last August. It aims to spur the development of thousands of flight-ready autonomous systems and related technology within the next 18 to 24 months.

Anduril issued a lengthy statement applauding the initiative when it was announced last year, suggesting it hopes to secure contracts through the program. DefenseScoop recently reported that the company will be among the first to receive Replicator grants, for a surveillance system capable of tracking drones called WISP. Fury could just be the start of Anduril’s role in the Pentagon’s plans for the swarm era of aerial warfare.

Updated 5-28-2024, 9:20 pm EDT: This article was updated to correct a mispelling of Northrop Grumman.