It is hard to remember a time when it did not feel like we were inundated with constant change and ‘unprecedented’ events. Ever since the initial stages of the COVID pandemic things have been in flux when it comes to market volatility, jobs, inflation, and technological advance. Some of you may see this more recent phenomenon with your children, relatives, or friends that are recent college graduates – increased unemployment. The trend and proliferation of Large Language Models (LLMs) and AI Agents into the workforce is a likely catalyst to this recent trend. AI is automating foundational, early career tasks that effectively eliminate some of the first few rungs on the career ladder. However, understanding this structural change is complicated by the policy uncertainty of Washington, leaving analysts and the Federal Reserve ‘flying blind’ in terms of getting current economic data.
Due to the government shutdown, we don’t have the most recent jobs numbers as I write this. With that caveat, there is a noticeable shift in the unemployment rate of recent college graduates. In the most current data, recent college graduates – defined by college graduates between 23 and 27 – have a 4.59% unemployment rate. Not only is that higher than the overall unemployment rate, but it is roughly forty percent higher than the 3.25% figure in 2019.[1] When companies lean into AI usage it often leads to increased productivity due to automation of specific tasks, allowing employees to put focus elsewhere.[2] While this is good for the bottom line, it often displaces entry level roles that focus on data processing – something that LLMs are currently fine-tuned for. The net effect is less hiring in entry level roles, which stunt career development and may shrink the future pipeline of new, competent workers.
AI technology is promising and is making many tasks easier to perform, allows workers to focus on revenue generating tasks, and may help boost productivity. With that in mind, there is chatter about AI potentially being another bubble, similar to dot com, but on a higher scale. The magnificent seven are directly or tangentially heavily invested in the AI infrastructure and innovation buildout. They also represent a significant percentage of the S&P 500 total weight (this varies based on market movement but has been between 35 and 40%).[3] A bubble does not suggest that AI is a dud of technology. We can see that it is useful. Similarly, the dot com bust did not predict online retail would not be successful in the future. In many ways it was ahead of its time, but consumers were not ready. The bubble potential is multifaceted, but comes down to capital expenditure vs revenues, depreciation schedules vs creative destruction, off balance sheet liabilities,[4] and energy requirement vs capacity.[5]
The first factor is revenue vs capital expenditures. According to Morningstar, we will need roughly $2 trillion in revenue to justify spending by 2030. It is currently estimated to be close to $20 billion. This is a sizeable gap to cross even factoring in an exponential growth rate. Another issue is depreciation schedules. Companies are pouring money into buying the latest chips, which have useful lives of just a few years, but are being depreciated 5-6 years. This opens a large accounting gap where the true costs are getting kicked down the road and not being recognized in the near term.[6] Additionally, there is a circular nature to some of the AI investment. One example is NVIDIA investing billions in OpenAI which is being used to buy chips from NVIDIA. Due to concentration in data center proximity, there is also a strain on localized power grids that is only expected to increase through 2030.[7]
Like many times in the last five years, we are left with a collision of forces: an economic challenge for recent college grads to get a foothold in the workforce, a trillion-dollar technology revolution built upon an aggressive if not unstainable financial foundation, and a central bank that is operating during an information blackout. The pressure on the labor market is likely to continue and evolve as policy and AI outcomes are still unclear.
[1] Ozkan, Serdar, and Nicholas Sullivan. “Recent College Grads Bear Brunt of Labor Market Shifts.” On the Economy. Federal Reserve Bank of St. Louis, August 25, 2025. https://www.stlouisfed.org/on-the-economy/2025/aug/recent-college-grads-bear-brunt-labor-market-shifts
[2] Murray, Seb. “How Artificial Intelligence Impacts the U.S. Labor Market.” MIT Sloan School of Management. https://mitsloan.mit.edu/ideas-made-to-matter/how-artificial-intelligence-impacts-us-labor-market
[3] Daly, Lyle. ” The Magnificent Seven’s Market Cap Vs. the S&P 500″ The Motley Fool. https://www.fool.com/research/magnificent-seven-sp-500/
[4] Farooque, Faizan. “Meta and Musk’s xAI Tap Off-Balance Sheet Financing for Billions in AI Infrastructure.” Yahoo Finance, November 10, 2025. https://finance.yahoo.com/news/meta-musks-xai-tap-off-175253347.html
[5] Swedroe, Larry. “Why the AI Spending Spree Could Spell Trouble for Investors.” Morningstar, November 3, 2025. https://www.morningstar.com/markets/why-ai-spending-spree-could-spell-trouble-investors
[6]Fox, Matthew. “There’s a hidden risk lurking for AI stocks in 2025.” Business Insider, October 22, 2024. https://markets.businessinsider.com/news/stocks/ai-stocks-risk-nvidia-gpus-blackwell-tech-outlook-depreciation-accounting-2024-8
[7] Leppert, Rebecca. “What we know about energy use at U.S. data centers amid the AI boom.” Pew Research Center, October 24, 2025. https://www.pewresearch.org/short-reads/2025/10/24/what-we-know-about-energy-use-at-us-data-centers-amid-the-ai-boom/
