AI & Economy News — 2026-09-22 (Tue) Evening News

Overview

Jamie Dimon (JPMorgan CEO) hinted that hyperscaler AI capex could reach $1 trillion next year, raising both caution and optimism about the scale of capex (CNBC, finance.biggo.com, LinkedIn). While AI, tariffs, and Iran are in focus ahead of the US-China summit (CNBC and others), friction over power infrastructure has also surfaced, with the state of Texas suspending permits for AI data centers (조선일보).

Macroeconomic/growth trends

Bank of Australia Governor Michele Bullock said AI could be a bubble and has not yet contributed to Australia's productivity gains (theguardian.com). JPMorgan's Dimon said on LinkedIn that hyperscaler AI investments could be on a scale that could boost GDP by 1% annually. In the United States, there are reports that rising inflation and concerns about AI are affecting economic sentiment ahead of the midterm elections (The New York Times). It was also reported that Australia is betting on AI as a multi-decade growth strategy (Bloomberg).

Productivity trends

Data shows that while the number of new graduate hires is decreasing, the number of senior hires is increasing at companies that have adopted AI (BigGo Finance). On the other hand, there have been reports of concerns that AI is undermining worker skills (Korea JoongAng Daily, THE Journal), and there are also debates about whether AI can be a solution to America's workforce challenges (Forbes, nationalaffairs.com). There is still little evidence of productivity improvement, and we are at a stage where both concerns and expectations coexist.

Trends in employment and labor market

There are reports that AI is creating a "recession-like" job market for new graduates (Inside Higher Ed). South Korea is planning a dashboard to detect AI-related employment shocks (Digital Watch Observatory). A behavioral scientific analysis found that workers are wary of AI (Hawaii Public Radio). Managing AI teams is becoming a new job skill, but some say it's not yet accompanied by compensation (Business Insider).

Trends in energy/power demand

Regulatory responses are divided, with Texas suspending new permits for AI data centers and California requiring information disclosure. ABB announces DC power portfolio for AI data centers (Engineering.com). The EU is working to understand the reality of AI's electricity demand (Cybernews). Some analysis suggests that the data center construction boom is reshaping America's infrastructure (qz.com, Emerald Book).

Remarks by central banks, international organizations, and VIPs

- Kristalina Georgieva (IMF Managing Director): At the Qatar Economic Forum, she cited inflation, debt service costs, and AI investment risks as challenges (IndexBox). - Michele Bullock, Governor of the Reserve Bank of Australia, not on the tracking list but related, says AI could be a bubble and has not yet boosted Australia's productivity (theguardian.com). - Specific statements regarding other tracked targets (Lagarde, Ueda, Bailey, Cormann, Banga, Acemoglu, Brynjolfsson) are not found in the headings of this list.

Trends in corporate capital investment

Jamie Dimon, JPMorgan CEO, said hyperscaler AI infrastructure spending could approach $1 trillion next year, up from $700 billion this year, which would boost GDP by 1% annually (CNBC, finance.biggo.com, LinkedIn). Goldman Sachs analyzes that Microsoft's AI gross profit margin has reached the same level as its cloud business, supporting AI transformation (Fukashio TechFlow). Although a $2.7 trillion AI boom is underway, it has been pointed out that only a small portion of the funds are allocated to the models themselves (Redmond Channel Partner). The CIO of Texas Teachers (a pension fund) warned that the AI capital investment boom is similar to past bubble bursts (PitchBook). Nvidia is reportedly looking to acquire Hugging Face for $13 billion to secure the "next AI economy" (Forbes JAPAN).

Trade and industrial policy

AI is expected to be a major topic of discussion at the Trump-Xi Jinping meeting, along with tariffs and Iran (CNBC, Kursiv Media, Oz Arab Media, myMotherLode.com). DeepSeek is said to be moving to Huawei chips for AI learning in order to reduce dependence on Nvidia (조선일보).

Integrated analysis: What is happening now

Pairing headlines on capital spending (JPMorgan, Goldman Sachs, $2.7 trillion boom) and power infrastructure (Texas license suspension, ABB, EU) suggests that the scale of AI investment is already starting to hit the limits of physical power supplies. On the other hand, on the macro level, Mr. Bullock and Mr. Georgieva of the IMF have mentioned the AI bubble and investment risks, and central banks and international organizations are expressing caution about rapidly expanding capital investment. On the employment front, there are concerns about the decline in the number of new graduates being hired and the deterioration of skills, and while there is still little empirical support for productivity, it appears that only investment is taking the lead. Furthermore, as AI is treated in conjunction with tariffs and chip regulations in the US-China talks, geopolitical risks in capital investment (such as DeepSeek's migration to Huawei) are becoming more apparent.

Future points of interest

- Will electricity infrastructure regulations such as the suspension of permits in Texas spread to other states and countries? - Will the gap between hyperscaler investments, which are said to be on the scale of "$1 trillion", and actual AI revenues (such as Microsoft's gross profit margin) widen or narrow? - How will the results of the US-China summit be reflected in AI-related tariffs and export regulations (including chips)?