No one has a good plan for how AI companies should work with the government

Read the full articleNo one has a good plan for how AI companies should work with the government on TechCrunch

What Happened

As OpenAI transitions from a wildly successful consumer startup into a piece of national security infrastructure, the company seems unequipped to manage its new responsibilities.

Our Take

OpenAI's a 7-year-old product company that somehow became national security infrastructure. That's broken, and nobody wants to admit it.

Neither OpenAI nor Anthropic wants real transparency or oversight—they just want the money and legitimacy. The government's equally lost. Everyone's pretending this was planned. It wasn't.

The real problem: there's no mechanism for this to actually work. You've got venture-backed companies, federal mandates, and nobody accountable to anyone.

What To Do

Don't expect government to figure this out—expect regulatory capture instead and plan your compliance accordingly.

Builder's Brief

Who

founders and executives at AI companies pursuing federal contracts

What changes

absence of established norms creates legal and reputational risk for any team in or entering the government market; governance structure becomes a procurement differentiator

When

months

Watch for

Congressional hearing specifically on AI contractor accountability standards producing draft legislation

What Skeptics Say

The premise that no framework exists ignores that traditional defense contractors navigated identical dual-use and ethics tensions for decades; OpenAI's governance gap is organizational immaturity, not a structurally novel problem, and framing it as unprecedented flatters the company while obscuring solvable institutional design questions.

Cited By

TechCrunch No one has a good plan for how AI companies should work with the government