Former Treasury Chief Warns AI Boom Shows 2008 Warning Signs
Hank Paulson looked at the AI spending boom and recognized a pattern he watched up close in 2008. His warning has nothing to do with a crash, and everything to do with which rope you happen to be tied to.
The man who ran Treasury during the last real financial crisis went on CNBC Tuesday from the Aspen Economic Strategy Group, sitting alongside Tim Geithner, and refused to pick a side. Hank Paulson steered a middle course between Bezos-style AI utopianism and Amodei-style civilizational alarm. He looked at the AI capex boom and saw a pattern from 2008, echoing its shape rather than its crash.
That shape is money moving in a circle. For a retail investor watching NVIDIA (NASDAQ:NVDA | NVDA Price Prediction)-fueled index gains alongside a job market where JOLTS openings just fell to 7.36 million from 7.585 million in April, the Paulson framing matters.
The Circular Money Loop Paulson Is Watching
Here is the setup in his own words. “The suppliers of chips and compute, you know, making investments in the customers. So they’ve got the capacity to buy and customers that need the chips, making investments in some of the suppliers to make sure they’ve got them.” He was careful to say nothing nefarious is happening. Everyone acts rationally inside their own contract.
The problem is what happens when the same dollar shows up on three different balance sheets. A chip vendor invests in an AI lab. The AI lab uses that money to buy chips from the vendor. The vendor books revenue, its stock rises, and it recycles the gain into the next round of customer financing. When demand is real, this looks like a flywheel. When demand softens, the same wiring runs in reverse, because the supplier’s revenue, the customer’s spending power, and the equity valuation propping up the arrangement are the same dollar wearing three hats.
Think of climbers roped together on a cliff. Fine on the way up. When one slips, everyone learns how tightly they were tied. Paulson’s point is that risk is concentrated inside the loop, which is why he says a correction is inevitable even if he is not calling a date.
The White Collar Jobs Shock
Then he turned to labor. “In terms of white collar workers, okay, there’s going to be a lot of a lot of displacement and it’s going to happen so quickly.” That word, quickly, is the load-bearing one.
Right now the aggregate labor picture looks fine. Unemployment sits at 4.2% as of June, and nonfarm payrolls hit 158.984 million in June, a fresh high. Aggregates lie about composition. If AI eats a specific slice of legal review, marketing copy, first-pass coding, and mid-level analyst work over eighteen months, the top-line unemployment number stays tame while entire career ladders quietly disappear. For a reader whose paycheck depends on producing exactly that kind of output, the realistic planning horizon is shorter than the last cycle’s was.
Why The Balance Sheets Look Different
The nuance keeping Paulson off the doom stage is balance sheets. The hyperscalers spending on GPUs throw off enormous cash flows and carry very little of the exotic leverage that broke Lehman. Information sector profits ran at $352.5 billion in the first quarter of 2026, up from $265 billion two years earlier, and total corporate profits grew 12.8% year over year. That is a very different animal from mortgage-backed paper stacked forty-to-one.
Markets seem to agree. The VIX closed at 15.86 on August 3, back in complacency territory after a March spike to 31.05. The 10-year minus 2-year spread sits at 0.45%, positive and not signaling recession. What he is describing is a painful mark-to-market correction in AI equities, paired with a fast, uneven jobs shock.
What To Watch
Paulson’s actual ask was for policy. “One of the things that we as a government need to focus on… what kinds of plans do we need to put in place? So when the jobs are lost… we’ve got programs to deal with.” Build the retraining and safety-net rails now, while markets look calm, because you cannot pour concrete during the storm. You can see the underlying data in FRED’s series for JOLTS job openings, which is where softness will show first.
For the retail investor scrolling their brokerage app. Do not confuse hyperscaler solvency with your own. The companies at the center of the AI trade can absorb a repricing. Your career, and the concentrated position in your 401(k) tracking the same names, might absorb it less gracefully. Watch three things. Watch whether vendor financing to AI customers keeps growing as a share of reported revenue. Watch JOLTS by white-collar category, not the headline. Watch the yield curve. Paulson’s message is to know which rope you are tied to.
Contact [email protected] for any questions or corrections.
Omor Ibne Ehsan is a writer at 24/7 Wall St. He is a self-taught investor with a focus on growth and cyclical stocks that have strong fundamentals, value, and long-term potential. He also has an interest in high-risk, high-reward investments such as cryptocurrencies and penny stocks.
