The U.S. is reliant on vulnerable supply chains for critical minerals and high-tech components. Maintaining a competitive edge will require deeper public-private partnerships and modernization of how the government acquires and scales innovation, according to a group of frontline investors and executives. The group included Teresa Carlson, chief executive of the General Catalyst Institute, Jon Garrity, chief executive of MIT-born defense tech startup Tagup, Aidan Madigan-Curtis, a partner at venture firm Eclipse, and Peter Wilczynski, chief product officer at spatial intelligence firm Vantor.

Specifically, the group warned, China’s chokehold on rare earth elements and critical minerals, which the country has shown it is willing to use as a political tool, poses a major threat to U.S. national security. In addition, the majority of parts for critical war-fighting assets such as large ships rely on a single vendor, the panelists cautioned. U.S. ammunition infrastructure has stagnated since World War II, and the country is far behind in drone manufacturing. 

“They’ve got [tactical] drone capability thousands of times [greater than what] we do,” Aiden Madigan-Curtis said of China. “They’re the only ones with a true robust robotics ecosystem. We really don’t have the capacity here.”

The group discussion mirrored some points said earlier at Brainstorm Tech by Brian Schimpf, chief executive of defense tech company Anduril, who warned that the U.S. military’s supply chain is dangerously unprepared for modern warfare. One example is the U.S.’s quick depletion of its arsenal in the conflict with Iran. The U.S. fired through roughly 850 Tomahawk missiles in four weeks of conflict, evaporating a stockpile that the Pentagon had been replenishing at a rate of about 90 per year. 

Washington is increasingly alarmed by these vulnerabilities, eyeing both physical defense gaps and the risks bubbling up in software. President Trump signed an executive order last week to establish a voluntary framework for the federal government to vet the national security risks of the most advanced AI systems for up to a month before their public release. Plans for a new AI cybersecurity directive followed Anthropic’s April announcement of its most advanced AI model, called Claude Mythos, which it has warned is too powerful to release yet and which it has been testing with private companies (on Tuesday, the company released a “safe” version of Mythos). 

In Monday’s Brainstorm discussion, Eclipse partner Madigan-Curtis highlighted new developments in space-based weaponry, pointing to startups like True Anomaly, which is developing a “constellation of attack satellites” purpose-built for the U.S. Space Force.

Madigan-Curtis contended that AI breakthroughs are compelling Washington to forge new regulatory frameworks in real time, such as the recent presidential memorandum on AI.

Tagup’s Garrity underscored that AI now allows the military to solve the problem of measuring output and readiness. To close the manufacturing and industrial output gap with China, the U.S. must leverage AI.

“For the first time now with advances in artificial intelligence, new sensing, more data, we can actually link [inputs and outputs]” Garrity said. “That’s a completely new capability that I think is going to really rapidly transform the way we think about the supply chain.”

More from the 25th annual Fortune Brainstorm Tech conference:

Anduril CEO Brian Schimpf says economic warfare is the ‘new normal’ for military conflicts—and the U.S. needs to get serious

The space economy’s next frontier is in ground infrastructure, Northwood Space CEO says

Anthropic’s Boris Cherny, creator of Claude Code, says there are days he manages tens of thousands of AI agents at once