Tesla and SpaceX CEO Elon Musk offered a rare, systematic reflection on his experience entering the Trump administration and leading the Department of Government Efficiency (DOGE) in an upcoming podcast, bluntly stating that the path of cutting spending from within the government to resolve the US national debt crisis has failed, and asserting that the only way out lies in artificial intelligence and robotics.

Musk made these remarks in a conversation with biographer Walter Isaacson. The new podcast, titled The Builders with Walter Isaacson, is set to launch next Tuesday (October 20). According to US news outlet Axios, Musk told Isaacson in the interview that the DOGE experiment had failed. “It turns out you can’t solve the budget deficit problem from within the government system. It just doesn’t work.”

“The only way to solve the US national debt problem is through AI and robots. The old approach cannot solve this problem,” Musk said. He added: “We can’t solve the problem simply by cutting spending.”

The US national debt has now surpassed $40 trillion, with Treasury Secretary Scott Bessent saying in early October that it had reached approximately $41 trillion. Musk’s remarks effectively amount to an admission that DOGE’s efforts to cut spending and curb waste from within the federal government fell short of expectations.

Sharp Criticism of Washington’s Political Ecosystem

In the interview, Musk expressed near-total disillusionment with the willingness to cut spending within the American political system. He argued that from the president to both chambers of Congress to the judiciary, there is a lack of genuine will to reduce government spending.

“The president, the House, the Senate, the courts—none of them have this in mind,” Musk said bluntly. “Frankly, almost no one in government genuinely wants to cut spending. There are only a handful of people in the world who do, like Republican Senator Rand Paul of Kentucky.”

When discussing President Trump, Musk’s analysis was more nuanced. “Maybe if the president made this his top priority, we could do it,” he said. “If you try to solve the deficit problem and eliminate wasteful spending, you make a lot of enemies. President Trump wants to be liked. He doesn’t like people blaming him or being upset with him because certain funding was cut.”

This observation reveals Musk’s deep understanding of White House decision-making logic: cutting spending politically means creating adversaries, which structurally conflicts with Trump’s own tendency to seek political popularity.

“Stayed in the White House Too Long”

Beyond his criticism of fiscal matters, Musk’s reflection on his political experience over the past year is equally striking. He admitted he had “gotten a bit sucked in” and had been away from his companies for too long.

“I should have gone back sooner and focused on AI,” Musk said. He stated that if he could choose again, he should have stayed more in the background and limited his formal government role to just one or two months, rather than the four months he actually served.

When Isaacson asked whether working in the West Wing was somewhat intoxicating, Musk did not shy away from the feeling. “A little bit, yeah. Because being in the Oval Office and witnessing those major decisions being made is actually pretty cool,” he replied.

However, reflecting on his Washington days does not mean Musk is exiting American politics. According to campaign finance disclosures and statements from a spokesperson for Musk’s political action committee, his political spending in this US election cycle totals approximately $366 million, with at least $276 million flowing to Republican-aligned organizations and his own super PAC since July of this year.

The Logic and Pitfalls of AI Solving the Debt Problem

Musk’s framing of AI and robots as the only solution to the US national debt problem echoes Bessent’s thinking. Bessent has previously said that if the US economy grows at 3%, it could grow its way out of the debt predicament. The logic is that AI-driven productivity gains can expand the size of the economy, increase corporate profits and government tax revenue, thereby easing debt pressure.

Musk’s own explanation of this mechanism is more specific. He believes that AI-led automation could lead to a surge in output, causing the supply of goods to grow faster than the money supply, which would trigger deflation; meanwhile, productivity gains could drive economic expansion and increase tax revenue, helping to curb the widening budget deficit and debt burden.

Musk’s chain of reasoning can be summarized as follows:

Jack Clark, co-founder of AI company Anthropic, also painted an optimistic picture of AI-driven economic growth on Saturday at the IMF-World Bank annual meetings. According to Anthropic’s economic scenario models, in an extreme scenario where AI capabilities advance rapidly and are widely adopted, US GDP could reach $44.4 trillion by 2030.

Anthropic’s Economic Scenario Explorer divides the 2030 US economy into three paths, with Clark citing the most optimistic one:

Scenario2030 US GDPvs. No-AI BaselineModerate$34.1 trillion+1.6%Significant$36.3 trillion+8.3%Extreme$44.4 trillion+32.4%

GDP figures in the table are in 2025 dollars; the “no-AI baseline” refers to a counterfactual path under the same population and productivity assumptions but without accounting for AI diffusion effects.

However, this outlook is not Anthropic’s baseline forecast, but rather rests on assumptions that AI can autonomously complete the vast majority of knowledge work and rapidly integrate into economic activity. Clark also acknowledged that such staggering economic growth would require a continuous stream of new industries and new inventions.

International Monetary Fund (IMF) Managing Director Kristalina Georgieva said on October 14 at a panel discussion during the IMF-World Bank annual meetings in Bangkok that growth “certainly helps, but it cannot be fast enough to bring about the necessary adjustment,” and that countries still need to introduce specific fiscal measures. In its latest World Economic Outlook, the IMF forecasts US economic growth of 2.3% for both 2026 and 2027, below the 3% threshold set by Bessent, and estimates that total US government debt will reach 142% of GDP by 2031.

The US Treasury Department does not share this assessment. David Zervos, an adviser to the Treasury Secretary, said in an email to media that the IMF’s forecasts have been “historically quite messy,” citing multiple instances of inaccuracy including 2008 global growth, the 1997 Southeast Asian financial crisis, and Russia’s economy in 2022. Market pressure is more direct: the 10-year US Treasury yield stood at approximately 5.2% on October 14, near its highest level since 2002.

More critically, a larger economy does not automatically mean the fiscal deficit disappears. The US Congressional Budget Office projects a federal fiscal deficit of $1.9 trillion in fiscal year 2026, widening to $3.1 trillion by 2036. Even if AI does boost productivity, whether the government can effectively convert new economic activity into tax revenue remains an open question.

Anthropic’s model reveals a structural shift worth watching: in the extreme growth scenario, capital’s share of GDP income could rise from roughly 40% currently to 54.8%, while labor’s income share falls to 45.2%. If AI primarily boosts corporate profits by substituting for labor without creating enough new jobs and labor income, the new wealth may not generate a corresponding scale of tax revenue. Meanwhile, massive investments in data centers, chips, and power infrastructure, as well as rising government debt interest payments, could offset part of the growth dividend.

This Washington debate is hardly a distant spectacle for Taiwan. US Treasury International Capital (TIC) data shows that Taiwan held $298.4 billion in US Treasuries as of August 2026, ranking tenth among foreign holders globally. The trajectory of US fiscal policy and Treasury yields directly affects the valuation of Taiwan’s life insurance industry’s massive US dollar bond positions, as well as the central bank’s foreign exchange reserve asset allocation.

Bessent revealed on the 9th of this month at a policy forum in Washington that the US government is formulating a fiscal consolidation plan to reduce the widening fiscal deficit and ease pressure on Treasury issuance, with the plan expected to be announced after the midterm elections. This timeline means that the debate over whether AI can serve as the “antidote” to America’s debt problem will continue until the fiscal consolidation plan is formally unveiled.