Deutsche Bank: The bursting of the AI bubble may become the biggest systemic risk in the market next year; U.S. Treasury bonds are expected to attract safe-haven funds.
George Saravelos, Global Head of FX Research at Deutsche Bank, stated that since the beginning of this year, global investors' pessimism toward the bond market has become excessive. He believes the market may be underestimating the likelihood of large-scale capital flows into the bond market if significant risks emerge in the artificial intelligence industry.
According to Zhitong Finance APP, George Saravelos, Global Head of FX Research at Deutsche Bank, stated that since the beginning of this year, pessimism among global investors towards the bond market has become excessively widespread, while the market may be underestimating the possibility of significant capital inflows into the bond market following a major risk event in the artificial intelligence (AI) industry. He believes the biggest systemic risk facing financial markets next year may not be the European debt issue, but an unexpected shock within the AI ecosystem.
In a report published on Friday, Saravelos pointed out that recent conversations with U.S. clients show investors are generally concerned that the AI investment frenzy is pushing bond yields higher, and believe the U.S. Treasury has gradually lost control over long-term Treasury yields.
Against this backdrop, the market has begun speculating that the U.S. Treasury might suspend issuance of 20-year Treasuries to ease pressure on the long-term bond market. At the same time, the recent sharp sell-off in French government bonds has also significantly eroded investor confidence. Saravelos noted that clients he interacted with are almost universally pessimistic about French bonds.
However, he believes current market sentiment may have swung to another extreme.
Saravelos pointed out that last year’s market perception of AI and bonds was completely different than now. At that time, investors generally believed that advances in AI technology would improve productivity and curb inflation, while also trusting that the U.S. Treasury could take effective measures to prevent long-term yields from rising excessively. Now, with AI infrastructure investment spurring economic growth and increasing financing needs, coupled with rising energy prices triggering inflation concerns, the market has begun to view AI as an important factor driving bond yields higher.
Saravelos stated that market views on the outlook for bonds may have become overly pessimistic, overlooking potential drivers for a future bond market rebound. He believes the real risk underestimated by the market is the possibility of a major negative event occurring within the AI industry itself. Saravelos noted, "The biggest systemic risk the market faces next year is not France, but 'something going wrong somewhere' in the AI ecosystem, such as a safety incident, IPO failure, or company revenue missing expectations."
He emphasized that the concentration risk in AI-related assets is extremely high at present. Should the AI investment frenzy encounter a major setback, it could prompt investors to reassess the valuations of related assets and shift funds from risk assets to bonds and other safe-haven assets. In his view, such an event could be clearly negative for the dollar but significantly benefit the bond market—a risk that financial markets have yet to fully reflect.
Recently, U.S. Treasury yields have continued climbing to their highest levels in decades. The Iran war has caused energy prices to surge, intensifying market concerns over inflation. Meanwhile, the U.S. economy remains robust, with AI infrastructure construction and resulting massive investments being one of the key drivers. However, while the AI investment boom supports economic growth, it has also increased financing demand, meaning investors must absorb more debt supply. Under the dual impact of inflation pressure and rising bond supply, U.S. long-term Treasuries remain under pressure.
As the bond market undergoes sell-offs, speculation is mounting that the U.S. Treasury may adjust the structure of bond issuance. Currently, one closely watched proposal is to reduce the scale of long-term bond issuance in favor of increased short-term debt financing. In particular, since the yield on 20-year Treasuries is higher than bonds of adjacent maturities, cutting or even canceling 20-year issuance has become one of the more radical options under discussion in the market.
However, the practical effect of this plan remains subject to debate. Some market participants, including strategists from BNP Paribas, believe that suspending 20-year Treasury issuance may not effectively lower long-term borrowing costs and could even have the opposite effect.
Saravelos, on the other hand, believes that rather than easing upward yield pressure by adjusting Treasury issuance, the repricing of AI-related risks could become an important catalyst for a bond market rebound. If a major negative event occurs in the AI industry, investors may rapidly reduce risk exposure and increase allocations to U.S. Treasuries, thereby pushing bond prices higher and yields lower.
Regarding the recent market attention on French bond risks, Saravelos similarly believes that some investor concerns may be exaggerated. He stated that Deutsche Bank has explained to clients why the current volatility in the French bond market should not be simply equated with the European sovereign debt crisis of 2010 to 2015.
However, Saravelos admitted that given the severe volatility seen in the French bond market last week, it will still take time for investor confidence to recover. He pointed out that turbulence in the French debt market has also brought new downward pressure on the euro, a risk that was not originally part of Deutsche Bank's expectations for this year's market trends.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
The "Digital Metrics Game" Behind the Trillion-Dollar Valuation: OpenAI and Anthropic's Revenues Are Actually Incomparable
Anthropic counts "total revenue" in its cloud sales, while OpenAI only recognizes the "net revenue" based on profit-sharing. This discrepancy in accounting standards has directly triggered a decline in tech stocks. Even more striking, the highly touted "annualized revenue" is significantly inflated, with OpenAI's actual revenue expected to be only half of this figure. Anthropic's actual revenue also exhibits nearly a 50% gap compared to its reported numbers.
Meanwhile raises $37.5 million as Bitcoin life insurance demand rises abroad
Nvidia quietly redefines free cash flow, is there something fishy in the $100 billion buyback plan?
Nvidia has classified its equity investments in 242 companies, including OpenAI and Anthropic, as "strategic use" and deducted them from its free cash flow. After this adjustment, free cash flow for the first half of the fiscal year plummeted from the official figure of $69.9 billion to $21.7 billion, a decrease of nearly 70%. Meanwhile, the company added $24.9 billion in long-term debt to support buybacks, casting doubt on the sustainability of its $235 billion buyback plan.
NEAR, XRP, SHIB and BTC retreat from key levels, bulls target new highs
