Jack Clark, co-founder of AI lab Anthropic, told an audience of finance ministers and central bankers that artificial intelligence could lift US gross domestic product to roughly $44 trillion by 2030 — up from just over $30 trillion today.

Clark presented the projection on October 17, 2026, at the annual meetings of the International Monetary Fund (IMF) and the World Bank in Bangkok, pairing the forecast with a call for international regulatory cooperation on AI.

Three Scenarios, One Striking Ceiling

The projection is drawn from Anthropic's internal economic model, which the company released in September 2026 as an economic scenarios explorer. The model maps three possible paths for the US economy by 2030:

• Modest: US GDP reaches $34.1 trillion, reflecting 1.6% growth.
• Substantial: GDP rises to $36.3 trillion, reflecting 8.3% growth.
• Extreme: GDP hits $44.4 trillion, a 32.4% boost.

The extreme scenario assumes AI attains high levels of autonomy and efficiency, and it requires recursively self-improving AI systems — AI capable of improving its own designs, with each round of gains feeding into the next — to deliver the full effect. The spread between the lowest and highest paths exceeds $10 trillion in annual output by 2030.

A Shorter Timeline for Self-Improving AI

Clark also pulled forward his own estimate for when that capability might arrive, saying he now expects full recursive self-improvement by 2027. He tied the updated estimate to faster progress in AI technology — a point that matters because the extreme GDP scenario depends on precisely this capability materializing.

Public expectations, however, sit below the most aggressive outcome. Surveys indicate that respondents' views align most closely with the substantial growth scenario, with those polled estimating a 10% GDP boost alongside a 5% unemployment rate.

Why the IMF Audience Was a Pointed One

Discussions at the meetings reflected broader concerns over how financial and employment policy should adapt to AI. For the finance ministers and central bankers in the room, GDP trajectories are not abstract: growth assumptions feed into revenue projections, debt sustainability, and the fiscal room governments have to respond to downturns. Anthropic's model points to concrete repercussions from heavier automation, including heightened displacement risks for knowledge workers and, in extreme cases, a surge in unemployment.

What It Means for Policymakers

For policymakers, the core message is that planning around a single outcome is not enough. Anthropic's framework describes three very different futures, each demanding a different response: a modest path to $34.1 trillion would look like steady progress, while an extreme path to $44.4 trillion — paired with possible unemployment spikes — would test social safety nets, tax systems, and labor policy all at once.

That context underpins Clark's push for international cooperation. The research is intended to fill gaps in what is known about AI's economic impact and to spark debate over what legislative frameworks should look like.

There is also the question of who is producing the forecast. Anthropic builds frontier AI systems, meaning a model projecting large economic gains from AI comes from a company with a direct stake in how the technology is perceived and regulated.

The key variable to watch is the 2027 self-improvement timeline, the hinge between the substantial and extreme outcomes. The second is labor data: if displacement among knowledge workers begins to appear in employment figures, the debate could move quickly from conference panels to legislation.