The US market narrative undergoes a "dramatic reversal": shifting from "AI-driven deflation and controllable US debt" to "AI squeezing the bond market, and Bassen unable to control long-term interest rates"

The mainstream narrative surrounding the US economy has undergone a fundamental reversal within one year, and the truly underestimated systemic risk may not lie in the bond market or fiscal issues, but instead lurks within the AI ecosystem itself.

According to Wind Trading Desk, after a series of intensive client meetings along the US East Coast, Deutsche Bank's Global Head of FX Research, George Saravelos, released a report highlighting that last year's mainstream market narrative was "AI has a deflationary impact, and the US Treasury would not allow bond sell-offs," but this year the narrative has completely flipped to "AI (debt-financing) is an important driver pushing yields higher, and the US Treasury has lost control over long-term rates."

He also noted that there has been discussion in the market about suspending 20-year US Treasury issuance and drastically shortening supply duration, but most clients remain extremely cautious about fixed income assets.

Saravelos disagrees with this. He believes that the narrative pendulum has swung "too far," and the market's currently most underestimated tail risk is the AI ecosystem itself encountering problems—be it a safety incident, a failed IPO, or disappointing revenues. Should this risk materialize, it would put significant pressure on the US dollar and provide strong support for the bond market. Yet the current market pricing barely reflects this risk at all.

Narrative Reversal: From "AI Lowers Inflation" to "AI Drives Up Yields"

According to the Deutsche Bank report, one year ago, Wall Street's prevailing consensus rested on two mutually reinforcing judgments: first, AI technology would reduce inflation by improving productivity; second, the US Treasury had the ability to manage the bond market, and long-term rates would not spiral out of control.

However, this narrative framework has been completely abandoned by the market over the past year. The current mainstream view holds that large-scale bond issuance by AI-related companies is now a key force pushing US Treasury yields higher, and that the US Treasury—whether via Besent or other policy tools—is no longer able to suppress the rise in long-term rates.

In the report, Saravelos recorded a specific signal circulating among clients: discussions in the market about suspending 20-year US Treasury issuance and drastically shortening overall supply duration. The mere existence of this rumor reflects deep market anxiety about the imbalance between long-term supply and demand. Nevertheless, Saravelos clearly states that he personally believes the narrative pendulum has "swung too far," and that the current pessimism toward fixed income could be an overcorrection.

French Risk: Additional Pressure on the Euro

During these client visits, the issue of France emerged as another important line of discussion. Saravelos noted that client sentiment on France was overwhelmingly pessimistic, mirroring the recent views of economist Paul Krugman, and tended to liken France’s current fiscal predicament to that of the Eurozone sovereign debt crisis between 2010 and 2015.

Deutsche Bank holds a different perspective. Saravelos said he explained in meetings why he does not agree with this comparison, though he acknowledged that, given last week's significant market dislocation, it will take time for market confidence to recover. He characterized the French issue as "an additional euro-specific pressure that was not originally expected this year," suggesting this factor will continue to weigh on the euro in the short term.

The Real Risk of AI: Data Centers Under Pressure, Revenue Logic Uncertain

On the AI discussion, Saravelos described an unexpectedly deep conversation. He participated in a panel with data center industry experts, where attendees unanimously expressed three concerns: declining underwriting standards, uncertainty about whether energy supply will meet construction demand, and whether the anticipated massive expansion in inference power will be supported by sufficient demand and revenue.

The report cites a research paper from Brookings Institution indicating that AI labs need to generate nearly $4 trillion in revenue to make data center investments pay off. This figure offers a quantitative benchmark for the sustainability of the current AI investment boom.

The Most Underestimated Systemic Risk: The Potential Collapse of the AI Ecosystem

Saravelos explicitly identifies what he believes is the current market's most dangerous blind spot in his report. He writes that the greatest systemic risk for the market next year is not France, but something going wrong with the AI ecosystem—whether that takes the form of a safety incident, a failed IPO, or disappointing revenue data.

He stressed that concentration risk in the current AI sector is extremely high. Should any of these risk events occur, the market impact would be: significant negative pressure on the US dollar and strongly positive support for the bond market. However, Saravelos believes that market pricing is currently severely underestimating this event risk.

This judgment stands in sharp contrast to the current mainstream market narrative—while the market broadly worries about AI-driven debt issuance pushing yields higher, Deutsche Bank is highlighting a tail risk in the opposite direction: the collapse of the AI narrative itself.