The U.S. Department of Defense Doesn't Innovate—and It's Missing the Robotics Revolution

I currently work in defense but work on open source projects on the side, some of which have morphed into something successful like Surfer. I've been reflecting on the fact that the type of things I build on the side would likely never see that light of day in terms of DOD funding... After musing with Grok a bit about this, it wrote me this essay which captures my sentiments rather well.

The U.S. Department of Defense (DoD) struts like it’s the godfather of innovation, dusting off DARPA’s ancient wins—ARPANET, GPS—to prop up its ego. Spoiler alert, Pentagon: your innovation engine’s been stalled for decades. While you’re fussing over fancier jets, China’s hijacking the robotics revolution, poised to dominate the future labor economy and leave the U.S. choking on their dust.

Take AI. OpenAI gambled $100M USD on ChatGPT’s 2021 training run, rewriting the tech playbook. DARPA, with its $4B budget, wouldn’t sniff a moonshot that "risky"—too abstract, too unproven. Now apply that to robotics. China’s “lights-out” factories, like Xiaomi’s, pump out smartphones 24/7 with zero humans, and their Unitree G1 humanoid—free of U.S. components—retails for a measly $16k. The U.S.? We’re shelling out 2.2x more to build a basic robot arm, leaning on Chinese parts slapped with “Made in USA” labels through supply chain trickery. China’s not just ahead; they’re rewriting the game.

The DoD’s problem is a culture allergic to risk. Program managers, sweating congressional oversight, greenlight safe bets—better radar, not game-changing robotics. It’s the same mindset as a $300M ARR company dodging $100k experiments to protect its cozy revenue stream. Meanwhile, China’s Made in China 2025 plan funnels billions into firms like Estun and UBTech, who are already mass-producing humanoids while Tesla’s Optimus gathers dust in a display case. China’s robotics flywheel—robots building robots, costs crashing, quality soaring—threatens to steamroll every industry, from electronics to healthcare.

This isn’t just about losing market share; it’s a national security gut-punch. If China locks in general-purpose robotics first, they’ll own global manufacturing, leaving the U.S. outgunned and outproduced. The DoD could’ve led, bankrolling the next ChatGPT or Unitree, but it’s too busy clutching its risk-averse pearls.

Get it together, DoD. Act like a venture capitalist: toss $5M at wild robotics startups, no strings. Partner with scrappy innovators before China scoops them up. Embrace the mess of failure like Silicon Valley does. Keep sleepwalking, and you’re not just failing to innovate—you’re handing China the future on a silver platter, one robot at a time.