The U.S. Military Is Not Ready for the New Era of Warfare
By Raj M. Shah and Christopher M. Kirchhoff
Mr. Shah is the managing partner of Shield Capital. Dr. Kirchhoff helped build the Pentagon’s Defense Innovation Unit.
The First Matabele War, fought between 1893 and 1894, foretold the future.
In its opening battle, roughly 700 soldiers, paramilitaries and African auxiliaries aligned with the British South Africa Company used five Maxim guns — the world’s first fully automatic weapon — to help repel over 5,000 Ndebele warriors, some 1,500 of whom were killed at a cost of only a handful of British soldiers. The brutal era of trench warfare that the Maxim gun ushered in didn’t become fully apparent until World War I. Yet initial accounts of its singular effectiveness correctly foretold the end of the cavalry, a critical piece of combat arms since the Iron Age.
We stand at the precipice of an even more consequential revolution in military affairs today. A new wave of war is bearing down on us. Artificial-intelligence-powered autonomous weapons systems are going global. And the U.S. military is not ready for them.
Weeks ago, the world experienced another Maxim gun moment: The Ukrainian military evacuated U.S.-provided M1A1 Abrams battle tanks from the front lines after many of them were reportedly destroyed by Russian kamikaze drones. The withdrawal of one of the world’s most advanced battle tanks in an A.I.-powered drone war foretells the end of a century of manned mechanized warfare as we know it. Like other unmanned vehicles that aim for a high level of autonomy, these Russian drones don’t rely on large language models or similar A.I. more familiar to civilian consumers, but rather on technology like machine learning to help identify, seek and destroy targets. Even those devices that are not entirely A.I.-driven increasingly use A.I. and adjacent technologies for targeting, sensing and guidance.
Techno-skeptics who argue against the use of A.I. in warfare are oblivious to the reality that autonomous systems are already everywhere — and the technology is increasingly being deployed to these systems’ benefit. Hezbollah’s alleged use of explosive-laden drones has displaced at least 60,000 Israelis south of the Lebanon border. Houthi rebels are using remotely controlled sea drones to threaten the 12 percent of global shipping value that passes through the Red Sea, including the supertanker Sounion, now abandoned, adrift and aflame, with four times as much oil as was carried by the Exxon Valdez. And in the attacks of Oct. 7, Hamas used quadcopter drones — which probably used some A.I. capabilities — to disable Israeli surveillance towers along the Gaza border wall, allowing at least 1,500 fighters to pour over a modern-day Maginot line and murder over 1,000 Israelis, precipitating the worst eruption of violence in Israel and Palestinian territories since the 1973 Arab-Israeli war.
Advertisement
Yet as this is happening, the Pentagon still overwhelmingly spends its dollars on legacy weapons systems. It continues to rely on an outmoded and costly technical production system to buy tanks, ships and aircraft carriers that new generations of weapons — autonomous and hypersonic — can demonstrably kill.
Take for example the F-35, the apex predator of the sky. The fifth-generation stealth fighter is known as a “flying computer” for its ability to fuse sensor data with advanced weapons.
Yet this $2 trillion program has fielded fighter airplanes with less processing power than many smartphones. It’s the result of a technology production system bespoke to the military and separate from the consumer technology ecosystem. The F-35 design was largely frozen in 2001, the year the Pentagon awarded its contract to Lockheed Martin. By the time the first F-35 was rolling down the runway, technology’s state of the art had already flown far past it. This year, the iPhone 16 arrives. Today, the F-35 is slowly progressing through its third technology upgrade with newer, but far from state-of-the-art, processors. The core issue is that this slow hardware refresh cycle prevents the F-35 from fully taking advantage of the accelerating advancements in A.I.
This is not an either/or argument. iPhones will not replace F-35s. The U.S. military requires unique platforms, such as stealth fighters and submarines, as well as newer technologies, including drones. All weapons systems, old or new, need to take full advantage of the software and A.I. revolution — a revolution driven forward primarily by Silicon Valley, not by large, traditional defense contractors.
There is progress. Merging these two systems of technological production — one for the military, another for everything else — is now a top Pentagon objective. Started in 2015, the Pentagon’s Defense Innovation Unit, based in Silicon Valley, brings innovative commercial technology into the military, including A.I. It functions more like a venture capital firm than a military program office. D.I.U. created a faster way for startups to contract with the military that today has been used by the Department of Defense to acquire $70 billion of technology. (We helped build D.I.U.; one of us, Mr. Shah, now runs a venture firm focused on national security startups, including some that have received federal funding.)
A new generation of defense unicorns powered by this investment are creating advanced A.I.-powered autonomous systems. Joby Aviation has deployed an electrically powered S4 flying air taxi. Anduril Industries just advanced to the final round in the Air Force’s collaborative combat mega-contract, in which 1,000 advanced stealth drones will fight alongside manned fighters. D.I.U. is also leading the Pentagon’s high-profile Replicator initiative, developing swarming autonomous weapons for air, land and sea.
Yet there is much left to do. While D.I.U.’s budget is more than 30 times as large as it was in its first full year of operations, now totaling almost $1 billion annually, the Pentagon spends only pennies on innovation for each dollar it throws at legacy systems. The Replicator initiative accounts for just 0.059 percent of the defense budget at a time when our adversaries are making monumental shifts.
China, of course, doesn’t need a D.I.U.; Xi Jinping and his predecessor, Hu Jintao, mandated that civilian technology be available to the People’s Liberation Army. This top-down, state-run economy is chasing quantum computers, nuclear-capable hypersonic weapons, and lofting into orbit its own 13,000-satellite equivalent to Starlink.
This is the civilizational race we’re in.
The way to win against both China and low-cost weapons in Ukraine and the Middle East is to unleash our market-based system so that scrappy, fast-moving product companies and the venture funds that back them revitalize our military’s technology pipeline. The good news is that market interest is robust: Venture capital funds deployed $120 billion of capital into national security startups over the last three years. Leading engineers are eager to work on problems that matter to preserving democracy. The question now is whether we can achieve this transformation in time to deter the next great power war and prevail in the more contained conflicts that threaten to envelop the U.S. and our allies.
“The history of failure in war can almost be summed up in two words: too late,” Douglas MacArthur declared hauntingly in 1940. Eighty-four years later, on the eve of tensions not unlike what preceded prior great power conflict, we would do well to heed MacArthur’s warning.
Comments
Post a Comment