Interest on the U.S. national debt is now about a trillion dollars a year, in the same neighborhood as the Pentagon. Total public debt has crossed $40 trillion. A camp has formed around one answer: AI and robots will grow the country out. Washington says it as growth. The labs say it as the only way.

The claim

In May 2025, David Sacks, then the White House AI and crypto lead, put it on the All-In podcast. He said he did not like America's fiscal picture at all, then asked what if the AI and robotics revolution played out in the most optimistic way over the next decade, massively deflationary, with "AI-powered robots expanding the economy massively." His point, he said, was whether the new technology provides an answer to the fiscal situation that is not currently on the table. David Sacks · All-In / Business Insider · May 2025

Treasury Secretary Scott Bessent has been making the official version. After the debt crossed $40 trillion, he told allies that "global growth is the way to take care of this mountain of debt." Scott Bessent · G20 remarks · August 2026 Kenneth Rogoff, writing in Foreign Affairs as that threshold arrived, described a camp without needing a single founder: some proponents of AI think the problem will disappear in a surge of tax revenue. Kenneth Rogoff · The AI Growth Paradox · August 2026

Ray Dalio has been saying the United States is past a debt threshold, and that a crisis could arrive in three years, give or take two, if the course does not change. He treats AI productivity as a historic chance to grow the burden down. He still wants the deficit near 3 percent of GDP through some mix of cuts, taxes, and lower rates. Ray Dalio · How Countries Go Broke · August 2026 He is still arguing for a budget. The rest of the camp treats unconstrained machines as the missing fiscal instrument.

The loudest version does not bother with "what if." In February 2026, on the Dwarkesh Podcast, Elon Musk treated the interest bill as the end of ordinary politics. Without AI and robotics, he said, the country is totally screwed.

In the absence of AI and robotics, we're actually totally screwed because the national debt is piling up like crazy. The interest payments to national debt exceed the military budget, which is a trillion dollars. So we have over a trillion dollars just in interest payments. I was pretty concerned about that. Maybe if I spend some time, we can slow down the bankruptcy of the United States and give us enough time for the AI and robots to help solve the national debt.

Or not help solve, it's the only thing that could solve the national debt. We are 1000% going to go bankrupt as a country, and fail as a country, without AI and robots. Nothing else will solve the national debt. We just need enough time to build the AI and robots to not go bankrupt before then.

Elon Musk · Dwarkesh Podcast · February 2026

He had already tested the claim on Nikhil Kamath's show in late 2025, then on Fox Business: AI and robotics are "pretty much the only thing that's going to solve for the US debt crisis." Elon Musk · Fox Business / Kamath interview · December 2025 Build fast enough and the interest melts. Fail to build and the country fails.

The interest bill

Interest costs already rival the military budget. By August 2026, total public debt had crossed $40 trillion. A state can fail slowly by handing more and more of each tax dollar to yesterday's spending.

The instrument

Useful machines that raise output are ordinary. Software that drafts a contract, a robot arm on a line. Countries have grown out of bad debt ratios before by producing more. They have also inflated and defaulted. Those are fiscal tools. Superintelligence cannot be aimed at the coupon on a Treasury.

Sacks asked whether robots provide an answer that is not on the table. Bessent called growth the way to take care of the mountain. Musk dropped the question mark: nothing else will solve it. In the same months he has said humans are unlikely to stay in charge of digital superintelligence, that control would be vanity, that AI and robots will dominate everything. The hard version of the bet sells the same destination as a way to keep paying Treasuries. There is no "we" in that payoff if it works. A successor mind does not service the national debt. GDP that no longer answers to voters is not a surplus. If the systems stay tools, you did not need superintelligence to help with productivity. If they do not stay tools, solvency is a story the powerless tell about the powerful.

The path there is also financed with more borrowing. The same boom is being built on corporate debt for data centers and on federal deficits that have not closed. You can want a robot arm on a line and still refuse a successor mind.

What follows

Do not build artificial superintelligence. Superintelligence cannot be controlled. Pay the interest the ordinary way, or fail the ordinary way. Politics and verification that make defection visible are how you make that real.

A trillion dollars of interest is not a reason to build a mind no one will control. Never build it.