Key Notes
- Jack Clark says AI could help expand annual US economic output to roughly $44 trillion by 2030.
- Anthropic’s highest-growth scenario depends on rapid adoption and is not a consensus economic forecast.
- Clark calls for international cooperation on AI rules as growth prospects raise questions about employment and shared prosperity.
Anthropic cofounder Jack Clark says artificial intelligence could help lift the US economy to roughly $44 trillion by 2030, linking the potential expansion to faster scientific discovery and increasingly capable AI systems.
Clark discussed the company’s economic modeling at the IMF–World Bank Annual Meetings in Bangkok, Reuters reported on October 17. He compared the potential outcome with an economy currently producing just over $30 trillion annually.
The figure describes an ambitious possibility rather than an established economic forecast. Clark paired his optimism about new industries and inventions with a call for countries to cooperate on stronger AI governance, arguing that the technology’s accelerating development warrants “unusual and serious rules.”
What the $44 Trillion Scenario Assumes
Anthropic’s published scenario model puts 2030 GDP at $34.1 trillion under modest change, $36.3 trillion under substantial change and $44.4 trillion in its extreme case. All three figures use constant 2025 prices.
The extreme result is 32.4% above the model’s no-AI baseline for 2030. The entire difference between today’s roughly $30 trillion economy and a future $44 trillion economy therefore cannot be counted as AI’s contribution.
That outcome assumes rapid adoption of AI capable of autonomously performing almost all knowledge-work tasks, likely requiring systems that improve their own successors. It also brings severe disruption for knowledge workers. The model excludes policy responses, business cycles and financial-market disruptions, among other forces that could change actual results.
Why Expectations Differ So Widely
A separate Anthropic survey report, released October 16, helps explain why the company’s researchers are considering such far-reaching outcomes. Surveys conducted in August covered 10,980 US adults and 1,334 Anthropic employees.
About 91% of staff expected AI to be capable of a Nobel-level scientific discovery by 2030, compared with 39% of US adults. Employees also anticipated faster adoption and more autonomous work than the public did.
When researchers fed respondents’ expectations into their economic model, the median implied annual GDP growth rate in 2030 was 17.1% for staff and 5.4% for the public. Those are model outputs based on respondents’ assumptions, not observed growth rates or an official company forecast.
The public survey was weighted to represent US adults, while the employee survey was unweighted. People working inside a frontier AI developer also have different experiences from the wider workforce, making the comparison informative without settling which outlook will prove correct.
Productivity Gains Must Reach the Wider Economy
Independent IMF research similarly identifies AI as a potential source of stronger productivity, while emphasizing uncertainty over the scale and timing of the gains.
The Fund highlights several conditions: technology must spread, businesses must reorganize production, complementary investment must arrive and workers must be able to move between jobs. A better-performing model does not automatically translate into higher output across every industry.
For investors, that distinction separates spending on AI capacity from the returns ultimately earned on it. As we reported, Nvidia Corp. (NASDAQ: NVDA) chief Jensen Huang has defended the industry’s longer-term prospects during the recent selloff. Clark’s macroeconomic scenario does not resolve the financing and profitability questions facing individual AI businesses.
Higher Output Does Not Guarantee Higher Pay
The IMF’s broader AI policy analysis stresses that workers’ ability to adapt will help determine whether productivity gains translate into broadly shared prosperity. Education, retraining and digital skills remain central to that adjustment.
The institution also warns that concentrated computing capacity and dependence on a small number of frontier models can widen gaps between economies. Its concerns extend to electricity supply, cybersecurity and financial stability, illustrating why a larger economy alone would not eliminate the transition’s costs.
Clark Calls for International Coordination
Clark leads the Anthropic Institute as the company’s head of public benefit. The institute combines research on model safety, societal effects and economics, connecting the company’s capability assessments with questions about governance and employment.
The official event program lists the Bangkok AI panel on October 16. The recording accompanies this article. Clark’s appeal for cooperation should be read as a policy position, rather than an announcement that governments have agreed on new global rules.
Anthropic itself is also preparing for a possible public listing, as we previously reported. Its own revenue, valuation and potential IPO are separate from the national GDP figures in this discussion. For households and investors alike, the question is how much of AI’s technical progress becomes durable income and productive economic activity.
