Speed, not sheer size, may determine the victor in the next major conflict. This central idea frames a conversation with Horacio Rozanski, CEO of Booz Allen Hamilton, and Gary Steele, CEO of Shield AI, who argue that the U.S. faces its most dangerous geopolitical moment in fifty years. The discussion revolves around readiness for a potential crisis over Taiwan, the transformative impact of autonomous systems observed in Ukraine, and the critical need to overhaul a sluggish defense procurement system. While both express cautious confidence in America’s current position—citing strong alliances, technological edge, and Taiwan’s own preparations—they stress that maintaining advantage hinges on moving faster than adversaries like China.

A significant portion of the dialogue dissects the lessons from Ukraine, which has served as a real-time laboratory for modern warfare. The conflict has underscored the vital importance of operating in GPS- and communications-denied environments, forcing rapid iteration and the agile integration of commercial technology. Open-source intelligence has proven nearly as valuable as classified channels due to its speed, while low-cost, attritable autonomous drones have redefined frontline tactics, reducing risk to human life and changing cost equations. This shift towards distributed, AI-enabled systems across all five domains (air, land, sea, space, and cyberspace) is now seen as the future of combat.

The most substantial barrier to capitalizing on this future, however, is identified as the U.S. government’s own procurement bureaucracy. Designed decades ago to minimize risk on platforms meant to last for generations, the system now stifles the speed essential for software-defined warfare. The CEOs advocate for a cultural and regulatory shift towards outcome-based contracts that allow for rapid experimentation and failure upstream, mirroring commercial practices and the iterative agility seen in Ukraine. Their companies’ partnership exemplifies the “all-of-nation” approach they believe is necessary: leveraging Silicon Valley’s innovation speed and the defense industry’s mission focus to field capable systems faster.

Surprising Insights

• Open-source intelligence (OSINT) can be as timely and valuable as classified intelligence. In modern conflicts, data from commercial satellites, social media, and cell phones often arrives faster than traditional classified feeds, making it crucial for rapid decision-making.
• The path to successful defense innovation requires accepting and accelerating failure. The current system, built to avoid failure at all costs, must learn to permit rapid prototyping and iteration—where multiple attempts fail quickly to find one that succeeds—to achieve necessary speed.
• A major bottleneck for U.S. military readiness isn’t just technology, but century-old procurement regulations. The core framework governing defense purchases is 75-100 years old, emphasizing cost-plus contracts and process compliance over outcomes, which is fundamentally mismatched with today’s pace of software and hardware innovation.
• Modern warfare increasingly depends on systems that can operate completely independently, without GPS or communications. The assumption of constant connectivity is a vulnerability; effective drones and other platforms must be able to navigate, make decisions, and complete missions autonomously in electronically contested environments.

Practical Takeaways

• Shift procurement focus from minimizing risk to rewarding outcomes. Government contracts should increasingly pay for specific, measurable mission results (e.g., a secured area, a disabled target) rather than micromanaging the development process and inputs.
• Adopt commercial buying practices for commercial technology. When purchasing commercial off-the-shelf software or hardware, the Defense Department should use standard commercial terms and upgrade cycles instead of layering on burdensome custom government requirements that nullify the speed advantage.
• Build partnerships that bridge innovation ecosystems. Traditional defense contractors, agile tech startups, and venture capital need to collaborate more seamlessly to pull the best commercial technology into national security missions at speed.
• Prioritize cybersecurity and “adversarial AI” testing from the start. For autonomous and AI-driven systems, building tamper-proof, resilient platforms and continuously testing them against cyber threats and algorithmic manipulation is as important as the core functionality itself.