Instead of a narrow 'defense tech' fund, General Catalyst invests through a wider lens called 'Global Resilience.' This thesis encompasses critical sectors like industrials, healthcare, and energy alongside defense, framing the investments around creating economically resilient and healthy societies to broaden their scope and appeal.

A significant paradox emerged from Norway's ethical guidelines: its sovereign fund was prohibited from investing in defense companies like Lockheed Martin, even while the Norwegian government was purchasing their F-35s for its own security. The war in Ukraine made this contradiction untenable, forcing a policy review.

Tech companies that refuse to work with the military are not taking a morally neutral position. They are making a moral choice to withhold technology that could increase precision, reduce civilian casualties, and protect service members. This abstention has real-world ethical consequences.

Lucrative civilian markets, not government deals, drive frontier tech. By making the defense side of a business a major political and legal liability, the Pentagon risks pushing top companies to completely shun government work, reversing a decades-long, successful dynamic for dual-use technology.

Investing in defense, energy, and public safety is not just another vertical. These foundational sectors uphold the stable democracy on which all other tech, like B2B SaaS, depends. A failure in these foundations renders investments in higher-level software and services worthless.

Investing in a hypersonic weapons company, once a career-ending move in Silicon Valley, is now seen as a crucial act of deterrence. This rapid cultural reversal, catalyzed by geopolitical events, signifies a profound sea change in the tech industry's values and its relationship with national security.

Seemingly reasonable terms like 'no autonomous lethal weapons' are impossible for a private company to enforce. They require moral and legal judgments about warfare—like defining a civilian or collateral damage—that are the exclusive and complex domain of a sovereign government, not a tech vendor.

The US government is labeling Anthropic a "supply chain risk" over ethical disputes while simultaneously using its AI model, Claude, for targeting and intelligence in strikes on Iran. This reveals a deep, contradictory dependence on the very technology it publicly rejects, undermining its own punitive measures.

The US Department of War is so committed to integrating AI into warfare that it blacklisted AI lab Anthropic for stipulating its models couldn't be used for autonomous weapons, revealing an intolerance for ethical limitations from suppliers.