A $1,000 drone can disable a million-dollar tank, and militaries now face a future where cheap, mass-produced systems, animated by advanced AI, have inverted the traditional economics of warfare. This conversation between Matt Cronin of a16z and drone industry pioneers Ryan Tsang (SHIELD AI) and Adam Breit (Skydio) delves into the urgent realities of this shift. The conflict in Ukraine has demonstrated that the decisive advantage now lies with the side that can deploy more autonomous drones, yet the United States is currently operating at a scale three orders of magnitude smaller than what’s being used on modern battlefields. The discussion centers on whether the U.S. can close this gap by leveraging its strengths in AI and software innovation while rebuilding a hollowed-out domestic manufacturing base for this critical technology.

The core challenge is twofold: industrial capacity and technological edge. China commands roughly 80% of the global drone market, built upon decades of dominance in consumer electronics manufacturing. The U.S. cannot match that industrial output overnight. Therefore, the path to competitiveness lies in marrying increased production of “attritable” systems—drones built for scale and acceptable loss—with superior autonomy. This means moving beyond drones that are merely remotely piloted to creating truly intelligent systems that can navigate, coordinate, and make decisions in GPS-denied, communications-jammed environments, a capability where American software expertise could provide a decisive advantage. As one founder noted, future conflict will resemble a “software writing fight,” where the speed of iterative updates in response to electronic warfare tactics will determine victory.

Beyond the technology, the conversation grapples with profound ethical and strategic questions. The “game theory” is stark: while the ideal is always keeping a “human in the loop,” defensive realities and the pace of autonomous engagements may force difficult choices. The participants argue that American leadership and values are essential for shaping the norms around these technologies. The ultimate goal is deterrence—using superior, precise capability to prevent conflict and, if necessary, to minimize suffering. The founders conclude that this mission, though extraordinarily difficult, is a foundational one for national security and stability, attracting those motivated by hard problems and profound purpose.

Surprising Insights

• Modern warfare as a software fight: The most critical battlefield advantage may soon be the speed at which engineers can push software updates to counter new electronic warfare tactics, sometimes requiring new code within 24 hours, as opposed to the Pentagon’s traditional multi-year cycle.
• Swarm tech spectacle vs. battlefield reality: While China’s massive, synchronized drone light shows are technically impressive, they rely on GPS and constant communication—two things easily jammed in combat. True battlefield-relevant swarm technology requires AI for navigation and decision-making without those crutches.
• Scale disparity is measured in orders of magnitude: The U.S. military procures small drones by the thousands, but evidence from Ukraine suggests future battlefields will require them by the millions annually—a thousand-fold difference in scale that current planning does not reflect.
• Failure as the best teacher: Skydio’s first drones sent to Ukraine failed immediately in the intense electronic warfare environment, a harsh lesson that ultimately drove more rapid and resilient innovation than traditional military requirements ever could.
• Autonomy can increase precision and reduce collateral damage: A counterintuitive argument was made that an AI-driven drone making some autonomous decisions to identify and strike a specific target could be more ethical than the status quo of dropping a 2,000-pound bomb that destroys an entire city block.

Practical Takeaways

• Shift procurement dollars to scale: Policymakers and military leaders should dramatically shift procurement spending from a handful of “exquisite” multi-billion-dollar platforms toward buying vast quantities of scalable, autonomous drone systems to achieve numerical and tactical superiority.
• Build for electronic warfare resilience from the start: Any drone system intended for modern conflict must be designed to operate without GPS and in heavily contested communications environments, using computer vision and onboard AI for navigation.
• Accelerate the software update pipeline: The Department of Defense must reform its acquisition and testing processes to allow for software updates at a pace measured in days or weeks, not years, to keep pace with adaptive adversaries.
• Use military purchasing power to rebuild industrial base: By committing to large-scale drone purchases, the U.S. government can provide the demand signal needed to catalyze domestic manufacturing for drones and the underlying robotics ecosystem.
• Engage with the problem despite its difficulty: For founders and engineers, the defense tech mission is described as brutally hard but uniquely meaningful, requiring a long-term commitment to solving complex challenges at the intersection of hardware, software, and national security.